US20200099619A1 - Packet header field extraction - Google Patents
Packet header field extraction Download PDFInfo
- Publication number
- US20200099619A1 US20200099619A1 US16/695,053 US201916695053A US2020099619A1 US 20200099619 A1 US20200099619 A1 US 20200099619A1 US 201916695053 A US201916695053 A US 201916695053A US 2020099619 A1 US2020099619 A1 US 2020099619A1
- Authority
- US
- United States
- Prior art keywords
- packet
- header
- network switch
- match
- circuitry
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Images
Classifications
-
- H04L45/7457—
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/54—Organization of routing tables
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/64—Routing or path finding of packets in data switching networks using an overlay routing layer
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/72—Routing based on the source address
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/74—Address processing for routing
- H04L45/745—Address table lookup; Address filtering
- H04L45/74591—Address table lookup; Address filtering using content-addressable memories [CAM]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L49/00—Packet switching elements
- H04L49/10—Packet switching elements characterised by the switching fabric construction
- H04L49/101—Packet switching elements characterised by the switching fabric construction using crossbar or matrix
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L49/00—Packet switching elements
- H04L49/15—Interconnection of switching modules
- H04L49/1515—Non-blocking multistage, e.g. Clos
- H04L49/1546—Non-blocking multistage, e.g. Clos using pipelined operation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/22—Parsing or analysis of headers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/38—Flow based routing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/42—Centralised routing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/74—Address processing for routing
Definitions
- OpenFlow Software Defined Networking
- the control plane is physically separated from the forwarding plane and communicates with the forwarding plane through an interface (e.g., the OpenFlow protocol).
- OpenFlow is an evolving networking standard that uses a match-action paradigm for network packet switching.
- the typical hardware switches that were (and still are) used to implement the match-action paradigm are not quite flexible since they process only a fixed set of fields with a limited repertoire of packet processing actions.
- a reconfigurable match table (RMT) architecture that overcomes the aforementioned shortcomings has recently been introduced to the market. This new architecture could be improved to make the hardware switch operate more efficiently.
- the hardware switch of some embodiments includes, among other elements, an ingress pipeline and an egress pipeline.
- Each pipeline includes a parser, a match-action unit (MAU), and a deparser.
- the parser separates the packet header data (e.g., Ethernet headers, IPv4 headers, TCP headers, etc.) from the payload of the packet and determines which fields of each packet header (e.g., source port address of the TCP header, destination port address of the TCP header, etc.) might be processed by the MAU.
- packet header data e.g., Ethernet headers, IPv4 headers, TCP headers, etc.
- the parser of some such embodiments (1) populates a primary packet header vector (PPHV) with only the fields of the different headers that may be used by the MAU, and (2) populates a secondary packet header vector (SPHV) with the rest of the fields of the different headers (i.e., the fields that will not be processed by the MAU).
- PPHV primary packet header vector
- SPHV secondary packet header vector
- the PPHV and SPHV each includes a set of registers that stores the packet header fields during processing by the MAU.
- the parser delivers the populated PPHV to the MAU, which includes a set of match-action stages for matching the different header fields of the packet against the different match tables and applying the required actions to the packet by modifying the packet header data (e.g., assigning the packet to an output port and queue, sending the packet to a network controller, dropping the packet, etc.).
- the parser of some such embodiments also delivers the populated SPHV directly to the deparser (without having the MAU process the SPHV) along with the payload of the packet.
- the deparser reassembles the packet with (1) the processed header data (e.g., modified header fields) of the PPHV that the deparser receives from the MAU, (2) the unprocessed header data stored in the SPHV that the deparser receives from the parser directly, and (3) the payload of the packet that the deparser also receives from the parser.
- the deparser sends the packet to a queuing system to be subsequently forwarded to the egress pipeline of the switch.
- the deparser sends the packet (with a header that might or might not have been modified) out of the switch.
- the parser of some embodiments determines which fields of each packet header may be processed and which fields will not be processed by the MAU, based on the information the parser receives from the packet headers themselves (e.g., the EtherType field of the packet, etc.), and based on configuration data received from the control plane.
- a compiler in the control plane receives the data required for configuring the pipeline (e.g., through a programing language code), generates a set of configuration data, and distributes the generated data to a configurator module (also part of the control plane). The configurator module then distributes the configuration data to both the parser and MAU of the pipeline in the switch.
- the parser of some embodiments based on the information it receives from the configurator, determines which sets of packet headers to extract from the packet and how to extract these packet headers.
- the parser analyzes the packet headers as it extracts them according to the configuration data configured by the control plane.
- This configuration data represents a parse graph that identifies how to parse the packet based on values of certain packet header fields as well as table flow and control flow graphs used to generate the configuration data for the MAU, which describe how packets will be processed through the stages of the MAU.
- the parser determines iteratively (i) whether each packet header field should be placed in the PPHV or the SPHV and (ii) how to extract the next set of packet header fields.
- the EtherType field of the layer 2 header of some embodiments specifies how to extract the layer 3 header (e.g., whether the packet is IPv4 or IPv6, etc.).
- a TCP packet header typically has a minimum length of five words (twenty bytes).
- the MAU of the pipeline processes only four bytes of the TCP header bytes (i.e., the source and destination port addresses) and the rest of the fields in the TCP header (sixteen bytes) are not used by any of the stages of the MAU. In such cases, by separating the unused data from useful data, the parser does not waste any of the valuable resources of the MAU to process the other sixteen bytes of the TCP header.
- separating the PPHV from the SPHV enables the parser to parse deeper into packets while sending the same amount of data to the MAU (via the PPHV).
- sending some of the unused fields into the SPHV allows the PPHV to carry only the more useful fields, and include deeper packet headers.
- newer tunneling protocols such as GENEVE add variable length option fields to the packet header, and may be quite long.
- packets sent using multi-layer tunneling may have significant amounts of packet header data that may need to be extracted in order to reach the layer 4 (e.g., TCP) data.
- the PPHV can include packet header fields that are located deeper into the packet, without sending additional data to the MAU.
- FIG. 1 illustrates a block diagram of a hardware switch and a block diagram of an ingress/egress interface of the hardware switch.
- FIG. 2 illustrates a block diagram that shows that the parser produces a secondary packet header vector (SPHV), besides the primary packet header vector (PPHV), that bypasses the MAU and is directly delivered to the deparser.
- SPHV secondary packet header vector
- PPHV primary packet header vector
- FIG. 3 illustrates an ingress/egress pipeline architecture that separates the participating header fields of the different packet headers from nonparticipating packet header fields of a packet.
- FIG. 4 conceptually illustrates the reason that the parser of some embodiments does not store the entire packet headers in the PPHV to be processed by the MAU.
- FIG. 5 conceptually illustrates a block diagram of a parser that is configured by a compiler to separate the participating (in an MAU process) packet header fields from nonparticipating header fields.
- FIG. 6 conceptually illustrates a process of some embodiments that populates the PPHV and SPHV with different header fields of a packet header.
- FIG. 7 conceptually illustrates a compiler of some embodiments for generating parser and MAU configuration data.
- FIG. 8 conceptually illustrates a process performed by the compiler of some embodiments to generate parser configuration data for a hardware switch that parses packet header data into a PPHV and a SPHV.
- FIG. 9 conceptually illustrates an electronic system with which some embodiments of the invention are implemented.
- Some embodiments provide a novel packet processing pipeline that enhances the match-action packet header processing performed by reconfigurable hardware switches.
- the hardware switch of some embodiments includes, among other elements, an ingress pipeline and an egress pipeline. Each of these pipelines includes a parser, a match-action unit (MAU), and a deparser.
- MAU match-action unit
- FIG. 1 illustrates a block diagram of a hardware switch 110 and a block diagram of an ingress/egress pipeline of the hardware switch. More specifically, the hardware switch 110 includes an ingress pipeline (datapath) 115 , a queuing system (e.g., a data buffer) 120 , and an egress pipeline 125 .
- an ingress pipeline datapath
- queuing system e.g., a data buffer
- the ingress pipeline 115 of some embodiments receives a packet 105 (e.g., through an I/O module), parses the packet into a packet header vector (PHV), sends the PHV through a set of match and action stages which may modify the PHV, deparses the packet headers back from the PHV into packet format, and enqueues the packet in a centralized data buffer (i.e., the data buffer 120 ).
- PGW packet header vector
- the ingress pipeline 115 receives a packet 105 (e.g., through an I/O module), parses the packet into a packet header vector (PHV), sends the PHV through a set of match and action stages which may modify the PHV, deparses the packet headers back from the PHV into packet format, and enqueues the packet in a centralized data buffer (i.e., the data buffer 120 ).
- a centralized data buffer i.e., the data buffer 120
- the queuing system 120 receives the packets that are processed by the ingress pipeline and provides a large shared buffer (storage) that accommodates the queuing delays due to oversubscription of the output channels of the ingress deparser.
- the data buffer stores packet data, while pointers to that data are kept in different queues per channel. Each channel in turn requests data from the common data buffer using a configurable queuing policy. When pointers to packets reach the head of the queues, the packets are read out of the data buffer 120 into the egress interface 125 .
- the egress interface 125 receives the packet from the queuing system 120 , separates the packet payload from the packet headers, stores the packets headers in a PHV, sends the PHV through an ingress pipeline of match and action stages, deparses the packet headers back from the PHV into packet format, and sends the packet 130 to an appropriate output port of the switch 110 to be driven off the switch (e.g., through one of the output channels).
- the output packet 130 may be the same packet as the input packet 105 (i.e., with identical packet headers), or it may have different packet headers compared to the input packet 105 based on the actions that are applied to the packet headers in the ingress and egress pipelines (e.g., different header field values for certain header fields and/or different sets of header fields).
- the illustrated blocks in switch 110 are exemplary only.
- the ingress, queuing, and egress blocks are simplified for simplicity of the description.
- the input signals are received by many different input channels (e.g., 64 channels) and the output signals are sent out of the switch from different output channels (e.g., 64 channels).
- the illustrated switch only one parser interface is shown for the ingress/egress pipeline 140 , some embodiments employ numerous parser blocks (e.g., 16 parser blocks) that feed the match-action unit (MAU) in each pipeline.
- MAU match-action unit
- FIG. 1 also shows a block diagram 140 of an ingress/egress interface of the hardware switch 110 .
- the interface includes a pipeline with three different units, namely a parser unit 145 , a match-action unit (MAU) 150 , and a deparser unit 155 .
- the parser 145 receives the incoming packet data and produces a PHV as its output. In other words, the parser 145 separates the packet headers from the packet payload by extracting the different packet headers and storing them in the PHV.
- the PHV includes a set of different size registers or containers.
- the PHV includes sixty-four 8-bit registers, ninety-six 16-bit registers, and sixty-four 32-bit registers (for a total of 224 registers containing 4096 bits), though other embodiments may have any different numbers of registers of different sizes.
- the parser 145 stores each extracted packet header in a particular subset of one ore more registers of the PHV. For example, the parser might store a first header field in one 16-bit register and a second header field in a combination of an 8-bit register and a 32-bit register (e.g., if the header field is 36 bits long).
- the PHV provides the input data to the match tables of the MAU.
- the MAU 150 includes a set of match-action stages (e.g., 32 match-action stages in some embodiments), each of which matches a particular set of header fields against a match table and takes an action based on the result of the match (e.g., assigning the packet to an output port and queue, dropping the packet, modifying one or more of the header fields, etc.).
- the MAU 150 and its different match-action stages are described in more detail below by reference to FIG. 3 .
- the PHV that the MAU outputs might include the same header data as the PHV that the MAU received from the parser, or the output PHV might contain different data than the input PPHV.
- the output PHV is then handed to the deparser 155 .
- the deparser 155 reassembles the packet by putting back together the output PHV (that might or might not have been modified) that the deparser receives from the MAU 150 and the payload of the packet that the deparser receives directly from the parser 145 in some embodiments.
- the deparser then sends the packet out of the ingress/egress pipeline (to the queuing system 120 or out of the switch, depending on whether it is the deparser for the ingress pipeline or the egress pipeline).
- FIG. 2 illustrates a block diagram similar to the block diagram 140 of FIG. 1 , except that this figure shows that the parser also produces a secondary packet header vector (SPHV), or tag-along PHV, that bypasses the MAU's pipeline and is directly delivered to the deparser. More specifically, after receiving a packet, the parser 145 of some embodiments produces two different packet header vectors, a PPHV and a SPHV, as its output.
- SPHV secondary packet header vector
- the parser separates the packet header data (e.g., Ethernet, IPv4, TCP, etc.) from the payload 220 of the packet and determines which fields of each packet header (e.g., source port address of the TCP header, destination port address of the TCP header, etc.) may be processed by the MAU (i.e., might be matched against, or might be modified by an action). Packets may contain packet headers and a payload, though the divide between the two depends on how deep into the packet an element is inspecting.
- packet header data e.g., Ethernet, IPv4, TCP, etc.
- the layer 3 e.g., IPv4, IPv6, etc.
- layer 4 e.g., TCP, UDP, etc.
- the layer 2 , layer 3 , and layer 4 portions are all part of the headers, and the remainder of the packet is treated as payload.
- Each header e.g., the layer 2 header, layer 3 header, layer 4 header, tunnel encapsulations, etc.
- the source and destination IP addresses of an IP header e.g., IPv4 header
- the source port and destination port addresses of a TCP header are two different header fields of the TCP header.
- Packet is used throughout this application to refer to a collection of bits in a particular format sent across a network. It should be understood that the term “packet” may be used herein to refer to various formatted collections of bits that may be sent across a network. A few examples of such formatted collections of bits are Ethernet frames, TCP segments, UDP datagrams, IP packets, etc.
- the parser of some embodiments (1) populates the primary packet header vector (PPHV) 230 with only the participating header fields of different headers (i.e., the fields that may be processed by the MAU), and (2) populates the secondary packet header vector (SPHV) 240 with the nonparticipating fields of the different packet headers (i.e., the fields that will not be processed by the MAU).
- PPHV primary packet header vector
- SPHV secondary packet header vector
- Each of the PPHV and SPHV includes a set of registers that stores the packet header fields in some embodiments.
- the PPHV 230 of some embodiments includes a combination of different size registers (e.g., with a total capacity of 4 k bits) that hold the packet header field values that may be processed the MAU.
- the SPHV 240 of some embodiments includes a set of the same or different size registers (e.g., shift registers, FIFO registers, etc.) for holding non-used header field values (e.g., with a total capacity of 2 k bits).
- the parser delivers the populated PPHV 230 to the MAU 150 , which includes a set of match-action stages for matching the different header fields of the packet against the different match tables and applying the required actions to the packet by modifying the packet header data (e.g., assigning the packet to an output port and queue, sending the packet to a network controller, dropping the packet, modifying the header field values of the PPHV, etc.).
- the parser of some such embodiments also delivers the populated SPHV 240 directly to the deparser 155 along with the payload of the packet and without having the MAU 150 process the SPHV 240 .
- the deparser 155 reassembles the packet with (1) the processed header data (that might or might not have been modified) which the deparser receives from the MAU 150 , (2) the unprocessed header data stored in the SPHV 240 that the deparser receives from the parser 145 , and (3) the payload 220 of the packet that the deparser also receives from the parser 145 .
- the deparser sends the packet to a queuing system to be subsequently forwarded to the egress pipeline of the switch.
- the deparser sends the packet (with a header that might or might not have been modified) out of the switch.
- the deparser recognizes the location in the reconstructed packet at which to place each value from the different registers (as an example, the 13 bits of the IPv4 fragment offset field that might be located in register X of the SPHV are inserted directly after the 3 bits of the flags field from register Y of the SPHV and directly before the 8 bits of the time to live field from register N of the PPHV).
- FIG. 3 illustrates an ingress/egress pipeline architecture 300 that separates the participating header fields (i.e., the header fields that may participate in the different match-action stages of the MAU) of the different packet headers from non-participating packet header fields of a packet. More specifically, the figure shows that the pipeline 300 includes a parser 310 , an MAU with multiple match-action stages 1 through m (where m is an integer such as 32 in some embodiment), and a deparser 360 . As illustrated, each match-action stage includes a Xbar 315 , a set of exact match tables 320 , a set of ternary match tables 325 , and a Very Long Instruction Word (VLIW) action engine 330 .
- VLIW Very Long Instruction Word
- the parser 310 receives a packet (e.g., through a set of input modules of the switch) and extracts the packet headers (e.g., Ethernet header, IPv4 header, IPv6 header, TCP header, UPD header, tunnel encapsulation headers, etc.) from the body of the packet.
- the parser 310 then separates the header field values that may be processed by the MAU from the rest of the header field values in each packet header.
- the header fields of a packet header are the different segments of the packet header, each of which stores a particular value representing a particular piece of data.
- an IPv4 packet header includes a source IP address header field, a destination IP address header field, an IPv4 identification field header field, a time-to-live header field, etc.
- the parser of some embodiments determines which header fields are participating header fields (i.e., the fields that may be processed by the MAU) and which header fields are non-participating header fields based on configuration data that the parser receives from a configurator module in the control plane.
- the configurator module receives the configuration data from a compiler (also part of the control plane) which generates the configuration data based on a set of instructions that the compiler receives (e.g., from a user designing the functionality of the hardware switch).
- the compiler and configurator modules of some embodiments are described in more detail below by reference to FIG. 5 .
- the parser 310 separates the header fields that may be processed by the MAU from the rest of the header fields of the different packet headers, the parser stores the participating header fields in the PPHV (in different registers of the PPHV) and delivers the PPHV to the MAU for further processing of the header data.
- the MAU of some embodiments includes a set of different stages, each of which is configured to process a particular set of header fields of the PPHV. In some embodiments, each of these stages is reconfigurable by a controller that is responsible for configuring (and reconfiguring) the stages of the match-action units of the hardware switch to match against various different packet header fields and
- the output PPHV from the parser 310 is delivered to both the Xbar 315 and the VLIW action engine 330 .
- the Xbar 315 of each stage includes a set of one or more crossbars (e.g., two crossbars) that identifies which field or fields of the PPHV are used by the match tables of the corresponding stage of the MAU, and delivers the identified fields to the exact match tables and ternary match tables of the stage.
- the Xbar 315 of that particular stage takes the IP source field from the PPHV register(s) and delivers the field to the match tables.
- the Xbar 315 operates at a bit level. That is, a particular header field that contains 20 bits might be carried by a 16-bit register and a portion of an 8-bit register of the PPHV. The Xbar of some such embodiments, therefore, passes only 20 bits of the data that is carried by the combination of the 16-bit register and the 8-bit register to its corresponding match tables.
- each match-action stage of the MAU includes a set of ternary match tables 325 and a set of exact match table 320 .
- Some embodiments implement (1) the ternary match tables with ternary content addressable memory (TCAM) units, and (2) the exact match tables with static random-access memory (SRAM) units, that operate as hash tables.
- Each one of the ternary match tables and exact match tables of some embodiments includes a number of entries (e.g., flow entries) each of which matches against the values of one or more packet header fields and instructs the action engine 330 of the same stage to take a particular action (e.g., an action stored in a SRAM unit of an action table).
- the ternary match table stores packet header data that can include ternary bits (i.e., bits with values of 0, 1 or wildcard). Wildcarding the bits allow a single table entry to match a wide variety of packet header fields.
- the exact match table of some embodiments does not allow any wildcarding, hence the packets must exactly match the table entries.
- the information includes, among other data, an instruction memory address for the action engine 320 and an action memory address and size.
- the desired actions include simple actions such as setting a field in the PPHV to a particular value or dropping a packet, as well as complex operations, such as adding encapsulations to the packet (e.g., Provider Backbone Bridges (PBB) encapsulation, GRE or GENEVE tunneling, etc.).
- PBB Provider Backbone Bridges
- the VLIW action engine 330 includes different action operations that may receive their sources from packet header fields (which is why a copy of the PPHV registers are directly delivered to the action unit), or from an action memory (not shown).
- An action indicated by a match may be simple and require only a small amount of data from the action memory or complex and require a large amount of data.
- the action engine 330 When the action engine 330 receives the action data (from the action memory or the PPHV), the action engine takes the required action which may include modifying a header field in the PPHV (e.g., to decrement the TTL field of the IP header, change the address of the packet, etc., or taking another action such as assigning the packet to an output port and/or queue, sending the packet to the controller, dropping the packet, etc.
- the required action is taken (e.g., when a match is found) in each stage of the MAU, the PPHV (as modified by the action, if such modifications are performed) is delivered to the next stage of the MAU for further processing.
- each successive PPHV that includes the packet header data of one packet progresses through the pipeline of match-action stages one clock cycle behind the previous PPHV.
- the last match-action stage of the MAU after processing the PPHV, delivers the processed PPHV to the deparser 360 .
- the deparser 360 of some embodiments reassembles the packet by putting together the processed header fields that are received from the MAU, with the unprocessed header fields that are received directly from the parser 310 in the SPHV 355 and the payload 350 of the packet that is also received directly from the parser 310 .
- the packet is then either sent out of the switch (if the deparser is an egress deparser), or sent to the queuing system (if the deparser is an ingress deparser) for further processing by the egress pipeline of the switch.
- FIG. 4 conceptually illustrates the reason that the parser of some embodiments does not store the entire packet headers in the PPHV to be processed by the MAU.
- This figure shows that a packet 410 is being received by the parser 420 to parse the packet headers into the PPHV 430 .
- not every header field of each packet header is needed by the MAU stages of the upcoming ingress or egress pipeline to which the parser sends the PPHV.
- some of the packet header fields will (i) not be matched against by any of the match entries of the match tables in the pipeline and (ii) not be modified by any possible action entry that could be performed in the pipeline.
- the parser 420 extracts each packet header from the packet 410 , it determines which of the header fields of the packet header might be processed by at least one of the match-action stages of the MAU.
- a packet header 415 (e.g., an Ethernet header, IPv4 header, etc.) of the packet 410 includes several participating header fields 450 that the MAU is configured (e.g., by a configurator module of the control plane) to potentially process.
- the packet header 415 also includes several other non-participating header fields 460 that the MAU is not configured to process.
- the parser 420 when the parser 420 extracts a particular packet header from the packet 410 , the parser must extract the entire contiguous packet header at once (i.e., the parser cannot leave certain fields of a packet header in the payload while placing the other fields of the packet header in the PHV). Because the different participating header fields of the packet header are often not placed next to each other in the packet header (as illustrated in the figure), the parser of some embodiments separates these participating header fields from nonparticipating fields during extraction of the packet header.
- the MAU might be configured to process only a particular set of header fields in a UDP packet header, which may not be the first two header fields of the packet header (i.e., the source and destination ports).
- the parser locates the particular header fields in the set, pulls these fields out of the packet header, and stores the header fields in the PPHV.
- the other nonparticipating header fields that are also extracted from the packet have to be dealt with as well. Therefore, in some embodiments, the parser 420 looks at each header field in the packet header and determines whether the identified header field might be processed by the MAU or will definitely not be processed by the MAU.
- none of the fields of a packet header might be placed in the PPHV, but the parser would extract these fields and place them in the SPHV in order to continue deeper into the packet and place the fields of a higher-layer packet into the PPHV.
- the parser 420 determines that the header field is one of the participating header fields, the parser stores the header field in the PPHV 430 (i.e., in a particular set of registers of the PPHV 430 designated for that header field). On the other hand, if the parser 420 determines that the identified header field is not supposed to be processed by the MAU, the parser stores the header field in the SPHV 440 to be subsequently sent directly to the deparser (not shown) of the pipeline without getting processed.
- the parser of some embodiments determines which fields of each packet header may be processed and which fields will not be processed by the MAU, based on the information the parser receives from the packet itself (e.g., by one or more particular packet header of the packet), and based on the configuration data received from a compiler in the control plane.
- the compiler receives the data required for configuring the pipeline (e.g., through a programing language code), generates a set of configuration data, and distributes the generated data to a configurator module (also in the control plane).
- the configurator module then distributes the configuration data to both parser and MAU of the pipeline in the switch (e.g., at run-time or during setup time).
- FIG. 5 conceptually illustrates a block diagram 500 of a parser that is configured by a compiler to separate the participating packet header fields from non-participating header fields.
- this figure shows a parser 510 , a compiler 520 , a configurator 530 , and an MAU 540 .
- the parser 510 of some embodiments includes an extractor state machine 560 .
- the compiler 520 and configurator 530 are separate from the hardware switch (i.e., not part of the chip), whereas the parser 510 and MAU 540 are part of the hardware switch circuitry.
- the configurator 530 in some embodiments, represents a runtime network controller stack that configures the hardware switch, while the compiler 520 operates at an earlier stage to generate the configuration data and provide this configuration data to the configurator 530 .
- the compiler 520 generates configuration data for the hardware switch based on a switch configuration received by the compiler 520 (e.g., through code written by a user, such as P4 code for designing switch operations).
- the configuration data specifies how to program the match-action stages of the MAU 540 to implement the switch configuration specified by the received configuration code, as well as how to program the extractor state machine 560 of the parser to extract the packet header fields correctly and to place the packet header fields into the appropriate PPHV or SPHV registers.
- the compiler 520 passes the configuration data to the configurator 530 , which provides the configuration data to the MAU 540 so that the MAU can implement the appropriate switch configuration.
- the configurator 530 provides the parser configuration data generated by the compiler 520 to the extractor state machine 560 so that the extractor state machine can extract header fields from the packets into the registers of the PPHV and the SPHV.
- the extractor state machine 560 receives the packet 585 and (i) separates the packet headers into the PPHV 570 and the SPHV 575 and (ii) separates out the payload 580 of the packet.
- the extractor state machine begins at the front of the packet (i.e., the outermost headers) and works inward, extracting each packet header field into either the PPHV or the SPHV. For certain header fields, the extractor state machine looks at the value and compares the value to configuration entries (e.g., TCAM entries), which specify how the next set of packet header fields should be extracted.
- configuration entries e.g., TCAM entries
- the extractor state machine 560 follows a parse graph that specifies numerous states at which a particular packet header field is extracted into either the PPHV or the SPHV, as well as what state of the parse graph to proceed to next. That is, in some embodiments each state specifies which bits of the header to extract for a particular header field, which register(s) of the PPHV or SPHV to put those bits into, and to which state to proceed next.
- the extractor state machine 560 of some embodiments may also read the data of certain header fields in order to determine which path to take through the set of states of the parse graph.
- the EtherType field of an Ethernet header indicates the type of protocol encapsulated by the Ethernet header (e.g., IPv4, IPv6, ARP, etc.).
- IPv4, IPv6, ARP, etc. The different fields of these protocols will be formatted differently, so for the extractor state machine 560 to properly identify the next state(s) in the parse graph, it needs to know the EtherType value.
- the protocol field of an IPv4 header specifies the transport layer protocol (e.g., TCP, UDP, etc.), which will determine how those fields are extracted.
- the parse graph is implemented in the parser 530 as a set of entries in TCAMs that the extractor state machine matches.
- the packet header is received as a set of bits stored in, e.g., a buffer, and at any time during the extraction process a pointer points to a particular location in the buffer.
- the TCAM entries of the parse graph of some embodiments specify which type of header is next based on the value of certain fields of a current header (e.g., the EtherType field of an Ethernet header, the protocol field of an IPv4 header, etc.), and how to extract the header fields of the next header.
- the header field value is matched against the TCAM entries to determine the next state or set of states based on the determination as to what type of packet headers are contained in the packet.
- the parser configuration specifies which packet headers will be extracted into the PPHV and which packet headers are extracted into the SPHV.
- any field that is either (i) used as a match field of any match entry in the MAU and/or (ii) modified by any action entry in the MAU is extracted into the PPHV, with only the header fields that are not involved in any way with the MAU extracted into the SPHV.
- some of the header fields in the PPHV will not actually be used or modified by the MAU (especially if one of the earlier stages instructs the switch to drop the packet).
- the extractor state machine 560 of some embodiments only reads and matches against the values for certain header fields that identify the structure of the subsequent packet headers, so that the subsequent packet headers can be extracted properly according to the parse graph.
- a TCP packet header typically has a minimum length of five words (twenty bytes).
- the MAU of the pipeline processes only four bytes of the TCP header bytes (i.e., the source and destination port addresses) and the rest of the fields in the TCP header (sixteen bytes) are not used by any of the stages of the MAU. In such cases, by separating the unused data from useful data, the parser does not waste any of the valuable resources of the MAU to process the other sixteen bytes of the TCP header.
- separating the PPHV form the SPHV enables the parser to parse deeper into packets while sending the same amount of data to the MAU (via the PPHV).
- sending some of the unused fields into the SPHV allows the PPHV to carry only the more useful fields, and include deeper packet headers.
- newer tunneling protocols such as GENEVE add variable length option fields to the packet header, and may be quite long.
- packets sent using multi-layer tunneling may have significant amounts of packet header data that may need to be extracted in order to reach the layer 4 (e.g., TCP) data.
- the PPHV can include packet header fields that are located deeper into the packet, without sending additional data to the MAU.
- FIG. 6 conceptually illustrates a process 600 of some embodiments that populates the PPHV and SPHV with different header field values (e.g., source and destination port addresses, fragmentation offset field, transport layer ports, encapsulation addresses, etc.) of a packet header (e.g., IPv4, TCP, UDP, etc.).
- the process 600 of some embodiments is performed by an extractor state machine of a ingress parser or egress parser of a hardware switch in some embodiments, such as the extractor state machine 560 illustrated in FIG. 5 .
- the process 600 initiates by receiving (at 610 ) a packet header.
- the process receives the packet header from a packet receiver module of the parser, such as the packet receiver 550 of FIG. 5 .
- a packet header could be an Ethernet header, an IP header, a TCP header, etc.
- the process has learned (e.g., from examining a previous packet header field), the format of the received packet header (i.e., what type of packet header it is, and therefore the lengths of the different fields of the packet header.
- the process then extracts (at 620 ) the next header field of the received packet header.
- the extractor state machine points to a particular location in a buffer that holds the packet header, and the parser configuration (implementing the parse graph) identifies the length of the next packet header, starting from that particular location.
- the process extracts the specified next number of bits (and, in some embodiments, moves the pointer to the new location after the last extracted bit).
- the process determines (at 630 ) whether the extracted field may be processed by any of the match-action stages of the MAU.
- the process determines that the extracted field is a participating header field that could be used at some point during the processing of the PPHV through the different stages of the MAU, the process identifies (at 640 ) a container (e.g., one or more registers of the PPHV) in the PPHV and stores the extracted header field in the identified container of the PPHV.
- a container e.g., one or more registers of the PPHV
- the process determines that the extracted field is a nonparticipating header field, which will not processed by any match-action stage of the MAU, the process identifies (at 650 ) a container (e.g., one or more registers of the SPHV) in the SPHV and stores the extracted header field in the identified container of the SPHV.
- a container e.g., one or more registers of the SPHV
- operations 620 - 640 or 620 - 650 are all performed as a single operation.
- the determination as to whether a particular field may be processed by the MAU is made during the generation of the parse graph, (according to the configuration data) and embedded in the parse graph based on whether the parse graph instructions specify to store the particular header field in the PPHV or the SPHV.
- the extractor state machine of some embodiments extracts the header field (i.e., the next X bits of the packet header) into a specific register of either the PPHV or the SPHV, without performing any specific analysis at packet processing time as to whether the header field is a participating or non-participating header field.
- the process of some embodiments determines (at 660 ) whether any additional header fields remain to be extracted.
- the parse graph proceeds to the next state according to the instructions of the previous state, and each subsequent state identifies the next set of bits (i.e., the next header field) to extract.
- the parse graph state specifies that the extraction process is done, at which point the process ends.
- the process 600 returns to 620 .
- the process 600 may not be performed in the exact order shown and described above. Additionally, the specific operations may not be performed in one continuous series of operations, and different specific operations may be performed in different embodiments. For instance, as mentioned above, in some embodiments the process actually extracts the field directly into one of the containers of the PPHV or the SPHV, with the determination at 630 being conceptual. In other embodiments, the process 600 identifies the packet headers first and then performs the extraction of the header fields in the identified packet headers.
- the process first extracts the first packet header (e.g., based on the configuration data and the parse graph) and after identifying the packet header, extracts the different header fields of the packet header and stores them either in the PPHV or the SPHV.
- the process could be implemented using several sub-processes, or as part of a larger macro process.
- the compiler of some embodiments generates the configuration data for the different switch elements including both the MAU and the parser.
- the MAU configuration data specifies entries for match and action tables (implemented in TCAMs, or a combination of TCAMs and RAM) of the packet processing pipelines that enable the switch to modify, forward, etc. packets that it receives.
- the parser configuration data specifies how the parser processes different types of packets, to extract header fields correctly, and to place the extracted header fields into the appropriate containers of the PPHV and SPHV.
- FIG. 7 conceptually illustrates a compiler 700 of some embodiments for generating parser and MAU configuration data.
- the compiler includes a code receiver 705 , control and table flow graph generator 710 , MAU configuration generator 715 , parse graph generator 720 , and parser configuration generator 725 .
- the code receiver 705 receives code defining the switch operations (e.g., as written by a switch designer, network administrator, etc.), and passes this code to the graph generators 710 and 720 (possibly after translating the code into a different format if needed).
- the code is received as P4 code.
- the control and table flow graph generator 710 uses the received code to generate the possible flows through a set of match and action tables that the switch will implement.
- the flow graph(s) may be linear or branched, depending on the nature of the code received by the compiler.
- the flow graph generator 710 outputs a flow graph 730 , that includes numerous match-action stages.
- Each match-action stage includes a set of match entries that specify one or more packet header fields (or other data that is not actually part of the packet header, such as ingress or egress ports) over which packets will be matched, as well as corresponding action entries that specify actions to take when a packet matches the associated match entry.
- the MAU configuration generator 715 uses the control and table flow graphs to generate the MAU configuration entries, in some embodiments.
- the parse graph generator 720 of some embodiments uses the received code to generate a parse graph specifying the possible paths that a parser can take to parse a received packet. Though shown here as generated from the code, in some embodiments the parse graph is information stored with the compiler, which specifies how to parse any packet. In other embodiments, some of the information (how any packet might be parsed) is stored with the compiler, but a parse graph for the particular switch is generated based on information in the code that indicates what types of packet may actually be processed by the switch (e.g., if all packets will only have Ethernet as the layer 2 protocol).
- the parse graph generator 720 outputs a parse graph 735 .
- the parse graph initially, conceptually describes the different types of packet headers that could be received by the switch, and the order in which those packet headers are arranged.
- the figure also shows a portion of this parse graph, including the Ethernet header.
- the parse graph 735 specifies (i) how to parse the current header into fields and (ii) how to proceed from the current header to the next header, depending on the value of one or more of the fields of the current header.
- the first eight bytes are the preamble (and start frame delimiter), the next twelve bytes are the destination MAC address and the source MAC address, and the last two bytes are the EtherType field.
- the edges of the parse graph exiting the Ethernet header node are dependent on the value of the EtherType field. For instance, if the EtherType field value is 800 , then the next header is IPv4 (with the IPv4 node of the parse graph specifying how to parse the IPv4 header); if the EtherType field value is 806 , then the next header is ARP, etc.
- the parse graph 735 continues in this manner until reaching a point at which additional data will not be parsed, and this additional data is treated as the payload. In some embodiments, the parse graph (prior to being converted into parser configuration data) continues through the entire packet.
- the parser configuration generator 725 uses the parse graph 735 as well as the control and table flow graphs 730 to generate configuration data used to configure the parser in the hardware switch pipeline.
- the parser configuration data includes TCAM entries for each edge of the parse graph that is relevant to the switch (i.e., for each packet header that the switch might need to process). Each of these entries matches over the packet header field values in the current packet header that specify the type of the next packet header (e.g., the EtherType field, the protocol field of an IP packet, etc.).
- the TCAM entries point to actions (e.g., in RAM) that specify how to parse the subsequent header (e.g., how to parse the various different fields of an IPv4 packet, an ARP packet, etc.).
- the parser configuration generator 725 determines which fields of each packet header might actually be used in the match-action unit based on the flow graphs 730 .
- any packet header field that appears in at least one match entry or is modified by at least one action entry is marked for possible use. While the switch will probably not use the field for every packet containing such a header field, it might for at least some such packets (and for certain fields, the switch may use the field for almost every packet that contains the field).
- Some embodiments annotate the parse graph to identify the participating and non-participating header fields.
- the configuration data for the parser specifies to send the participating header fields to particular registers of the PPHV and the non-participating header fields to particular registers of the SPHV.
- the configuration data also specifies when to stop parsing the packet headers and treat the remainder of the packet as payload. This decision may be made based on the amount of data already stored in the PHVs for the current traversal through the nodes of the parse graph, or because the end of relevant packet headers for the switch have been reached (e.g., the end of the innermost layer 4 headers).
- FIG. 8 conceptually illustrates a process 800 performed by the compiler of some embodiments to generate parser configuration data for a hardware switch that parses packet header data into a PPHV and a SPHV.
- the process 800 begins by receiving (at 805 ) code for programming the hardware switch.
- the code is received as P4 code, describing how the switch should function.
- the process then generates (at 810 ) a parse graph from the code and/or pre-defined packet specifications.
- the parse graph is information stored with the compiler, which specifies how to parse any packet.
- some of the information (how any packet might be parsed) is stored with the compiler, but a parse graph for the particular switch is generated based on information in the code that indicates what types of packets may actually be processed by the switch (e.g., if all packets will only have Ethernet as the layer 2 protocol).
- the parse graph of some embodiments conceptually describes the different types of packet headers that could be received by the switch, and the order in which those packet headers are arranged.
- Each node in the parse graph represents a packet header, and the parse graph specifies (i) how to parse the header into packet header fields and (ii) how to proceed from the current header to the next header, depending on the value of one or more of the fields of the current header.
- the Ethernet header for example, the first eight bytes are the preamble (and start frame delimiter), the next twelve bytes are the destination MAC address and the source MAC address, and the last two bytes are the EtherType field.
- the edges of the parse graph exiting each node are dependent on the value of one or more of the field values of the current field.
- the process 800 also generates (at 815 ) control flow and table flow graphs for defining the match-action stages of the switch.
- the flow graph(s) may be linear or branched, depending on the nature of the code received by the compiler.
- Each match-action stage in the table flow graph includes a set of match entries that specify one or more packet header fields (or other data that is not actually part of the packet header, such as ingress or egress ports) over which packets will be matched, as well as corresponding action entries that specify actions to take when a packet matches the associated match entry. These actions may include modifying certain packet header fields, in addition to other actions (e.g., dropping the packet, modifying non-header field data about the packet).
- the process 800 determines (at 820 ) whether each packet header field specified in the parse graph participates in the match-action stages.
- the process annotates (at 825 ) the parse graph with whether each header field will be placed into the PPHV or the SPHV.
- any packet header field that appears in at least one match entry or is modified by at least one action entry is marked for possible use. While the switch will probably not use the field for every packet containing such a header field, it might for at least some such packets (and for certain fields, the switch may use the field for almost every packet that contains the field). Participating packet header fields are marked to be sent to the PPHV, while non-participating packet header fields are marked to be sent to the SPHV.
- the process 800 generates (at 830 ) configuration entries for the parser based on the annotated parse graph.
- the parser configuration data includes TCAM entries for each edge of the parse graph that is relevant to the switch (i.e., for each packet header that the switch might need to process). Each of these entries matches over the packet header field values in the current packet header that specify the type of the next packet header (e.g., the EtherType field, the protocol field of an IP packet, etc.).
- the TCAM entries point to actions (e.g., in RAM) that specify how to parse the subsequent header (e.g., how to parse the various different fields of an IPv4 packet, an ARP packet, etc.).
- the configuration data for the parser also specifies to send the participating header fields to particular registers of the PPHV and the non-participating header fields to particular registers of the SPHV.
- the configuration data also specifies when to stop parsing the packet headers and treat the remainder of the packet as payload. This decision may be made based on the amount of data already stored in the PHVs for the current traversal through the nodes of the parse graph, or because the end of relevant packet headers for the switch have been reached (e.g., the end of the innermost layer 4 headers).
- Some of the above-described features and applications are implemented as software processes that are specified as a set of instructions recorded on a computer readable storage medium (also referred to as computer readable medium).
- a computer readable storage medium also referred to as computer readable medium.
- processing unit(s) e.g., one or more processors, cores of processors, or other processing units
- Examples of computer readable media include, but are not limited to, CD-ROMs, flash drives, RAM chips, hard drives, EPROMs, etc.
- the computer readable media does not include carrier waves and electronic signals passing wirelessly or over wired connections.
- the term “software” is meant to include firmware residing in read-only memory or applications stored in magnetic storage, which can be read into memory for processing by a processor.
- multiple software inventions can be implemented as sub-parts of a larger program while remaining distinct software inventions.
- multiple software inventions can also be implemented as separate programs.
- any combination of separate programs that together implement a software invention described here is within the scope of the invention.
- the software programs when installed to operate on one or more electronic systems, define one or more specific machine implementations that execute and perform the operations of the software programs.
- FIG. 9 conceptually illustrates an electronic system 900 with which some embodiments of the invention are implemented.
- the electronic system 900 can be used to execute any of the control, virtualization, or operating system applications described above.
- the electronic system 900 may be a computer (e.g., a desktop computer, personal computer, tablet computer, server computer, mainframe, a blade computer etc.), phone, PDA, or any other sort of electronic device.
- Such an electronic system includes various types of computer readable media and interfaces for various other types of computer readable media.
- Electronic system 900 includes a bus 905 , processing unit(s) 910 , a system memory 925 , a read-only memory 930 , a permanent storage device 935 , input devices 940 , and output devices 945 .
- the bus 905 collectively represents all system, peripheral, and chipset buses that communicatively connect the numerous internal devices of the electronic system 900 .
- the bus 905 communicatively connects the processing unit(s) 910 with the read-only memory 930 , the system memory 925 , and the permanent storage device 935 .
- the processing unit(s) 910 retrieves instructions to execute and data to process in order to execute the processes of the invention.
- the processing unit(s) may be a single processor or a multi-core processor in different embodiments.
- the read-only-memory (ROM) 930 stores static data and instructions that are needed by the processing unit(s) 910 and other modules of the electronic system.
- the permanent storage device 935 is a read-and-write memory device. This device is a non-volatile memory unit that stores instructions and data even when the electronic system 900 is off. Some embodiments of the invention use a mass-storage device (such as a magnetic or optical disk and its corresponding disk drive) as the permanent storage device 935 .
- the system memory 925 is a read-and-write memory device. However, unlike storage device 935 , the system memory is a volatile read-and-write memory, such a random access memory.
- the system memory stores some of the instructions and data that the processor needs at runtime.
- the invention's processes are stored in the system memory 925 , the permanent storage device 935 , and/or the read-only memory 930 . From these various memory units, the processing unit(s) 910 retrieves instructions to execute and data to process in order to execute the processes of some embodiments.
- the bus 905 also connects to the input and output devices 940 and 945 .
- the input devices enable the user to communicate information and select commands to the electronic system.
- the input devices 940 include alphanumeric keyboards and pointing devices (also called “cursor control devices”).
- the output devices 945 display images generated by the electronic system.
- the output devices include printers and display devices, such as cathode ray tubes (CRT) or liquid crystal displays (LCD). Some embodiments include devices such as a touchscreen that function as both input and output devices.
- bus 905 also couples electronic system 900 to a network 965 through a network adapter (not shown).
- the computer can be a part of a network of computers (such as a local area network (“LAN”), a wide area network (“WAN”), or an Intranet, or a network of networks, such as the Internet. Any or all components of electronic system 900 may be used in conjunction with the invention.
- Some embodiments include electronic components, such as microprocessors, storage and memory that store computer program instructions in a machine-readable or computer-readable medium (alternatively referred to as computer-readable storage media, machine-readable media, or machine-readable storage media).
- computer-readable media include RAM, ROM, read-only compact discs (CD-ROM), recordable compact discs (CD-R), rewritable compact discs (CD-RW), read-only digital versatile discs (e.g., DVD-ROM, dual-layer DVD-ROM), a variety of recordable/rewritable DVDs (e.g., DVD-RAM, DVD-RW, DVD+RW, etc.), flash memory (e.g., SD cards, mini-SD cards, micro-SD cards, etc.), magnetic and/or solid state hard drives, read-only and recordable Blu-Ray® discs, ultra density optical discs, any other optical or magnetic media, and floppy disks.
- CD-ROM compact discs
- CD-R recordable compact discs
- the computer-readable media may store a computer program that is executable by at least one processing unit and includes sets of instructions for performing various operations.
- Examples of computer programs or computer code include machine code, such as is produced by a compiler, and files including higher-level code that are executed by a computer, an electronic component, or a microprocessor using an interpreter.
- ASICs application specific integrated circuits
- FPGAs field programmable gate arrays
- the terms “computer”, “server”, “processor”, and “memory” all refer to electronic or other technological devices. These terms exclude people or groups of people.
- display or displaying means displaying on an electronic device.
- the terms “computer readable medium,” “computer readable media,” and “machine readable medium” are entirely restricted to tangible, physical objects that store information in a form that is readable by a computer. These terms exclude any wireless signals, wired download signals, and any other ephemeral signals.
- FIGS. 6 and 8 conceptually illustrate processes. The specific operations of these processes may not be performed in the exact order shown and described. The specific operations may not be performed in one continuous series of operations, and different specific operations may be performed in different embodiments. Furthermore, the process could be implemented using several sub-processes, or as part of a larger macro process. Thus, one of ordinary skill in the art would understand that the invention is not to be limited by the foregoing illustrative details, but rather is to be defined by the appended claims.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computer Security & Cryptography (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
Description
- In Software Defined Networking (SDN), the control plane is physically separated from the forwarding plane and communicates with the forwarding plane through an interface (e.g., the OpenFlow protocol). OpenFlow is an evolving networking standard that uses a match-action paradigm for network packet switching. The typical hardware switches that were (and still are) used to implement the match-action paradigm, however, are not quite flexible since they process only a fixed set of fields with a limited repertoire of packet processing actions. A reconfigurable match table (RMT) architecture that overcomes the aforementioned shortcomings has recently been introduced to the market. This new architecture could be improved to make the hardware switch operate more efficiently.
- Some embodiments provide a novel packet processing pipeline that enhances the match-action packet header processing performed by reconfigurable hardware switches. The hardware switch of some embodiments includes, among other elements, an ingress pipeline and an egress pipeline. Each pipeline includes a parser, a match-action unit (MAU), and a deparser. In some embodiments, for each packet that is processed by the switch, the parser separates the packet header data (e.g., Ethernet headers, IPv4 headers, TCP headers, etc.) from the payload of the packet and determines which fields of each packet header (e.g., source port address of the TCP header, destination port address of the TCP header, etc.) might be processed by the MAU. The parser of some such embodiments (1) populates a primary packet header vector (PPHV) with only the fields of the different headers that may be used by the MAU, and (2) populates a secondary packet header vector (SPHV) with the rest of the fields of the different headers (i.e., the fields that will not be processed by the MAU). In some embodiments, the PPHV and SPHV each includes a set of registers that stores the packet header fields during processing by the MAU.
- In some embodiments, the parser delivers the populated PPHV to the MAU, which includes a set of match-action stages for matching the different header fields of the packet against the different match tables and applying the required actions to the packet by modifying the packet header data (e.g., assigning the packet to an output port and queue, sending the packet to a network controller, dropping the packet, etc.). The parser of some such embodiments also delivers the populated SPHV directly to the deparser (without having the MAU process the SPHV) along with the payload of the packet.
- In some embodiments, after the different packet headers are processed by the MAU, the deparser reassembles the packet with (1) the processed header data (e.g., modified header fields) of the PPHV that the deparser receives from the MAU, (2) the unprocessed header data stored in the SPHV that the deparser receives from the parser directly, and (3) the payload of the packet that the deparser also receives from the parser. After reassembling the packet, if the packet is reassembled by an ingress deparser, the deparser sends the packet to a queuing system to be subsequently forwarded to the egress pipeline of the switch. On the other hand, if the packet is reassembled by an egress deparser, the deparser sends the packet (with a header that might or might not have been modified) out of the switch.
- The parser of some embodiments determines which fields of each packet header may be processed and which fields will not be processed by the MAU, based on the information the parser receives from the packet headers themselves (e.g., the EtherType field of the packet, etc.), and based on configuration data received from the control plane. In some embodiments, a compiler in the control plane receives the data required for configuring the pipeline (e.g., through a programing language code), generates a set of configuration data, and distributes the generated data to a configurator module (also part of the control plane). The configurator module then distributes the configuration data to both the parser and MAU of the pipeline in the switch.
- The parser of some embodiments, based on the information it receives from the configurator, determines which sets of packet headers to extract from the packet and how to extract these packet headers. The parser analyzes the packet headers as it extracts them according to the configuration data configured by the control plane. This configuration data represents a parse graph that identifies how to parse the packet based on values of certain packet header fields as well as table flow and control flow graphs used to generate the configuration data for the MAU, which describe how packets will be processed through the stages of the MAU. Based on the configuration data and the values of certain packet header fields, the parser determines iteratively (i) whether each packet header field should be placed in the PPHV or the SPHV and (ii) how to extract the next set of packet header fields. For instance, the EtherType field of the layer 2 header of some embodiments specifies how to extract the layer 3 header (e.g., whether the packet is IPv4 or IPv6, etc.).
- Separating the fields that may be processed by the MAU from the rest of the fields in each packet header improves the overall efficiency of the switch in many ways. For instance, this separation helps avoid wasting the valuable resources of the MAU by enabling these resources to only (or primarily) process data that is actually needed for making the packet processing determinations. For example, a TCP packet header typically has a minimum length of five words (twenty bytes). In many cases, the MAU of the pipeline processes only four bytes of the TCP header bytes (i.e., the source and destination port addresses) and the rest of the fields in the TCP header (sixteen bytes) are not used by any of the stages of the MAU. In such cases, by separating the unused data from useful data, the parser does not waste any of the valuable resources of the MAU to process the other sixteen bytes of the TCP header.
- In addition, separating the PPHV from the SPHV enables the parser to parse deeper into packets while sending the same amount of data to the MAU (via the PPHV). While using a single PPHV of a fixed size only allows the parser to extract the packet headers up to a certain point (the size of the PPHV), sending some of the unused fields into the SPHV allows the PPHV to carry only the more useful fields, and include deeper packet headers. For example, newer tunneling protocols such as GENEVE add variable length option fields to the packet header, and may be quite long. Similarly, packets sent using multi-layer tunneling may have significant amounts of packet header data that may need to be extracted in order to reach the layer 4 (e.g., TCP) data. By sending the fields that will not be processed by the MAU to the SPHV, the PPHV can include packet header fields that are located deeper into the packet, without sending additional data to the MAU.
- The preceding Summary is intended to serve as a brief introduction to some embodiments of the invention. It is not meant to be an introduction or overview of all inventive subject matter disclosed in this document. The Detailed Description that follows and the Drawings that are referred to in the Detailed Description will further describe the embodiments described in the Summary as well as other embodiments. Accordingly, to understand all the embodiments described by this document, a full review of the Summary, Detailed Description and the Drawings is needed. Moreover, the claimed subject matters are not to be limited by the illustrative details in the Summary, Detailed Description and the Drawings, but rather are to be defined by the appended claims, because the claimed subject matters can be embodied in other specific forms without departing from the spirit of the subject matters.
- The novel features of the invention are set forth in the appended claims. However, for purpose of explanation, several embodiments of the invention are set forth in the following figures.
-
FIG. 1 illustrates a block diagram of a hardware switch and a block diagram of an ingress/egress interface of the hardware switch. -
FIG. 2 illustrates a block diagram that shows that the parser produces a secondary packet header vector (SPHV), besides the primary packet header vector (PPHV), that bypasses the MAU and is directly delivered to the deparser. -
FIG. 3 illustrates an ingress/egress pipeline architecture that separates the participating header fields of the different packet headers from nonparticipating packet header fields of a packet. -
FIG. 4 conceptually illustrates the reason that the parser of some embodiments does not store the entire packet headers in the PPHV to be processed by the MAU. -
FIG. 5 conceptually illustrates a block diagram of a parser that is configured by a compiler to separate the participating (in an MAU process) packet header fields from nonparticipating header fields. -
FIG. 6 conceptually illustrates a process of some embodiments that populates the PPHV and SPHV with different header fields of a packet header. -
FIG. 7 conceptually illustrates a compiler of some embodiments for generating parser and MAU configuration data. -
FIG. 8 conceptually illustrates a process performed by the compiler of some embodiments to generate parser configuration data for a hardware switch that parses packet header data into a PPHV and a SPHV. -
FIG. 9 conceptually illustrates an electronic system with which some embodiments of the invention are implemented. - In the following description, numerous details are set forth for the purpose of explanation. However, one of ordinary skill in the art will realize that the invention may be practiced without the use of these specific details. In other instances, well-known structures and devices are shown in block diagram form in order not to obscure the description of the invention with unnecessary detail.
- Some embodiments provide a novel packet processing pipeline that enhances the match-action packet header processing performed by reconfigurable hardware switches. The hardware switch of some embodiments includes, among other elements, an ingress pipeline and an egress pipeline. Each of these pipelines includes a parser, a match-action unit (MAU), and a deparser.
-
FIG. 1 illustrates a block diagram of ahardware switch 110 and a block diagram of an ingress/egress pipeline of the hardware switch. More specifically, thehardware switch 110 includes an ingress pipeline (datapath) 115, a queuing system (e.g., a data buffer) 120, and anegress pipeline 125. Theingress pipeline 115 of some embodiments receives a packet 105 (e.g., through an I/O module), parses the packet into a packet header vector (PHV), sends the PHV through a set of match and action stages which may modify the PHV, deparses the packet headers back from the PHV into packet format, and enqueues the packet in a centralized data buffer (i.e., the data buffer 120). Each one of the aforementioned operations is described in more detail below by reference to the ingress/egress pipeline 140. - In some embodiments the
queuing system 120 receives the packets that are processed by the ingress pipeline and provides a large shared buffer (storage) that accommodates the queuing delays due to oversubscription of the output channels of the ingress deparser. In some embodiments, the data buffer stores packet data, while pointers to that data are kept in different queues per channel. Each channel in turn requests data from the common data buffer using a configurable queuing policy. When pointers to packets reach the head of the queues, the packets are read out of thedata buffer 120 into theegress interface 125. - Similar to the
ingress interface 115, theegress interface 125 of some embodiments receives the packet from thequeuing system 120, separates the packet payload from the packet headers, stores the packets headers in a PHV, sends the PHV through an ingress pipeline of match and action stages, deparses the packet headers back from the PHV into packet format, and sends thepacket 130 to an appropriate output port of theswitch 110 to be driven off the switch (e.g., through one of the output channels). Theoutput packet 130 may be the same packet as the input packet 105 (i.e., with identical packet headers), or it may have different packet headers compared to theinput packet 105 based on the actions that are applied to the packet headers in the ingress and egress pipelines (e.g., different header field values for certain header fields and/or different sets of header fields). - One of ordinary skill in the art would realize that the illustrated blocks in
switch 110 are exemplary only. The ingress, queuing, and egress blocks are simplified for simplicity of the description. For example, although the figure shows only one entry point to the ingress parser and one exit point from the egress deparser, in some embodiments the input signals are received by many different input channels (e.g., 64 channels) and the output signals are sent out of the switch from different output channels (e.g., 64 channels). Additionally, although for the illustrated switch only one parser interface is shown for the ingress/egress pipeline 140, some embodiments employ numerous parser blocks (e.g., 16 parser blocks) that feed the match-action unit (MAU) in each pipeline. -
FIG. 1 also shows a block diagram 140 of an ingress/egress interface of thehardware switch 110. The interface includes a pipeline with three different units, namely aparser unit 145, a match-action unit (MAU) 150, and adeparser unit 155. Theparser 145 of some embodiments receives the incoming packet data and produces a PHV as its output. In other words, theparser 145 separates the packet headers from the packet payload by extracting the different packet headers and storing them in the PHV. In some embodiments the PHV includes a set of different size registers or containers. For instance, in some embodiments the PHV includes sixty-four 8-bit registers, ninety-six 16-bit registers, and sixty-four 32-bit registers (for a total of 224 registers containing 4096 bits), though other embodiments may have any different numbers of registers of different sizes. In some embodiments, theparser 145 stores each extracted packet header in a particular subset of one ore more registers of the PHV. For example, the parser might store a first header field in one 16-bit register and a second header field in a combination of an 8-bit register and a 32-bit register (e.g., if the header field is 36 bits long). - The PHV provides the input data to the match tables of the MAU. In some embodiments the
MAU 150 includes a set of match-action stages (e.g., 32 match-action stages in some embodiments), each of which matches a particular set of header fields against a match table and takes an action based on the result of the match (e.g., assigning the packet to an output port and queue, dropping the packet, modifying one or more of the header fields, etc.). TheMAU 150 and its different match-action stages are described in more detail below by reference toFIG. 3 . Based on the actions taken on different header data during the different stages of theMAU 150, the PHV that the MAU outputs might include the same header data as the PHV that the MAU received from the parser, or the output PHV might contain different data than the input PPHV. - The output PHV is then handed to the
deparser 155. In some embodiments, thedeparser 155 reassembles the packet by putting back together the output PHV (that might or might not have been modified) that the deparser receives from theMAU 150 and the payload of the packet that the deparser receives directly from theparser 145 in some embodiments. The deparser then sends the packet out of the ingress/egress pipeline (to thequeuing system 120 or out of the switch, depending on whether it is the deparser for the ingress pipeline or the egress pipeline). -
FIG. 2 illustrates a block diagram similar to the block diagram 140 ofFIG. 1 , except that this figure shows that the parser also produces a secondary packet header vector (SPHV), or tag-along PHV, that bypasses the MAU's pipeline and is directly delivered to the deparser. More specifically, after receiving a packet, theparser 145 of some embodiments produces two different packet header vectors, a PPHV and a SPHV, as its output. - As described above, in some embodiments, for each packet that is processed by the switch, the parser separates the packet header data (e.g., Ethernet, IPv4, TCP, etc.) from the
payload 220 of the packet and determines which fields of each packet header (e.g., source port address of the TCP header, destination port address of the TCP header, etc.) may be processed by the MAU (i.e., might be matched against, or might be modified by an action). Packets may contain packet headers and a payload, though the divide between the two depends on how deep into the packet an element is inspecting. For instance, for a network element that only looks at and operates on the layer 2 (e.g., Ethernet) headers, the layer 3 (e.g., IPv4, IPv6, etc.), layer 4 (e.g., TCP, UDP, etc.) portions are part of the packet payload. However, for a network element that uses the layer 4 headers, the layer 2, layer 3, and layer 4 portions are all part of the headers, and the remainder of the packet is treated as payload. Each header (e.g., the layer 2 header, layer 3 header, layer 4 header, tunnel encapsulations, etc.) may include one or more header fields, which contain data values. For example, the source and destination IP addresses of an IP header (e.g., IPv4 header) are two different header fields of the IP header, while the source port and destination port addresses of a TCP header are two different header fields of the TCP header. - The term “packet” is used throughout this application to refer to a collection of bits in a particular format sent across a network. It should be understood that the term “packet” may be used herein to refer to various formatted collections of bits that may be sent across a network. A few examples of such formatted collections of bits are Ethernet frames, TCP segments, UDP datagrams, IP packets, etc.
- While separating the different packet headers from the payload and separating the participating header fields from the nonparticipating header fields, the parser of some embodiments (1) populates the primary packet header vector (PPHV) 230 with only the participating header fields of different headers (i.e., the fields that may be processed by the MAU), and (2) populates the secondary packet header vector (SPHV) 240 with the nonparticipating fields of the different packet headers (i.e., the fields that will not be processed by the MAU).
- Each of the PPHV and SPHV includes a set of registers that stores the packet header fields in some embodiments. For example, and as described above, the
PPHV 230 of some embodiments includes a combination of different size registers (e.g., with a total capacity of 4 k bits) that hold the packet header field values that may be processed the MAU. Similarly, theSPHV 240 of some embodiments includes a set of the same or different size registers (e.g., shift registers, FIFO registers, etc.) for holding non-used header field values (e.g., with a total capacity of 2 k bits). - In some embodiments, the parser delivers the
populated PPHV 230 to theMAU 150, which includes a set of match-action stages for matching the different header fields of the packet against the different match tables and applying the required actions to the packet by modifying the packet header data (e.g., assigning the packet to an output port and queue, sending the packet to a network controller, dropping the packet, modifying the header field values of the PPHV, etc.). The parser of some such embodiments also delivers thepopulated SPHV 240 directly to thedeparser 155 along with the payload of the packet and without having theMAU 150 process theSPHV 240. By carrying the SPHV directly to the deparser without having the registers sent to the MAU, the wiring of the switch with respect to these packet header registers is greatly simplified. - In some embodiments, after the packet headers are processed by the
MAU 150, thedeparser 155 reassembles the packet with (1) the processed header data (that might or might not have been modified) which the deparser receives from theMAU 150, (2) the unprocessed header data stored in theSPHV 240 that the deparser receives from theparser 145, and (3) thepayload 220 of the packet that the deparser also receives from theparser 145. As described before, after reassembling the packet, if the packet is reassembled by an ingress deparser, the deparser sends the packet to a queuing system to be subsequently forwarded to the egress pipeline of the switch. On the other hand, if the packet is reassembled by an egress deparser, the deparser sends the packet (with a header that might or might not have been modified) out of the switch. In some embodiments, the deparser recognizes the location in the reconstructed packet at which to place each value from the different registers (as an example, the 13 bits of the IPv4 fragment offset field that might be located in register X of the SPHV are inserted directly after the 3 bits of the flags field from register Y of the SPHV and directly before the 8 bits of the time to live field from register N of the PPHV). -
FIG. 3 illustrates an ingress/egress pipeline architecture 300 that separates the participating header fields (i.e., the header fields that may participate in the different match-action stages of the MAU) of the different packet headers from non-participating packet header fields of a packet. More specifically, the figure shows that thepipeline 300 includes aparser 310, an MAU with multiple match-action stages 1 through m (where m is an integer such as 32 in some embodiment), and adeparser 360. As illustrated, each match-action stage includes aXbar 315, a set of exact match tables 320, a set of ternary match tables 325, and a Very Long Instruction Word (VLIW)action engine 330. - As described above, the
parser 310 receives a packet (e.g., through a set of input modules of the switch) and extracts the packet headers (e.g., Ethernet header, IPv4 header, IPv6 header, TCP header, UPD header, tunnel encapsulation headers, etc.) from the body of the packet. Theparser 310 then separates the header field values that may be processed by the MAU from the rest of the header field values in each packet header. As stated above, the header fields of a packet header are the different segments of the packet header, each of which stores a particular value representing a particular piece of data. For example, an IPv4 packet header includes a source IP address header field, a destination IP address header field, an IPv4 identification field header field, a time-to-live header field, etc. - The parser of some embodiments determines which header fields are participating header fields (i.e., the fields that may be processed by the MAU) and which header fields are non-participating header fields based on configuration data that the parser receives from a configurator module in the control plane. In some such embodiments, the configurator module receives the configuration data from a compiler (also part of the control plane) which generates the configuration data based on a set of instructions that the compiler receives (e.g., from a user designing the functionality of the hardware switch). The compiler and configurator modules of some embodiments are described in more detail below by reference to
FIG. 5 . - While the
parser 310 separates the header fields that may be processed by the MAU from the rest of the header fields of the different packet headers, the parser stores the participating header fields in the PPHV (in different registers of the PPHV) and delivers the PPHV to the MAU for further processing of the header data. As described above, the MAU of some embodiments includes a set of different stages, each of which is configured to process a particular set of header fields of the PPHV. In some embodiments, each of these stages is reconfigurable by a controller that is responsible for configuring (and reconfiguring) the stages of the match-action units of the hardware switch to match against various different packet header fields and - In some embodiments, the output PPHV from the
parser 310 is delivered to both theXbar 315 and theVLIW action engine 330. TheXbar 315 of each stage, in some embodiments, includes a set of one or more crossbars (e.g., two crossbars) that identifies which field or fields of the PPHV are used by the match tables of the corresponding stage of the MAU, and delivers the identified fields to the exact match tables and ternary match tables of the stage. For instance if the match tables of a particular stage are configured to match against a source IP address field that is carried by one or more PPHV registers, theXbar 315 of that particular stage takes the IP source field from the PPHV register(s) and delivers the field to the match tables. In some embodiments, theXbar 315 operates at a bit level. That is, a particular header field that contains 20 bits might be carried by a 16-bit register and a portion of an 8-bit register of the PPHV. The Xbar of some such embodiments, therefore, passes only 20 bits of the data that is carried by the combination of the 16-bit register and the 8-bit register to its corresponding match tables. - As illustrated in
FIG. 3 , each match-action stage of the MAU includes a set of ternary match tables 325 and a set of exact match table 320. Some embodiments implement (1) the ternary match tables with ternary content addressable memory (TCAM) units, and (2) the exact match tables with static random-access memory (SRAM) units, that operate as hash tables. Each one of the ternary match tables and exact match tables of some embodiments includes a number of entries (e.g., flow entries) each of which matches against the values of one or more packet header fields and instructs theaction engine 330 of the same stage to take a particular action (e.g., an action stored in a SRAM unit of an action table). - In some embodiments the ternary match table stores packet header data that can include ternary bits (i.e., bits with values of 0, 1 or wildcard). Wildcarding the bits allow a single table entry to match a wide variety of packet header fields. The exact match table of some embodiments does not allow any wildcarding, hence the packets must exactly match the table entries. When the header field is matched against an entry of a match table (a ternary table or an exact match table), several pointers which together contain the required information to perform the desired actions are retrieved from that entry of the match table. The information, in some embodiments, includes, among other data, an instruction memory address for the
action engine 320 and an action memory address and size. The desired actions include simple actions such as setting a field in the PPHV to a particular value or dropping a packet, as well as complex operations, such as adding encapsulations to the packet (e.g., Provider Backbone Bridges (PBB) encapsulation, GRE or GENEVE tunneling, etc.). - In some embodiments, the
VLIW action engine 330 includes different action operations that may receive their sources from packet header fields (which is why a copy of the PPHV registers are directly delivered to the action unit), or from an action memory (not shown). An action indicated by a match may be simple and require only a small amount of data from the action memory or complex and require a large amount of data. When theaction engine 330 receives the action data (from the action memory or the PPHV), the action engine takes the required action which may include modifying a header field in the PPHV (e.g., to decrement the TTL field of the IP header, change the address of the packet, etc., or taking another action such as assigning the packet to an output port and/or queue, sending the packet to the controller, dropping the packet, etc. After the required action is taken (e.g., when a match is found) in each stage of the MAU, the PPHV (as modified by the action, if such modifications are performed) is delivered to the next stage of the MAU for further processing. In some embodiments, each successive PPHV that includes the packet header data of one packet progresses through the pipeline of match-action stages one clock cycle behind the previous PPHV. - The last match-action stage of the MAU, after processing the PPHV, delivers the processed PPHV to the
deparser 360. As described above, after receiving the PPHV of a particular packet from the MAU, thedeparser 360 of some embodiments reassembles the packet by putting together the processed header fields that are received from the MAU, with the unprocessed header fields that are received directly from theparser 310 in theSPHV 355 and thepayload 350 of the packet that is also received directly from theparser 310. The packet is then either sent out of the switch (if the deparser is an egress deparser), or sent to the queuing system (if the deparser is an ingress deparser) for further processing by the egress pipeline of the switch. -
FIG. 4 conceptually illustrates the reason that the parser of some embodiments does not store the entire packet headers in the PPHV to be processed by the MAU. This figure shows that apacket 410 is being received by theparser 420 to parse the packet headers into thePPHV 430. However, not every header field of each packet header is needed by the MAU stages of the upcoming ingress or egress pipeline to which the parser sends the PPHV. For instance, some of the packet header fields will (i) not be matched against by any of the match entries of the match tables in the pipeline and (ii) not be modified by any possible action entry that could be performed in the pipeline. Thus, as theparser 420 extracts each packet header from thepacket 410, it determines which of the header fields of the packet header might be processed by at least one of the match-action stages of the MAU. - The illustrated example shows that a packet header 415 (e.g., an Ethernet header, IPv4 header, etc.) of the
packet 410 includes several participatingheader fields 450 that the MAU is configured (e.g., by a configurator module of the control plane) to potentially process. At the same time, thepacket header 415 also includes several other non-participating header fields 460 that the MAU is not configured to process. In some embodiments, when theparser 420 extracts a particular packet header from thepacket 410, the parser must extract the entire contiguous packet header at once (i.e., the parser cannot leave certain fields of a packet header in the payload while placing the other fields of the packet header in the PHV). Because the different participating header fields of the packet header are often not placed next to each other in the packet header (as illustrated in the figure), the parser of some embodiments separates these participating header fields from nonparticipating fields during extraction of the packet header. - For example, the MAU might be configured to process only a particular set of header fields in a UDP packet header, which may not be the first two header fields of the packet header (i.e., the source and destination ports). In such a case, the parser locates the particular header fields in the set, pulls these fields out of the packet header, and stores the header fields in the PPHV. However, the other nonparticipating header fields that are also extracted from the packet have to be dealt with as well. Therefore, in some embodiments, the
parser 420 looks at each header field in the packet header and determines whether the identified header field might be processed by the MAU or will definitely not be processed by the MAU. Though not the case in this example, in some embodiments none of the fields of a packet header might be placed in the PPHV, but the parser would extract these fields and place them in the SPHV in order to continue deeper into the packet and place the fields of a higher-layer packet into the PPHV. - If the
parser 420 determines that the header field is one of the participating header fields, the parser stores the header field in the PPHV 430 (i.e., in a particular set of registers of thePPHV 430 designated for that header field). On the other hand, if theparser 420 determines that the identified header field is not supposed to be processed by the MAU, the parser stores the header field in theSPHV 440 to be subsequently sent directly to the deparser (not shown) of the pipeline without getting processed. - The parser of some embodiments determines which fields of each packet header may be processed and which fields will not be processed by the MAU, based on the information the parser receives from the packet itself (e.g., by one or more particular packet header of the packet), and based on the configuration data received from a compiler in the control plane. In some embodiments, the compiler receives the data required for configuring the pipeline (e.g., through a programing language code), generates a set of configuration data, and distributes the generated data to a configurator module (also in the control plane). The configurator module then distributes the configuration data to both parser and MAU of the pipeline in the switch (e.g., at run-time or during setup time).
-
FIG. 5 conceptually illustrates a block diagram 500 of a parser that is configured by a compiler to separate the participating packet header fields from non-participating header fields. Specifically, this figure shows aparser 510, acompiler 520, aconfigurator 530, and anMAU 540. Theparser 510 of some embodiments includes anextractor state machine 560. In some embodiments, thecompiler 520 andconfigurator 530 are separate from the hardware switch (i.e., not part of the chip), whereas theparser 510 andMAU 540 are part of the hardware switch circuitry. Theconfigurator 530, in some embodiments, represents a runtime network controller stack that configures the hardware switch, while thecompiler 520 operates at an earlier stage to generate the configuration data and provide this configuration data to theconfigurator 530. - In some embodiments, the
compiler 520 generates configuration data for the hardware switch based on a switch configuration received by the compiler 520 (e.g., through code written by a user, such as P4 code for designing switch operations). The configuration data specifies how to program the match-action stages of theMAU 540 to implement the switch configuration specified by the received configuration code, as well as how to program theextractor state machine 560 of the parser to extract the packet header fields correctly and to place the packet header fields into the appropriate PPHV or SPHV registers. - The
compiler 520 passes the configuration data to theconfigurator 530, which provides the configuration data to theMAU 540 so that the MAU can implement the appropriate switch configuration. In addition, theconfigurator 530 provides the parser configuration data generated by thecompiler 520 to theextractor state machine 560 so that the extractor state machine can extract header fields from the packets into the registers of the PPHV and the SPHV. - The
extractor state machine 560 receives thepacket 585 and (i) separates the packet headers into thePPHV 570 and theSPHV 575 and (ii) separates out thepayload 580 of the packet. In some embodiments, the extractor state machine begins at the front of the packet (i.e., the outermost headers) and works inward, extracting each packet header field into either the PPHV or the SPHV. For certain header fields, the extractor state machine looks at the value and compares the value to configuration entries (e.g., TCAM entries), which specify how the next set of packet header fields should be extracted. - The
extractor state machine 560 follows a parse graph that specifies numerous states at which a particular packet header field is extracted into either the PPHV or the SPHV, as well as what state of the parse graph to proceed to next. That is, in some embodiments each state specifies which bits of the header to extract for a particular header field, which register(s) of the PPHV or SPHV to put those bits into, and to which state to proceed next. Theextractor state machine 560 of some embodiments may also read the data of certain header fields in order to determine which path to take through the set of states of the parse graph. For example, the EtherType field of an Ethernet header indicates the type of protocol encapsulated by the Ethernet header (e.g., IPv4, IPv6, ARP, etc.). The different fields of these protocols will be formatted differently, so for theextractor state machine 560 to properly identify the next state(s) in the parse graph, it needs to know the EtherType value. Similarly, the protocol field of an IPv4 header specifies the transport layer protocol (e.g., TCP, UDP, etc.), which will determine how those fields are extracted. - In some embodiments, the parse graph is implemented in the
parser 530 as a set of entries in TCAMs that the extractor state machine matches. The packet header is received as a set of bits stored in, e.g., a buffer, and at any time during the extraction process a pointer points to a particular location in the buffer. The TCAM entries of the parse graph of some embodiments specify which type of header is next based on the value of certain fields of a current header (e.g., the EtherType field of an Ethernet header, the protocol field of an IPv4 header, etc.), and how to extract the header fields of the next header. This entails specifying how many bits to extract for each header field in the next header, to where in the sets of registers of the PPHV and the SPHV each header field should be extracted, and to where in the buffer to move the pointer for the next header field. As mentioned, for certain fields, the header field value is matched against the TCAM entries to determine the next state or set of states based on the determination as to what type of packet headers are contained in the packet. - The parser configuration specifies which packet headers will be extracted into the PPHV and which packet headers are extracted into the SPHV. In some embodiments, any field that is either (i) used as a match field of any match entry in the MAU and/or (ii) modified by any action entry in the MAU is extracted into the PPHV, with only the header fields that are not involved in any way with the MAU extracted into the SPHV. Thus, for some packets, some of the header fields in the PPHV will not actually be used or modified by the MAU (especially if one of the earlier stages instructs the switch to drop the packet). Identifying all of the fields that will be used on a packet-by-packet basis would effectively entail performing the match-action pipeline (because the values of the header fields would have to be considered for each stage to determine which actions to take, and the actions performed to find the new values for the header fields in order to identify which fields are considered at subsequent stages). Thus, the
extractor state machine 560 of some embodiments only reads and matches against the values for certain header fields that identify the structure of the subsequent packet headers, so that the subsequent packet headers can be extracted properly according to the parse graph. - Separating the fields that may be processed by the MAU from the rest of the fields in each packet header improves the overall efficiency of the switch in many ways. For instance, this separation helps avoid wasting the valuable resources of the MAU by enabling these resources to only (or primarily) process data that is actually needed for making the packet processing determinations. For example, a TCP packet header typically has a minimum length of five words (twenty bytes). In many cases, the MAU of the pipeline processes only four bytes of the TCP header bytes (i.e., the source and destination port addresses) and the rest of the fields in the TCP header (sixteen bytes) are not used by any of the stages of the MAU. In such cases, by separating the unused data from useful data, the parser does not waste any of the valuable resources of the MAU to process the other sixteen bytes of the TCP header.
- In addition, separating the PPHV form the SPHV enables the parser to parse deeper into packets while sending the same amount of data to the MAU (via the PPHV). While using a single PPHV of a fixed size only allows the parser to extract the packet headers up to a certain point (the size of the PPHV), sending some of the unused fields into the SPHV allows the PPHV to carry only the more useful fields, and include deeper packet headers. For example, newer tunneling protocols such as GENEVE add variable length option fields to the packet header, and may be quite long. Similarly, packets sent using multi-layer tunneling may have significant amounts of packet header data that may need to be extracted in order to reach the layer 4 (e.g., TCP) data. By sending the fields that will not be processed by the MAU to the SPHV, the PPHV can include packet header fields that are located deeper into the packet, without sending additional data to the MAU.
-
FIG. 6 conceptually illustrates aprocess 600 of some embodiments that populates the PPHV and SPHV with different header field values (e.g., source and destination port addresses, fragmentation offset field, transport layer ports, encapsulation addresses, etc.) of a packet header (e.g., IPv4, TCP, UDP, etc.). Theprocess 600 of some embodiments is performed by an extractor state machine of a ingress parser or egress parser of a hardware switch in some embodiments, such as theextractor state machine 560 illustrated inFIG. 5 . - The
process 600 initiates by receiving (at 610) a packet header. In some embodiments the process receives the packet header from a packet receiver module of the parser, such as the packet receiver 550 ofFIG. 5 . A packet header could be an Ethernet header, an IP header, a TCP header, etc. In addition, in some embodiments, the process has learned (e.g., from examining a previous packet header field), the format of the received packet header (i.e., what type of packet header it is, and therefore the lengths of the different fields of the packet header. - The process then extracts (at 620) the next header field of the received packet header. As described above, in some embodiments the extractor state machine points to a particular location in a buffer that holds the packet header, and the parser configuration (implementing the parse graph) identifies the length of the next packet header, starting from that particular location. Thus, to extract the next header field, the process extracts the specified next number of bits (and, in some embodiments, moves the pointer to the new location after the last extracted bit).
- The process then determines (at 630) whether the extracted field may be processed by any of the match-action stages of the MAU. When the process determines that the extracted field is a participating header field that could be used at some point during the processing of the PPHV through the different stages of the MAU, the process identifies (at 640) a container (e.g., one or more registers of the PPHV) in the PPHV and stores the extracted header field in the identified container of the PPHV. On the other hand, if the process determines that the extracted field is a nonparticipating header field, which will not processed by any match-action stage of the MAU, the process identifies (at 650) a container (e.g., one or more registers of the SPHV) in the SPHV and stores the extracted header field in the identified container of the SPHV.
- As described above, in some embodiments operations 620-640 or 620-650 are all performed as a single operation. Specifically, the determination as to whether a particular field may be processed by the MAU is made during the generation of the parse graph, (according to the configuration data) and embedded in the parse graph based on whether the parse graph instructions specify to store the particular header field in the PPHV or the SPHV. Thus, the extractor state machine of some embodiments extracts the header field (i.e., the next X bits of the packet header) into a specific register of either the PPHV or the SPHV, without performing any specific analysis at packet processing time as to whether the header field is a participating or non-participating header field.
- After storing the extracted header field in one of the PPHV or the SPHV, the process of some embodiments determines (at 660) whether any additional header fields remain to be extracted. In some embodiments, the parse graph proceeds to the next state according to the instructions of the previous state, and each subsequent state identifies the next set of bits (i.e., the next header field) to extract. Upon reaching the last field to extract, the parse graph state specifies that the extraction process is done, at which point the process ends. Thus, when the parse graph state moves the pointer and specifies the next set of bits to extract, the
process 600 returns to 620. - One of ordinary skill in the art will recognize that the specific operations of the
process 600 may not be performed in the exact order shown and described above. Additionally, the specific operations may not be performed in one continuous series of operations, and different specific operations may be performed in different embodiments. For instance, as mentioned above, in some embodiments the process actually extracts the field directly into one of the containers of the PPHV or the SPHV, with the determination at 630 being conceptual. In other embodiments, theprocess 600 identifies the packet headers first and then performs the extraction of the header fields in the identified packet headers. More specifically, in some such embodiments, the process first extracts the first packet header (e.g., based on the configuration data and the parse graph) and after identifying the packet header, extracts the different header fields of the packet header and stores them either in the PPHV or the SPHV. Furthermore, the process could be implemented using several sub-processes, or as part of a larger macro process. - As mentioned, the compiler of some embodiments generates the configuration data for the different switch elements including both the MAU and the parser. The MAU configuration data specifies entries for match and action tables (implemented in TCAMs, or a combination of TCAMs and RAM) of the packet processing pipelines that enable the switch to modify, forward, etc. packets that it receives. The parser configuration data specifies how the parser processes different types of packets, to extract header fields correctly, and to place the extracted header fields into the appropriate containers of the PPHV and SPHV.
-
FIG. 7 conceptually illustrates acompiler 700 of some embodiments for generating parser and MAU configuration data. As shown, the compiler includes acode receiver 705, control and tableflow graph generator 710,MAU configuration generator 715, parsegraph generator 720, andparser configuration generator 725. Thecode receiver 705 receives code defining the switch operations (e.g., as written by a switch designer, network administrator, etc.), and passes this code to thegraph generators 710 and 720 (possibly after translating the code into a different format if needed). In some embodiments, the code is received as P4 code. - The control and table
flow graph generator 710 uses the received code to generate the possible flows through a set of match and action tables that the switch will implement. In some embodiments, the flow graph(s) may be linear or branched, depending on the nature of the code received by the compiler. As shown in the figure, theflow graph generator 710 outputs aflow graph 730, that includes numerous match-action stages. Each match-action stage includes a set of match entries that specify one or more packet header fields (or other data that is not actually part of the packet header, such as ingress or egress ports) over which packets will be matched, as well as corresponding action entries that specify actions to take when a packet matches the associated match entry. These actions may include modifying certain packet header fields, in addition to other actions (e.g., dropping the packet, modifying non-header field data about the packet). TheMAU configuration generator 715 uses the control and table flow graphs to generate the MAU configuration entries, in some embodiments. - The parse
graph generator 720 of some embodiments uses the received code to generate a parse graph specifying the possible paths that a parser can take to parse a received packet. Though shown here as generated from the code, in some embodiments the parse graph is information stored with the compiler, which specifies how to parse any packet. In other embodiments, some of the information (how any packet might be parsed) is stored with the compiler, but a parse graph for the particular switch is generated based on information in the code that indicates what types of packet may actually be processed by the switch (e.g., if all packets will only have Ethernet as the layer 2 protocol). - As shown, the parse
graph generator 720 outputs a parsegraph 735. The parse graph, initially, conceptually describes the different types of packet headers that could be received by the switch, and the order in which those packet headers are arranged. The figure also shows a portion of this parse graph, including the Ethernet header. As shown, the parsegraph 735 specifies (i) how to parse the current header into fields and (ii) how to proceed from the current header to the next header, depending on the value of one or more of the fields of the current header. For the Ethernet header, the first eight bytes are the preamble (and start frame delimiter), the next twelve bytes are the destination MAC address and the source MAC address, and the last two bytes are the EtherType field. In addition, the edges of the parse graph exiting the Ethernet header node are dependent on the value of the EtherType field. For instance, if the EtherType field value is 800, then the next header is IPv4 (with the IPv4 node of the parse graph specifying how to parse the IPv4 header); if the EtherType field value is 806, then the next header is ARP, etc. The parsegraph 735 continues in this manner until reaching a point at which additional data will not be parsed, and this additional data is treated as the payload. In some embodiments, the parse graph (prior to being converted into parser configuration data) continues through the entire packet. - The
parser configuration generator 725 uses the parsegraph 735 as well as the control and table flowgraphs 730 to generate configuration data used to configure the parser in the hardware switch pipeline. The parser configuration data, in some embodiments, includes TCAM entries for each edge of the parse graph that is relevant to the switch (i.e., for each packet header that the switch might need to process). Each of these entries matches over the packet header field values in the current packet header that specify the type of the next packet header (e.g., the EtherType field, the protocol field of an IP packet, etc.). In some embodiments, the TCAM entries point to actions (e.g., in RAM) that specify how to parse the subsequent header (e.g., how to parse the various different fields of an IPv4 packet, an ARP packet, etc.). - In addition, the
parser configuration generator 725 determines which fields of each packet header might actually be used in the match-action unit based on theflow graphs 730. In some embodiments, any packet header field that appears in at least one match entry or is modified by at least one action entry is marked for possible use. While the switch will probably not use the field for every packet containing such a header field, it might for at least some such packets (and for certain fields, the switch may use the field for almost every packet that contains the field). Some embodiments annotate the parse graph to identify the participating and non-participating header fields. This information is then included in the configuration data for the parser, which specifies to send the participating header fields to particular registers of the PPHV and the non-participating header fields to particular registers of the SPHV. In some embodiments, the configuration data also specifies when to stop parsing the packet headers and treat the remainder of the packet as payload. This decision may be made based on the amount of data already stored in the PHVs for the current traversal through the nodes of the parse graph, or because the end of relevant packet headers for the switch have been reached (e.g., the end of the innermost layer 4 headers). -
FIG. 8 conceptually illustrates aprocess 800 performed by the compiler of some embodiments to generate parser configuration data for a hardware switch that parses packet header data into a PPHV and a SPHV. As shown, theprocess 800 begins by receiving (at 805) code for programming the hardware switch. In some embodiments, the code is received as P4 code, describing how the switch should function. - The process then generates (at 810) a parse graph from the code and/or pre-defined packet specifications. In some embodiments the parse graph is information stored with the compiler, which specifies how to parse any packet. In other embodiments, some of the information (how any packet might be parsed) is stored with the compiler, but a parse graph for the particular switch is generated based on information in the code that indicates what types of packets may actually be processed by the switch (e.g., if all packets will only have Ethernet as the layer 2 protocol).
- The parse graph of some embodiments conceptually describes the different types of packet headers that could be received by the switch, and the order in which those packet headers are arranged. Each node in the parse graph represents a packet header, and the parse graph specifies (i) how to parse the header into packet header fields and (ii) how to proceed from the current header to the next header, depending on the value of one or more of the fields of the current header. For the Ethernet header, for example, the first eight bytes are the preamble (and start frame delimiter), the next twelve bytes are the destination MAC address and the source MAC address, and the last two bytes are the EtherType field. The edges of the parse graph exiting each node are dependent on the value of one or more of the field values of the current field.
- The
process 800 also generates (at 815) control flow and table flow graphs for defining the match-action stages of the switch. In some embodiments, the flow graph(s) may be linear or branched, depending on the nature of the code received by the compiler. Each match-action stage in the table flow graph includes a set of match entries that specify one or more packet header fields (or other data that is not actually part of the packet header, such as ingress or egress ports) over which packets will be matched, as well as corresponding action entries that specify actions to take when a packet matches the associated match entry. These actions may include modifying certain packet header fields, in addition to other actions (e.g., dropping the packet, modifying non-header field data about the packet). - Based on the control and table flow graphs, the
process 800 determines (at 820) whether each packet header field specified in the parse graph participates in the match-action stages. The process annotates (at 825) the parse graph with whether each header field will be placed into the PPHV or the SPHV. In some embodiments, any packet header field that appears in at least one match entry or is modified by at least one action entry is marked for possible use. While the switch will probably not use the field for every packet containing such a header field, it might for at least some such packets (and for certain fields, the switch may use the field for almost every packet that contains the field). Participating packet header fields are marked to be sent to the PPHV, while non-participating packet header fields are marked to be sent to the SPHV. - Finally, the
process 800 generates (at 830) configuration entries for the parser based on the annotated parse graph. The parser configuration data, in some embodiments, includes TCAM entries for each edge of the parse graph that is relevant to the switch (i.e., for each packet header that the switch might need to process). Each of these entries matches over the packet header field values in the current packet header that specify the type of the next packet header (e.g., the EtherType field, the protocol field of an IP packet, etc.). In some embodiments, the TCAM entries point to actions (e.g., in RAM) that specify how to parse the subsequent header (e.g., how to parse the various different fields of an IPv4 packet, an ARP packet, etc.). - The configuration data for the parser also specifies to send the participating header fields to particular registers of the PPHV and the non-participating header fields to particular registers of the SPHV. In some embodiments, the configuration data also specifies when to stop parsing the packet headers and treat the remainder of the packet as payload. This decision may be made based on the amount of data already stored in the PHVs for the current traversal through the nodes of the parse graph, or because the end of relevant packet headers for the switch have been reached (e.g., the end of the innermost layer 4 headers).
- Some of the above-described features and applications are implemented as software processes that are specified as a set of instructions recorded on a computer readable storage medium (also referred to as computer readable medium). When these instructions are executed by one or more processing unit(s) (e.g., one or more processors, cores of processors, or other processing units), they cause the processing unit(s) to perform the actions indicated in the instructions. Examples of computer readable media include, but are not limited to, CD-ROMs, flash drives, RAM chips, hard drives, EPROMs, etc. The computer readable media does not include carrier waves and electronic signals passing wirelessly or over wired connections.
- In this specification, the term “software” is meant to include firmware residing in read-only memory or applications stored in magnetic storage, which can be read into memory for processing by a processor. Also, in some embodiments, multiple software inventions can be implemented as sub-parts of a larger program while remaining distinct software inventions. In some embodiments, multiple software inventions can also be implemented as separate programs. Finally, any combination of separate programs that together implement a software invention described here is within the scope of the invention. In some embodiments, the software programs, when installed to operate on one or more electronic systems, define one or more specific machine implementations that execute and perform the operations of the software programs.
-
FIG. 9 conceptually illustrates anelectronic system 900 with which some embodiments of the invention are implemented. Theelectronic system 900 can be used to execute any of the control, virtualization, or operating system applications described above. Theelectronic system 900 may be a computer (e.g., a desktop computer, personal computer, tablet computer, server computer, mainframe, a blade computer etc.), phone, PDA, or any other sort of electronic device. Such an electronic system includes various types of computer readable media and interfaces for various other types of computer readable media.Electronic system 900 includes abus 905, processing unit(s) 910, asystem memory 925, a read-only memory 930, apermanent storage device 935,input devices 940, andoutput devices 945. - The
bus 905 collectively represents all system, peripheral, and chipset buses that communicatively connect the numerous internal devices of theelectronic system 900. For instance, thebus 905 communicatively connects the processing unit(s) 910 with the read-only memory 930, thesystem memory 925, and thepermanent storage device 935. - From these various memory units, the processing unit(s) 910 retrieves instructions to execute and data to process in order to execute the processes of the invention. The processing unit(s) may be a single processor or a multi-core processor in different embodiments.
- The read-only-memory (ROM) 930 stores static data and instructions that are needed by the processing unit(s) 910 and other modules of the electronic system. The
permanent storage device 935, on the other hand, is a read-and-write memory device. This device is a non-volatile memory unit that stores instructions and data even when theelectronic system 900 is off. Some embodiments of the invention use a mass-storage device (such as a magnetic or optical disk and its corresponding disk drive) as thepermanent storage device 935. - Other embodiments use a removable storage device (such as a floppy disk, flash drive, etc.) as the permanent storage device. Like the
permanent storage device 935, thesystem memory 925 is a read-and-write memory device. However, unlikestorage device 935, the system memory is a volatile read-and-write memory, such a random access memory. The system memory stores some of the instructions and data that the processor needs at runtime. In some embodiments, the invention's processes are stored in thesystem memory 925, thepermanent storage device 935, and/or the read-only memory 930. From these various memory units, the processing unit(s) 910 retrieves instructions to execute and data to process in order to execute the processes of some embodiments. - The
bus 905 also connects to the input andoutput devices input devices 940 include alphanumeric keyboards and pointing devices (also called “cursor control devices”). Theoutput devices 945 display images generated by the electronic system. The output devices include printers and display devices, such as cathode ray tubes (CRT) or liquid crystal displays (LCD). Some embodiments include devices such as a touchscreen that function as both input and output devices. - Finally, as shown in
FIG. 9 ,bus 905 also coupleselectronic system 900 to anetwork 965 through a network adapter (not shown). In this manner, the computer can be a part of a network of computers (such as a local area network (“LAN”), a wide area network (“WAN”), or an Intranet, or a network of networks, such as the Internet. Any or all components ofelectronic system 900 may be used in conjunction with the invention. - Some embodiments include electronic components, such as microprocessors, storage and memory that store computer program instructions in a machine-readable or computer-readable medium (alternatively referred to as computer-readable storage media, machine-readable media, or machine-readable storage media). Some examples of such computer-readable media include RAM, ROM, read-only compact discs (CD-ROM), recordable compact discs (CD-R), rewritable compact discs (CD-RW), read-only digital versatile discs (e.g., DVD-ROM, dual-layer DVD-ROM), a variety of recordable/rewritable DVDs (e.g., DVD-RAM, DVD-RW, DVD+RW, etc.), flash memory (e.g., SD cards, mini-SD cards, micro-SD cards, etc.), magnetic and/or solid state hard drives, read-only and recordable Blu-Ray® discs, ultra density optical discs, any other optical or magnetic media, and floppy disks. The computer-readable media may store a computer program that is executable by at least one processing unit and includes sets of instructions for performing various operations. Examples of computer programs or computer code include machine code, such as is produced by a compiler, and files including higher-level code that are executed by a computer, an electronic component, or a microprocessor using an interpreter.
- While the above discussion primarily refers to microprocessor or multi-core processors that execute software, some embodiments are performed by one or more integrated circuits, such as application specific integrated circuits (ASICs) or field programmable gate arrays (FPGAs). In some embodiments, such integrated circuits execute instructions that are stored on the circuit itself
- As used in this specification, the terms “computer”, “server”, “processor”, and “memory” all refer to electronic or other technological devices. These terms exclude people or groups of people. For the purposes of the specification, the terms display or displaying means displaying on an electronic device. As used in this specification, the terms “computer readable medium,” “computer readable media,” and “machine readable medium” are entirely restricted to tangible, physical objects that store information in a form that is readable by a computer. These terms exclude any wireless signals, wired download signals, and any other ephemeral signals.
- While the invention has been described with reference to numerous specific details, one of ordinary skill in the art will recognize that the invention can be embodied in other specific forms without departing from the spirit of the invention. In addition, a number of the figures (including
FIGS. 6 and 8 ) conceptually illustrate processes. The specific operations of these processes may not be performed in the exact order shown and described. The specific operations may not be performed in one continuous series of operations, and different specific operations may be performed in different embodiments. Furthermore, the process could be implemented using several sub-processes, or as part of a larger macro process. Thus, one of ordinary skill in the art would understand that the invention is not to be limited by the foregoing illustrative details, but rather is to be defined by the appended claims.
Claims (19)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/695,053 US20200099619A1 (en) | 2015-08-26 | 2019-11-25 | Packet header field extraction |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US14/836,850 US9825862B2 (en) | 2015-08-26 | 2015-08-26 | Packet header field extraction |
US15/729,555 US10432527B1 (en) | 2015-08-26 | 2017-10-10 | Packet header field extraction |
US16/573,847 US11411870B2 (en) | 2015-08-26 | 2019-09-17 | Packet header field extraction |
US16/695,053 US20200099619A1 (en) | 2015-08-26 | 2019-11-25 | Packet header field extraction |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US16/573,847 Continuation US11411870B2 (en) | 2015-08-26 | 2019-09-17 | Packet header field extraction |
Publications (1)
Publication Number | Publication Date |
---|---|
US20200099619A1 true US20200099619A1 (en) | 2020-03-26 |
Family
ID=58096266
Family Applications (7)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/836,850 Active 2036-02-15 US9825862B2 (en) | 2015-08-26 | 2015-08-26 | Packet header field extraction |
US15/729,555 Active 2036-01-03 US10432527B1 (en) | 2015-08-26 | 2017-10-10 | Packet header field extraction |
US16/573,847 Active US11411870B2 (en) | 2015-08-26 | 2019-09-17 | Packet header field extraction |
US16/695,044 Active 2036-01-02 US11425038B2 (en) | 2015-08-26 | 2019-11-25 | Packet header field extraction |
US16/695,049 Active 2035-11-02 US11425039B2 (en) | 2015-08-26 | 2019-11-25 | Packet header field extraction |
US16/695,053 Abandoned US20200099619A1 (en) | 2015-08-26 | 2019-11-25 | Packet header field extraction |
US17/859,722 Pending US20220345405A1 (en) | 2015-08-26 | 2022-07-07 | Packet header field extraction |
Family Applications Before (5)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/836,850 Active 2036-02-15 US9825862B2 (en) | 2015-08-26 | 2015-08-26 | Packet header field extraction |
US15/729,555 Active 2036-01-03 US10432527B1 (en) | 2015-08-26 | 2017-10-10 | Packet header field extraction |
US16/573,847 Active US11411870B2 (en) | 2015-08-26 | 2019-09-17 | Packet header field extraction |
US16/695,044 Active 2036-01-02 US11425038B2 (en) | 2015-08-26 | 2019-11-25 | Packet header field extraction |
US16/695,049 Active 2035-11-02 US11425039B2 (en) | 2015-08-26 | 2019-11-25 | Packet header field extraction |
Family Applications After (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/859,722 Pending US20220345405A1 (en) | 2015-08-26 | 2022-07-07 | Packet header field extraction |
Country Status (1)
Country | Link |
---|---|
US (7) | US9825862B2 (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11223520B1 (en) | 2017-01-31 | 2022-01-11 | Intel Corporation | Remote control plane directing data plane configurator |
US11362967B2 (en) | 2017-09-28 | 2022-06-14 | Barefoot Networks, Inc. | Expansion of packet data within processing pipeline |
US11388053B2 (en) | 2014-12-27 | 2022-07-12 | Intel Corporation | Programmable protocol parser for NIC classification and queue assignments |
US11411870B2 (en) | 2015-08-26 | 2022-08-09 | Barefoot Networks, Inc. | Packet header field extraction |
US11425058B2 (en) | 2017-04-23 | 2022-08-23 | Barefoot Networks, Inc. | Generation of descriptive data for packet fields |
US11503141B1 (en) | 2017-07-23 | 2022-11-15 | Barefoot Networks, Inc. | Stateful processing unit with min/max capability |
US11677851B2 (en) | 2015-12-22 | 2023-06-13 | Intel Corporation | Accelerated network packet processing |
Families Citing this family (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9826071B2 (en) | 2015-08-26 | 2017-11-21 | Barefoot Networks, Inc. | Configuring a switch for extracting packet header fields |
US10911579B1 (en) * | 2016-03-01 | 2021-02-02 | Amazon Technologies, Inc. | Generating programmatically defined fields of metadata for network packets |
CN105812365B (en) * | 2016-03-12 | 2018-09-28 | 武汉芯泰科技有限公司 | A kind of programmable restructural Frame Handler |
US10594574B2 (en) * | 2016-11-07 | 2020-03-17 | Hewlett Packard Enterprise Development Lp | Middlebox modeling |
US10404594B2 (en) | 2016-12-13 | 2019-09-03 | Oracle International Corporation | System and method for providing partitions of classification resources in a network device |
US10341242B2 (en) * | 2016-12-13 | 2019-07-02 | Oracle International Corporation | System and method for providing a programmable packet classification framework for use in a network device |
US10764176B1 (en) | 2017-07-09 | 2020-09-01 | Barefoot Networks, Inc. | Compiler and hardware interactions to reuse register fields in the data plane of a network forwarding element |
US10511532B2 (en) | 2018-03-06 | 2019-12-17 | Barefoot Networks, Inc. | Algorithmic longest prefix matching in programmable switch |
WO2019178264A1 (en) * | 2018-03-14 | 2019-09-19 | Fungible, Inc. | Flexible processing of network packets |
CN113542125B (en) * | 2018-03-31 | 2022-11-25 | 华为技术有限公司 | Method and device for forwarding message based on integrated flow table |
US11218574B2 (en) | 2018-06-08 | 2022-01-04 | Fungible, Inc. | Directed graph traversal using content-addressable memory |
US11483245B2 (en) * | 2018-09-13 | 2022-10-25 | Intel Corporation | Technologies for filtering network traffic on ingress |
US10958770B2 (en) | 2018-10-15 | 2021-03-23 | Fungible, Inc. | Realization of a programmable forwarding pipeline through packet header summaries in a data processing unit |
US11431829B2 (en) | 2019-03-06 | 2022-08-30 | Parsons Corporation | Multi-tiered packet processing |
US11258726B2 (en) | 2019-03-27 | 2022-02-22 | Fungible, Inc. | Low latency packet switch architecture |
US11301231B2 (en) * | 2019-04-05 | 2022-04-12 | Arista Networks, Inc. | Dynamic run time programming of hardware tables |
US11656992B2 (en) | 2019-05-03 | 2023-05-23 | Western Digital Technologies, Inc. | Distributed cache with in-network prefetch |
US11153360B2 (en) * | 2019-05-21 | 2021-10-19 | Genetec Inc. | Methods and systems for codec detection in video streams |
US11388078B1 (en) * | 2019-06-10 | 2022-07-12 | Keysight Technologies, Inc. | Methods, systems, and computer readable media for generating and using statistically varying network traffic mixes to test network devices |
US11616850B1 (en) | 2019-08-23 | 2023-03-28 | Fitbit, Inc. | Connection management techniques |
US11579802B2 (en) | 2019-10-04 | 2023-02-14 | Fungible, Inc. | Pipeline using match-action blocks |
US11765250B2 (en) | 2020-06-26 | 2023-09-19 | Western Digital Technologies, Inc. | Devices and methods for managing network traffic for a distributed cache |
US11675706B2 (en) | 2020-06-30 | 2023-06-13 | Western Digital Technologies, Inc. | Devices and methods for failure detection and recovery for a distributed cache |
US11736417B2 (en) | 2020-08-17 | 2023-08-22 | Western Digital Technologies, Inc. | Devices and methods for network message sequencing |
TWI745034B (en) * | 2020-08-18 | 2021-11-01 | 國立陽明交通大學 | Method of aggregating and disaggregating packet |
US20230147422A1 (en) * | 2021-10-26 | 2023-05-11 | Arista Networks, Inc. | Filtering vlan packets having malformed tpids |
Family Cites Families (265)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5243596A (en) | 1992-03-18 | 1993-09-07 | Fischer & Porter Company | Network architecture suitable for multicasting and resource locking |
JP2541460B2 (en) | 1993-07-30 | 1996-10-09 | 日本電気株式会社 | User status broadcasting with transmission restrictions |
US5784003A (en) | 1996-03-25 | 1998-07-21 | I-Cube, Inc. | Network switch with broadcast support |
US6442172B1 (en) | 1996-07-11 | 2002-08-27 | Alcatel Internetworking, Inc. | Input buffering and queue status-based output control for a digital traffic switch |
US6157955A (en) | 1998-06-15 | 2000-12-05 | Intel Corporation | Packet processing system including a policy engine having a classification unit |
US6836483B1 (en) | 1998-06-24 | 2004-12-28 | Research Investment Network, Inc. | Message system for asynchronous transfer |
WO2000003522A1 (en) | 1998-07-08 | 2000-01-20 | Broadcom Corporation | A method of sending packets between trunk ports of network switches |
US6912637B1 (en) | 1998-07-08 | 2005-06-28 | Broadcom Corporation | Apparatus and method for managing memory in a network switch |
US7430171B2 (en) | 1998-11-19 | 2008-09-30 | Broadcom Corporation | Fibre channel arbitrated loop bufferless switch circuitry to increase bandwidth without significant increase in cost |
JP3684308B2 (en) | 1998-12-15 | 2005-08-17 | 富士通株式会社 | Scheduling control device and exchange |
US6453360B1 (en) | 1999-03-01 | 2002-09-17 | Sun Microsystems, Inc. | High performance network interface |
US6948099B1 (en) | 1999-07-30 | 2005-09-20 | Intel Corporation | Re-loading operating systems |
JP3643507B2 (en) | 1999-09-20 | 2005-04-27 | 株式会社東芝 | Packet processing apparatus and packet processing method |
US7203740B1 (en) | 1999-12-22 | 2007-04-10 | Intel Corporation | Method and apparatus for allowing proprietary forwarding elements to interoperate with standard control elements in an open architecture for network devices |
US6549604B2 (en) | 1999-12-28 | 2003-04-15 | Symmetricom, Inc. | Clock recovery and detection of rapid phase transients |
US6977930B1 (en) | 2000-02-14 | 2005-12-20 | Cisco Technology, Inc. | Pipelined packet switching and queuing architecture |
FI109061B (en) | 2000-05-10 | 2002-05-15 | Nokia Corp | Resource reservation on a packet network |
US7062571B1 (en) * | 2000-06-30 | 2006-06-13 | Cisco Technology, Inc. | Efficient IP load-balancing traffic distribution using ternary CAMs |
US7180857B2 (en) | 2000-11-24 | 2007-02-20 | Matsushita Electric Industrial Co., Ltd | Apparatus and method for flow control |
US7418523B2 (en) | 2001-01-25 | 2008-08-26 | Ericsson Ab | Operation of a multiplicity of time sorted queues with reduced memory |
US6976149B1 (en) | 2001-02-22 | 2005-12-13 | Cisco Technology, Inc. | Mapping technique for computing addresses in a memory of an intermediate network node |
US6904057B2 (en) * | 2001-05-04 | 2005-06-07 | Slt Logic Llc | Method and apparatus for providing multi-protocol, multi-stage, real-time frame classification |
US7082104B2 (en) | 2001-05-18 | 2006-07-25 | Intel Corporation | Network device switch |
US20030009466A1 (en) | 2001-06-21 | 2003-01-09 | Ta John D. C. | Search engine with pipeline structure |
US7418536B2 (en) | 2001-07-30 | 2008-08-26 | Cisco Technology, Inc. | Processor having systolic array pipeline for processing data packets |
US20030046429A1 (en) * | 2001-08-30 | 2003-03-06 | Sonksen Bradley Stephen | Static data item processing |
US6909713B2 (en) | 2001-09-05 | 2005-06-21 | Intel Corporation | Hash-based data frame distribution for web switches |
US20030063345A1 (en) | 2001-10-01 | 2003-04-03 | Dan Fossum | Wayside user communications over optical supervisory channel |
AU2002361830A1 (en) | 2001-12-21 | 2003-07-15 | Chip Engines | Reconfigurable data packet header processor |
KR100464406B1 (en) | 2002-02-08 | 2005-01-03 | 삼성전자주식회사 | Apparatus and method for dispatching very long instruction word with variable length |
US7100011B2 (en) | 2002-03-01 | 2006-08-29 | Arris International, Inc. | Method and system for reducing storage requirements for program code in a communication device |
US9203135B2 (en) | 2002-04-23 | 2015-12-01 | Piedek Technical Laboratory | Method for manufacturing quartz crystal resonator |
US20120155466A1 (en) | 2002-05-06 | 2012-06-21 | Ian Edward Davis | Method and apparatus for efficiently processing data packets in a computer network |
US7142513B2 (en) | 2002-05-23 | 2006-11-28 | Yea-Li Sun | Method and multi-queue packet scheduling system for managing network packet traffic with minimum performance guarantees and maximum service rate control |
US7415723B2 (en) | 2002-06-11 | 2008-08-19 | Pandya Ashish A | Distributed network security system and a hardware processor therefor |
US8015303B2 (en) | 2002-08-02 | 2011-09-06 | Astute Networks Inc. | High data rate stateful protocol processing |
US20040031029A1 (en) | 2002-08-06 | 2004-02-12 | Kyu-Woong Lee | Methods and systems for automatically updating software components in a network |
US7245586B2 (en) | 2002-08-30 | 2007-07-17 | Lucent Technologies Inc. | Buffer management based on buffer sharing across ports and per-port minimum buffer guarantee |
US7539777B1 (en) | 2002-10-25 | 2009-05-26 | Cisco Technology, Inc. | Method and system for network time protocol forwarding |
US7324460B2 (en) | 2002-11-28 | 2008-01-29 | International Business Machines Corporation | Event-driven flow control for a very high-speed switching node |
US7050394B2 (en) | 2002-12-18 | 2006-05-23 | Intel Corporation | Framer |
US7492714B1 (en) | 2003-02-04 | 2009-02-17 | Pmc-Sierra, Inc. | Method and apparatus for packet grooming and aggregation |
US7522523B2 (en) | 2003-04-25 | 2009-04-21 | Alcatel-Lucent Usa Inc. | Assigning packet queue priority |
US20050108518A1 (en) | 2003-06-10 | 2005-05-19 | Pandya Ashish A. | Runtime adaptable security processor |
US20050013251A1 (en) | 2003-07-18 | 2005-01-20 | Hsuan-Wen Wang | Flow control hub having scoreboard memory |
EP1509011A3 (en) | 2003-08-16 | 2011-01-19 | Samsung Electronics Co., Ltd. | Method and apparatus for assigning scheduling for uplink packet transmission in a mobile communication system |
US8018852B2 (en) | 2003-08-22 | 2011-09-13 | Alcatel Lucent | Equal-cost source-resolved routing system and method |
US7464181B2 (en) | 2003-09-11 | 2008-12-09 | International Business Machines Corporation | Method for caching lookups based upon TCP traffic flow characteristics |
US7111102B2 (en) | 2003-10-06 | 2006-09-19 | Cisco Technology, Inc. | Port adapter for high-bandwidth bus |
US7787471B2 (en) * | 2003-11-10 | 2010-08-31 | Broadcom Corporation | Field processor for a network device |
JP4376040B2 (en) | 2003-11-27 | 2009-12-02 | 株式会社日立製作所 | Apparatus and method for performing information processing using a plurality of processors |
CN1625176B (en) | 2003-12-03 | 2010-04-28 | 华为技术有限公司 | Realization method of edge to edge pseudo-line simulation protocol |
KR20050057698A (en) | 2003-12-10 | 2005-06-16 | 삼성전자주식회사 | Apparatus and method for generating checksum |
US7685434B2 (en) * | 2004-03-02 | 2010-03-23 | Advanced Micro Devices, Inc. | Two parallel engines for high speed transmit IPsec processing |
US7889750B1 (en) | 2004-04-28 | 2011-02-15 | Extreme Networks, Inc. | Method of extending default fixed number of processing cycles in pipelined packet processor architecture |
US8665719B2 (en) | 2004-05-03 | 2014-03-04 | Alcatel Lucent | Variable packet-size backplanes for switching and routing systems |
US7760719B2 (en) * | 2004-06-30 | 2010-07-20 | Conexant Systems, Inc. | Combined pipelined classification and address search method and apparatus for switching environments |
US7453797B2 (en) | 2004-09-29 | 2008-11-18 | Intel Corporation | Method to provide high availability in network elements using distributed architectures |
US7826470B1 (en) | 2004-10-19 | 2010-11-02 | Broadcom Corp. | Network interface device with flow-oriented bus interface |
US8953432B2 (en) | 2004-11-01 | 2015-02-10 | Alcatel Lucent | Softrouter dynamic binding protocol |
US8000324B2 (en) | 2004-11-30 | 2011-08-16 | Broadcom Corporation | Pipeline architecture of a network device |
US8170019B2 (en) | 2004-11-30 | 2012-05-01 | Broadcom Corporation | CPU transmission of unmodified packets |
US7873959B2 (en) | 2005-02-01 | 2011-01-18 | Microsoft Corporation | Publishing the status of and updating firmware components |
US20070050426A1 (en) | 2005-06-20 | 2007-03-01 | Dubal Scott P | Platform with management agent to receive software updates |
US7505410B2 (en) | 2005-06-30 | 2009-03-17 | Intel Corporation | Method and apparatus to support efficient check-point and role-back operations for flow-controlled queues in network devices |
US7499941B2 (en) | 2005-09-05 | 2009-03-03 | Cisco Technology, Inc. | Pipeline regular expression matching |
US7953002B2 (en) | 2005-11-10 | 2011-05-31 | Broadcom Corporation | Buffer management and flow control mechanism including packet-based dynamic thresholding |
US8218546B2 (en) | 2005-11-10 | 2012-07-10 | Broadcom Corporation | Interleaved processing of dropped packets in a network device |
US20070153796A1 (en) | 2005-12-30 | 2007-07-05 | Intel Corporation | Packet processing utilizing cached metadata to support forwarding and non-forwarding operations on parallel paths |
US7593330B1 (en) | 2006-01-30 | 2009-09-22 | Juniper Networks, Inc. | Processing of partial frames and partial superframes |
US7809009B2 (en) * | 2006-02-21 | 2010-10-05 | Cisco Technology, Inc. | Pipelined packet switching and queuing architecture |
RU2491737C2 (en) | 2006-03-31 | 2013-08-27 | Квэлкомм Инкорпорейтед | Memory management for high-speed medium access control |
JP4236672B2 (en) | 2006-04-05 | 2009-03-11 | シャープ株式会社 | Image processing device |
US8665892B2 (en) | 2006-05-30 | 2014-03-04 | Broadcom Corporation | Method and system for adaptive queue and buffer control based on monitoring in a packet network switch |
US7701849B1 (en) | 2006-06-23 | 2010-04-20 | Juniper Networks, Inc. | Flow-based queuing of network traffic |
US8077611B2 (en) | 2006-07-27 | 2011-12-13 | Cisco Technology, Inc. | Multilevel coupled policer |
US7835277B2 (en) | 2006-12-05 | 2010-11-16 | Samsung Electronics Co., Ltd. | Method and apparatus for managing a buffer in a communication system |
US7746901B2 (en) | 2006-12-14 | 2010-06-29 | Oracle America, Inc. | Method and system for offloaded transport layer protocol switching |
CN101563908B (en) * | 2006-12-19 | 2013-01-09 | 国际商业机器公司 | Apparatus and method for analysing a network flow |
US20080175449A1 (en) | 2007-01-19 | 2008-07-24 | Wison Technology Corp. | Fingerprint-based network authentication method and system thereof |
US7904642B1 (en) * | 2007-02-08 | 2011-03-08 | Netlogic Microsystems, Inc. | Method for combining and storing access control lists |
US8458282B2 (en) | 2007-06-26 | 2013-06-04 | International Business Machines Corporation | Extended write combining using a write continuation hint flag |
US8094659B1 (en) * | 2007-07-09 | 2012-01-10 | Marvell Israel (M.I.S.L) Ltd. | Policy-based virtual routing and forwarding (VRF) assignment |
US8458497B2 (en) | 2007-10-11 | 2013-06-04 | Qualcomm Incorporated | Demand based power control in a graphics processing unit |
US8156309B2 (en) | 2007-10-18 | 2012-04-10 | Cisco Technology, Inc. | Translation look-aside buffer with variable page sizes |
JP5088145B2 (en) | 2008-01-10 | 2012-12-05 | 富士通株式会社 | Packet relay apparatus, control method, and packet relay program |
US7961734B2 (en) | 2008-09-30 | 2011-06-14 | Juniper Networks, Inc. | Methods and apparatus related to packet classification associated with a multi-stage switch |
WO2010056723A1 (en) | 2008-11-12 | 2010-05-20 | Collective Media, Inc. | Method and system for semantic distance measurement |
RU2009144127A (en) | 2008-12-01 | 2011-06-10 | РАЗУМ, Инк. (US) | SERVICE QUALITY MANAGEMENT BASED ON MONITORING THE STATE OF THE STREAM STREAM WITHOUT USER SIGNALING |
US8118238B2 (en) | 2008-12-10 | 2012-02-21 | Honeywell International Inc. | User interface for building controller |
US8214060B2 (en) | 2008-12-10 | 2012-07-03 | Honeywell International Inc. | Building appliance controller with safety feature |
JP5168166B2 (en) | 2009-01-21 | 2013-03-21 | 富士通株式会社 | Communication apparatus and communication control method |
EP2389630A4 (en) | 2009-01-26 | 2012-08-01 | Assa Abloy Ab | Provisioned firmware updates using object identifiers |
US8391163B2 (en) | 2009-03-23 | 2013-03-05 | Cisco Technology, Inc. | Operating MPLS label switched paths and MPLS pseudowire in loopback mode |
US8638793B1 (en) | 2009-04-06 | 2014-01-28 | Marvell Israle (M.I.S.L) Ltd. | Enhanced parsing and classification in a packet processor |
US9049271B1 (en) | 2009-07-16 | 2015-06-02 | Teradici Corporation | Switch-initiated congestion management method |
US8571031B2 (en) * | 2009-10-07 | 2013-10-29 | Intel Corporation | Configurable frame processing pipeline in a packet switch |
US20110149960A1 (en) | 2009-12-17 | 2011-06-23 | Media Patents, S.L. | Method and apparatus for filtering multicast packets |
KR101292873B1 (en) * | 2009-12-21 | 2013-08-02 | 한국전자통신연구원 | Network interface card device and method of processing traffic by using the network interface card device |
US9112818B1 (en) | 2010-02-05 | 2015-08-18 | Marvell Isreal (M.I.S.L) Ltd. | Enhanced tail dropping in a switch |
US8869138B2 (en) | 2011-11-11 | 2014-10-21 | Wyse Technology L.L.C. | Robust firmware update with recovery logic |
EP2381622B1 (en) | 2010-04-23 | 2012-06-20 | Alcatel Lucent | Update of a cumulative residence time of a packet in a packet-switched communication network |
US8514855B1 (en) * | 2010-05-04 | 2013-08-20 | Sandia Corporation | Extensible packet processing architecture |
CN103119900B (en) | 2010-06-23 | 2016-01-27 | 日本电气株式会社 | Communication system, control appliance, node control method and node control program |
US8880468B2 (en) | 2010-07-06 | 2014-11-04 | Nicira, Inc. | Secondary storage architecture for a network control system that utilizes a primary network information base |
US8738860B1 (en) | 2010-10-25 | 2014-05-27 | Tilera Corporation | Computing in parallel processing environments |
US9047251B2 (en) | 2010-12-20 | 2015-06-02 | Citrix Systems, Inc. | Systems and methods for implementing connection mirroring in a multi-core system |
US9319276B2 (en) * | 2010-12-21 | 2016-04-19 | Cisco Technology, Inc. | Client modeling in a forwarding plane |
US8798077B2 (en) | 2010-12-29 | 2014-08-05 | Juniper Networks, Inc. | Methods and apparatus for standard protocol validation mechanisms deployed over a switch fabric system |
US9003057B2 (en) | 2011-01-04 | 2015-04-07 | Cisco Technology, Inc. | System and method for exchanging information in a mobile wireless network environment |
CN103004158B (en) * | 2011-01-06 | 2016-12-28 | 马维尔以色列(M.I.S.L)有限公司 | There is the network equipment of programmable core |
US20150142991A1 (en) | 2011-04-21 | 2015-05-21 | Efficiency3 Corp. | Electronic hub appliances used for collecting, storing, and processing potentially massive periodic data streams indicative of real-time or other measuring parameters |
WO2012143953A2 (en) | 2011-04-21 | 2012-10-26 | Ineda Systems Pvt. Ltd | Optimized multi-root input output virtualization aware switch |
WO2012154596A1 (en) | 2011-05-06 | 2012-11-15 | Xcelemor, Inc. | Computing system with data and control planes and method of operation thereof |
JP6072782B2 (en) | 2011-06-10 | 2017-02-01 | フィリップス ライティング ホールディング ビー ヴィ | Executing secure protocols on the network |
US8442056B2 (en) | 2011-06-28 | 2013-05-14 | Marvell International Ltd. | Scheduling packets in a packet-processing pipeline |
US8798047B1 (en) * | 2011-08-29 | 2014-08-05 | Qlogic, Corporation | Methods and systems for processing network information |
US9204207B2 (en) | 2011-11-01 | 2015-12-01 | Plexxi Inc. | Hierarchy of control in a data center network |
US8718064B2 (en) * | 2011-12-22 | 2014-05-06 | Telefonaktiebolaget L M Ericsson (Publ) | Forwarding element for flexible and extensible flow processing software-defined networks |
US8711860B2 (en) * | 2011-12-22 | 2014-04-29 | Telefonaktiebolaget L M Ericsson (Publ) | Controller for flexible and extensible flow processing in software-defined networks |
US8521905B2 (en) * | 2011-12-22 | 2013-08-27 | Telefonaktiebolaget L M Ericsson (Publ) | System for flexible and extensible flow processing in software-defined networks |
US9055114B1 (en) * | 2011-12-22 | 2015-06-09 | Juniper Networks, Inc. | Packet parsing and control packet classification |
WO2013101765A1 (en) | 2011-12-27 | 2013-07-04 | Cisco Technology, Inc. | System and method for management of network-based services |
US8971338B2 (en) | 2012-01-09 | 2015-03-03 | Telefonaktiebolaget L M Ericsson (Publ) | Expanding network functionalities for openflow based split-architecture networks |
US9519614B2 (en) | 2012-01-10 | 2016-12-13 | Verizon Digital Media Services Inc. | Multi-layer multi-hit caching for long tail content |
US20160342510A1 (en) | 2012-01-17 | 2016-11-24 | Google Inc. | Remote management of data planes and configuration of networking devices |
WO2013115177A1 (en) | 2012-01-30 | 2013-08-08 | 日本電気株式会社 | Network system and topology management method |
US9608893B2 (en) | 2012-02-27 | 2017-03-28 | The Boeing Company | Methods and systems for parsing data objects |
US9438505B1 (en) * | 2012-03-29 | 2016-09-06 | Google Inc. | System and method for increasing capacity in router forwarding tables |
US8886879B2 (en) * | 2012-04-27 | 2014-11-11 | Hewlett-Packard Development Company, L.P. | TCAM action updates |
US8788512B2 (en) | 2012-05-23 | 2014-07-22 | International Business Machines Corporation | Generating data feed specific parser circuits |
US9438527B2 (en) | 2012-05-24 | 2016-09-06 | Marvell World Trade Ltd. | Flexible queues in a network switch |
US8856600B2 (en) | 2012-06-21 | 2014-10-07 | Breakingpoint Systems, Inc. | JTAG-based programming and debug |
US9065755B2 (en) | 2012-08-07 | 2015-06-23 | Broadcom Corporation | Low-latency switching |
US8943268B2 (en) * | 2012-08-17 | 2015-01-27 | Broadcom Corporation | Ternary content addressable memory (TCAM) storage system |
US9015427B2 (en) | 2012-09-14 | 2015-04-21 | Xerox Corporation | Systems and methods for employing an electronically-readable monitoring module associated with a customer replaceable component to update a non-volatile memory in an image forming device |
US9055004B2 (en) | 2012-09-18 | 2015-06-09 | Cisco Technology, Inc. | Scalable low latency multi-protocol networking device |
CN103777971A (en) | 2012-10-23 | 2014-05-07 | 新加坡商华科国际股份有限公司 | Electronic device, method for updating firmware, and recording medium capable of being read by computer |
US10104004B2 (en) | 2012-11-08 | 2018-10-16 | Texas Instruments Incorporated | Openflow match and action pipeline structure |
US9419903B2 (en) | 2012-11-08 | 2016-08-16 | Texas Instruments Incorporated | Structure for implementing openflow all group buckets using egress flow table entries |
US8711855B1 (en) | 2012-12-18 | 2014-04-29 | Juniper Networks, Inc. | Topology discovery, control channel establishment, and datapath provisioning within an aggregation network with centralized control |
US9384072B2 (en) | 2012-12-20 | 2016-07-05 | Oracle International Corporation | Distributed queue pair state on a host channel adapter |
US9350656B2 (en) | 2013-01-24 | 2016-05-24 | Knuedge Incorporated | Systems and methods for packet routing |
US20150363522A1 (en) | 2013-01-31 | 2015-12-17 | Hewlett-Packard Development Company, L.P. | Network switch simulation |
US9906445B2 (en) | 2013-02-01 | 2018-02-27 | Texas Instruments Incorporated | Packet processing match and action pipeline structure with dependency calculation removing false dependencies |
US10645032B2 (en) | 2013-02-28 | 2020-05-05 | Texas Instruments Incorporated | Packet processing match and action unit with stateful actions |
US9258224B2 (en) | 2013-02-28 | 2016-02-09 | Texas Instruments Incorporated | Packet processing VLIW action unit with or-multi-ported instruction memory |
US10009276B2 (en) | 2013-02-28 | 2018-06-26 | Texas Instruments Incorporated | Packet processing match and action unit with a VLIW action engine |
US9712439B2 (en) | 2013-02-28 | 2017-07-18 | Texas Instruments Incorporated | Packet processing match and action unit with configurable memory allocation |
US9826067B2 (en) | 2013-02-28 | 2017-11-21 | Texas Instruments Incorporated | Packet processing match and action unit with configurable bit allocation |
KR20150132151A (en) | 2013-03-12 | 2015-11-25 | 휴렛-팩커드 디벨롭먼트 컴퍼니, 엘.피. | Programmable address mapping and memory access operations |
US9276846B2 (en) | 2013-03-15 | 2016-03-01 | Cavium, Inc. | Packet extraction optimization in a network processor |
US9450817B1 (en) | 2013-03-15 | 2016-09-20 | Juniper Networks, Inc. | Software defined network controller |
US9419919B2 (en) | 2013-03-15 | 2016-08-16 | Cisco Technology, Inc. | VPC auto configuration |
US9973429B2 (en) | 2013-04-05 | 2018-05-15 | Futurewei Technologies, Inc. | Software defined networking (SDN) controller orchestration and network virtualization for data center interconnection |
US20140321473A1 (en) | 2013-04-26 | 2014-10-30 | Mediatek Inc. | Active output buffer controller for controlling packet data output of main buffer in network device and related method |
US9667447B2 (en) | 2013-07-08 | 2017-05-30 | Nicira, Inc. | Managing context identifier assignment across multiple physical domains |
US9571386B2 (en) * | 2013-07-08 | 2017-02-14 | Nicira, Inc. | Hybrid packet processing |
US9124644B2 (en) | 2013-07-14 | 2015-09-01 | Netronome Systems, Inc. | Script-controlled egress packet modifier |
US9712431B2 (en) | 2013-07-17 | 2017-07-18 | Kt Corporation | Methods for managing transaction in software defined network |
CN104348750B (en) | 2013-07-31 | 2019-07-26 | 中兴通讯股份有限公司 | The implementation method and device of QoS in OpenFlow network |
CN103428094B (en) * | 2013-08-12 | 2016-08-17 | 杭州华三通信技术有限公司 | Message forwarding method in open flows OpenFlow system and device |
WO2015031866A1 (en) | 2013-08-30 | 2015-03-05 | Clearpath Networks, Inc. | System and method of network functions virtualization of network services within and across clouds |
US9602398B2 (en) * | 2013-09-15 | 2017-03-21 | Nicira, Inc. | Dynamically generating flows with wildcard fields |
US9722916B2 (en) | 2013-09-30 | 2017-08-01 | Cisco Technology, Inc. | Data-plane driven fast protection mechanism for MPLS pseudowire services |
CN104579968B (en) | 2013-10-26 | 2018-03-09 | 华为技术有限公司 | SDN switch obtains accurate flow table item method and SDN switch, controller, system |
US9298446B2 (en) | 2013-10-28 | 2016-03-29 | International Business Machines Corporation | Unified update tool for multi-protocol network adapter |
KR20150053620A (en) * | 2013-11-08 | 2015-05-18 | 한국전자통신연구원 | Apparatus and method for transmitting packet |
US9577924B2 (en) * | 2013-11-14 | 2017-02-21 | Electronics And Telecommunications Research Institute | SDN-based network device with extended function and method of processing packet in the same device |
FI20136138L (en) | 2013-11-18 | 2015-05-19 | Tellabs Oy | A network element and a controller to manage the network element |
US9973599B2 (en) | 2013-12-04 | 2018-05-15 | Mediatek Inc. | Parser for parsing header in packet and related packet processing apparatus |
US9674102B2 (en) * | 2013-12-18 | 2017-06-06 | Marvell Israel (M.I.S.L.) Ltd. | Methods and network device for oversubscription handling |
US20150178747A1 (en) | 2013-12-20 | 2015-06-25 | Zumur, LLC | System and method for determining and distributing consumer items according to dynamic demand levels |
US20150180769A1 (en) | 2013-12-20 | 2015-06-25 | Alcatel-Lucent Usa Inc. | Scale-up of sdn control plane using virtual switch based overlay |
US9825884B2 (en) | 2013-12-30 | 2017-11-21 | Cavium, Inc. | Protocol independent programmable switch (PIPS) software defined data center networks |
CN104753818B (en) | 2013-12-30 | 2019-04-30 | 深圳市中兴微电子技术有限公司 | A kind of array dispatching method and device |
US9299434B2 (en) * | 2014-01-09 | 2016-03-29 | Netronome Systems, Inc. | Dedicated egress fast path for non-matching packets in an OpenFlow switch |
US9686204B2 (en) | 2014-02-05 | 2017-06-20 | Verizon Patent And Licensing Inc. | Capacity management based on backlog information |
CN106416148A (en) | 2014-02-19 | 2017-02-15 | 日本电气株式会社 | Network control method, network system, device, and program |
US20150249572A1 (en) | 2014-03-03 | 2015-09-03 | Futurewei Technologies, Inc. | Software-Defined Network Control Using Functional Objects |
US20150256465A1 (en) * | 2014-03-04 | 2015-09-10 | Futurewei Technologies, Inc. | Software-Defined Network Control Using Control Macros |
US9985896B2 (en) * | 2014-03-31 | 2018-05-29 | Nicira, Inc. | Caching of service decisions |
WO2015167462A1 (en) | 2014-04-29 | 2015-11-05 | Hewlett-Packard Development Company, L.P. | Network re-convergence point |
US9986434B2 (en) | 2014-04-30 | 2018-05-29 | Avago Technologies General Ip (Singapore) Pte. Ltd. | System for accelerated network route update through exclusive access to routing tables |
WO2015180761A1 (en) | 2014-05-27 | 2015-12-03 | Telecom Italia S.P.A. | System and method for network apparatus management |
EP3139549B1 (en) | 2014-05-30 | 2021-03-03 | Huawei Technologies Co. Ltd. | Packet editing method and related device |
SG10201402893YA (en) | 2014-06-04 | 2016-01-28 | Lantiq Deutschland Gmbh | Data packet processing system on a chip |
US20150381418A1 (en) | 2014-06-27 | 2015-12-31 | iPhotonix | Remote Orchestration of Virtual Machine Updates |
US9838268B1 (en) | 2014-06-27 | 2017-12-05 | Juniper Networks, Inc. | Distributed, adaptive controller for multi-domain networks |
US9419897B2 (en) | 2014-06-30 | 2016-08-16 | Nicira, Inc. | Methods and systems for providing multi-tenancy support for Single Root I/O Virtualization |
WO2016000170A1 (en) * | 2014-06-30 | 2016-01-07 | 华为技术有限公司 | Data processing method executed by network apparatus, and associated device |
US10044795B2 (en) | 2014-07-11 | 2018-08-07 | Vmware Inc. | Methods and apparatus for rack deployments for virtual computing environments |
WO2016012051A1 (en) | 2014-07-25 | 2016-01-28 | Telefonaktiebolaget L M Ericsson (Publ) | Packet processing in an openflow switch |
CN105684382A (en) | 2014-08-11 | 2016-06-15 | 华为技术有限公司 | Packet control method, switch and controller |
US9755932B1 (en) | 2014-09-26 | 2017-09-05 | Juniper Networks, Inc. | Monitoring packet residence time and correlating packet residence time to input sources |
US11178051B2 (en) * | 2014-09-30 | 2021-11-16 | Vmware, Inc. | Packet key parser for flow-based forwarding elements |
US9960956B1 (en) * | 2014-10-15 | 2018-05-01 | The United States Of America, As Represented By The Secretary Of The Navy | Network monitoring method using phantom nodes |
US9606781B2 (en) * | 2014-11-14 | 2017-03-28 | Cavium, Inc. | Parser engine programming tool for programmable network devices |
US9838286B2 (en) | 2014-11-20 | 2017-12-05 | Telefonaktiebolaget L M Ericsson (Publ) | Passive performance measurement for inline service chaining |
US10044646B1 (en) | 2014-12-02 | 2018-08-07 | Adtran, Inc. | Systems and methods for efficiently storing packet data in network switches |
CN105721332A (en) | 2014-12-03 | 2016-06-29 | 深圳市中兴微电子技术有限公司 | Congestion control method and device based on improved WRED |
EP3654593B1 (en) | 2014-12-11 | 2021-11-24 | AT&T Intellectual Property I, L.P. | Multilayered distributed router architecture |
US10003544B2 (en) | 2014-12-11 | 2018-06-19 | Futurewei Technologies, Inc. | Method and apparatus for priority flow and congestion control in ethernet network |
US9553807B2 (en) * | 2014-12-24 | 2017-01-24 | Nicira, Inc. | Batch processing of packets |
US10015048B2 (en) | 2014-12-27 | 2018-07-03 | Intel Corporation | Programmable protocol parser for NIC classification and queue assignments |
US9794263B2 (en) * | 2014-12-27 | 2017-10-17 | Intel Corporation | Technologies for access control |
US9846576B2 (en) | 2014-12-27 | 2017-12-19 | Intel Corporation | Technologies for reprogramming network interface cards over a network |
US9602387B2 (en) | 2014-12-29 | 2017-03-21 | Juniper Networks, Inc. | Network topology optimization |
US10193783B2 (en) * | 2014-12-31 | 2019-01-29 | Nicira, Inc. | System for aggregating statistics associated with interfaces |
CN107113637B (en) | 2015-01-12 | 2021-02-26 | 瑞典爱立信有限公司 | Method and module for managing packets in a software defined network |
US10812632B2 (en) | 2015-02-09 | 2020-10-20 | Avago Technologies International Sales Pte. Limited | Network interface controller with integrated network flow processing |
US9660904B2 (en) * | 2015-02-10 | 2017-05-23 | Alcatel Lucent | Method and system for inserting an openflow flow entry into a flow table using openflow protocol |
US9686137B2 (en) * | 2015-02-10 | 2017-06-20 | Alcatel Lucent | Method and system for identifying an outgoing interface using openflow protocol |
US9660903B2 (en) * | 2015-02-10 | 2017-05-23 | Alcatel Lucent | Method and system for inserting an openflow flow entry into a flow table using openflow protocol |
US9596173B2 (en) | 2015-04-09 | 2017-03-14 | Telefonaktiebolaget L M Ericsson (Publ) | Method and system for traffic pattern generation in a software-defined networking (SDN) system |
US9754217B2 (en) * | 2015-05-01 | 2017-09-05 | Cirius Messaging Inc. | Data leak protection system and processing methods thereof |
US10116557B2 (en) | 2015-05-22 | 2018-10-30 | Gray Research LLC | Directional two-dimensional router and interconnection network for field programmable gate arrays, and other circuits and applications of the router and network |
US10089089B2 (en) | 2015-06-03 | 2018-10-02 | The Mathworks, Inc. | Data type reassignment |
US10419242B1 (en) | 2015-06-04 | 2019-09-17 | Netronome Systems, Inc. | Low-level programming language plugin to augment high-level programming language setup of an SDN switch |
US10505819B2 (en) | 2015-06-04 | 2019-12-10 | Cisco Technology, Inc. | Method and apparatus for computing cell density based rareness for use in anomaly detection |
US9891898B1 (en) * | 2015-06-04 | 2018-02-13 | Netronome Systems, Inc. | Low-level programming language plugin to augment high-level programming language setup of an SDN switch |
US20180191640A1 (en) | 2015-06-30 | 2018-07-05 | Hewlett Packard Enterprise Development Lp | Action references |
US9942171B2 (en) | 2015-07-02 | 2018-04-10 | Arista Networks, Inc. | Network data processor having per-input port virtual output queues |
US10367728B2 (en) * | 2015-07-15 | 2019-07-30 | Netsia, Inc. | Methods for forwarding rule hopping based secure communications |
US9654424B2 (en) * | 2015-07-31 | 2017-05-16 | Nicira, Inc. | Managed forwarding element with conjunctive match flow entries |
US9692690B2 (en) * | 2015-08-03 | 2017-06-27 | Telefonaktiebolaget Lm Ericsson (Publ) | Method and system for path monitoring in a software-defined networking (SDN) system |
US20170048144A1 (en) | 2015-08-13 | 2017-02-16 | Futurewei Technologies, Inc. | Congestion Avoidance Traffic Steering (CATS) in Datacenter Networks |
US10496680B2 (en) | 2015-08-17 | 2019-12-03 | Mellanox Technologies Tlv Ltd. | High-performance bloom filter array |
US9826071B2 (en) | 2015-08-26 | 2017-11-21 | Barefoot Networks, Inc. | Configuring a switch for extracting packet header fields |
US9825862B2 (en) | 2015-08-26 | 2017-11-21 | Barefoot Networks, Inc. | Packet header field extraction |
US20170070416A1 (en) | 2015-09-04 | 2017-03-09 | Telefonaktiebolaget L M Ericsson (Publ) | Method and apparatus for modifying forwarding states in a network device of a software defined network |
US10255074B2 (en) | 2015-09-11 | 2019-04-09 | Qualcomm Incorporated | Selective flushing of instructions in an instruction pipeline in a processor back to an execution-resolved target address, in response to a precise interrupt |
US9923816B2 (en) | 2015-09-24 | 2018-03-20 | Barefoot Networks, Inc. | Data-plane stateful processing units in packet processing pipelines |
US9912610B2 (en) | 2015-09-24 | 2018-03-06 | Barefoot Networks, Inc. | Data-plane stateful processing units in packet processing pipelines |
US10523764B2 (en) | 2015-09-24 | 2019-12-31 | Barefoot Networks, Inc. | Data-plane stateful processing units in packet processing pipelines |
US10361899B2 (en) | 2015-09-30 | 2019-07-23 | Nicira, Inc. | Packet processing rule versioning |
US10284389B2 (en) | 2015-10-21 | 2019-05-07 | Arris Enterprises Llc | High availability for distributed network services in an extended bridge |
DE112016004860T5 (en) | 2015-10-21 | 2018-07-26 | Arris International Ip Ltd. | Distributed rule deployment in an extended bridge |
US10282220B2 (en) | 2015-11-05 | 2019-05-07 | Dell Products, L.P. | Dynamic allocation of queue depths for virtual functions in a converged infrastructure |
US20170134282A1 (en) | 2015-11-10 | 2017-05-11 | Ciena Corporation | Per queue per service differentiation for dropping packets in weighted random early detection |
US10291555B2 (en) | 2015-11-17 | 2019-05-14 | Telefonaktiebolaget Lm Ericsson (Publ) | Service based intelligent packet-in buffering mechanism for openflow switches by having variable buffer timeouts |
US9912774B2 (en) | 2015-12-22 | 2018-03-06 | Intel Corporation | Accelerated network packet processing |
US10135734B1 (en) | 2015-12-28 | 2018-11-20 | Amazon Technologies, Inc. | Pipelined evaluations for algorithmic forwarding route lookup |
US10484285B2 (en) | 2015-12-30 | 2019-11-19 | Argela Yazilim ve Bilisim Teknolojileri San. ve Tic. A.S. | System, method and article of manufacture for using control plane for critical data communications in software-defined networks |
EP3400688B1 (en) | 2016-01-04 | 2020-05-20 | Gray Research LLC | Massively parallel computer, accelerated computing clusters, and two dimensional router and interconnection network for field programmable gate arrays, and applications |
CN108702326B (en) | 2016-01-05 | 2021-03-19 | 瑞典爱立信有限公司 | Method, device and non-transitory machine-readable medium for detecting SDN control plane loops |
US10574555B2 (en) | 2016-01-28 | 2020-02-25 | Arista Networks, Inc. | Network data stream tracer |
US10609597B2 (en) | 2016-01-29 | 2020-03-31 | Arista Networks, Inc. | System and method of a pause watchdog |
US10778809B2 (en) | 2016-02-26 | 2020-09-15 | Arista Networks, Inc. | Per-input port, per-control plane network data traffic class control plane policing |
US10250530B2 (en) | 2016-03-08 | 2019-04-02 | Mellanox Technologies Tlv Ltd. | Flexible buffer allocation in a network switch |
US10333848B2 (en) | 2016-07-01 | 2019-06-25 | Intel Corporation | Technologies for adaptive routing using throughput estimation |
US10305805B2 (en) | 2016-07-01 | 2019-05-28 | Intel Corporation | Technologies for adaptive routing using aggregated congestion information |
US10305799B2 (en) | 2016-08-17 | 2019-05-28 | Cisco Technology, Inc. | Re-configurable lookup pipeline architecture for packet forwarding |
US10212065B2 (en) | 2016-10-20 | 2019-02-19 | Gatesair, Inc. | Extended time reference generation |
US20180124183A1 (en) | 2016-11-03 | 2018-05-03 | Futurewei Technologies, Inc. | Method and Apparatus for Stateful Control of Forwarding Elements |
US10341242B2 (en) | 2016-12-13 | 2019-07-02 | Oracle International Corporation | System and method for providing a programmable packet classification framework for use in a network device |
US10091137B2 (en) | 2017-01-30 | 2018-10-02 | Cavium, Inc. | Apparatus and method for scalable and flexible wildcard matching in a network switch |
US11223520B1 (en) | 2017-01-31 | 2022-01-11 | Intel Corporation | Remote control plane directing data plane configurator |
US10412018B1 (en) | 2017-03-21 | 2019-09-10 | Barefoot Networks, Inc. | Hierarchical queue scheduler |
WO2018178906A1 (en) | 2017-03-28 | 2018-10-04 | Marvell World Trade Ltd. | Flexible processor of a port extender device |
US10757028B1 (en) | 2017-04-23 | 2020-08-25 | Barefoot Networks, Inc. | Configurable forwarding element deparser |
JP2020113924A (en) | 2019-01-15 | 2020-07-27 | 富士通株式会社 | Monitoring program, programmable device, and monitoring method |
US10833987B2 (en) | 2019-01-29 | 2020-11-10 | Cisco Technology, Inc. | Supporting asynchronous packet operations in a deterministic network |
US20200280428A1 (en) | 2019-10-18 | 2020-09-03 | Intel Corporation | Configuration scheme for link establishment |
US20200280518A1 (en) | 2020-01-28 | 2020-09-03 | Intel Corporation | Congestion management techniques |
-
2015
- 2015-08-26 US US14/836,850 patent/US9825862B2/en active Active
-
2017
- 2017-10-10 US US15/729,555 patent/US10432527B1/en active Active
-
2019
- 2019-09-17 US US16/573,847 patent/US11411870B2/en active Active
- 2019-11-25 US US16/695,044 patent/US11425038B2/en active Active
- 2019-11-25 US US16/695,049 patent/US11425039B2/en active Active
- 2019-11-25 US US16/695,053 patent/US20200099619A1/en not_active Abandoned
-
2022
- 2022-07-07 US US17/859,722 patent/US20220345405A1/en active Pending
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11388053B2 (en) | 2014-12-27 | 2022-07-12 | Intel Corporation | Programmable protocol parser for NIC classification and queue assignments |
US11394611B2 (en) | 2014-12-27 | 2022-07-19 | Intel Corporation | Programmable protocol parser for NIC classification and queue assignments |
US11394610B2 (en) | 2014-12-27 | 2022-07-19 | Intel Corporation | Programmable protocol parser for NIC classification and queue assignments |
US11425038B2 (en) | 2015-08-26 | 2022-08-23 | Barefoot Networks, Inc. | Packet header field extraction |
US11411870B2 (en) | 2015-08-26 | 2022-08-09 | Barefoot Networks, Inc. | Packet header field extraction |
US11425039B2 (en) | 2015-08-26 | 2022-08-23 | Barefoot Networks, Inc. | Packet header field extraction |
US11677851B2 (en) | 2015-12-22 | 2023-06-13 | Intel Corporation | Accelerated network packet processing |
US11245572B1 (en) | 2017-01-31 | 2022-02-08 | Barefoot Networks, Inc. | Messaging between remote controller and forwarding element |
US11223520B1 (en) | 2017-01-31 | 2022-01-11 | Intel Corporation | Remote control plane directing data plane configurator |
US11463385B2 (en) | 2017-01-31 | 2022-10-04 | Barefoot Networks, Inc. | Messaging between remote controller and forwarding element |
US11606318B2 (en) | 2017-01-31 | 2023-03-14 | Barefoot Networks, Inc. | Messaging between remote controller and forwarding element |
US11425058B2 (en) | 2017-04-23 | 2022-08-23 | Barefoot Networks, Inc. | Generation of descriptive data for packet fields |
US11503141B1 (en) | 2017-07-23 | 2022-11-15 | Barefoot Networks, Inc. | Stateful processing unit with min/max capability |
US11750526B2 (en) | 2017-07-23 | 2023-09-05 | Barefoot Networks, Inc. | Using stateful traffic management data to perform packet processing |
US11362967B2 (en) | 2017-09-28 | 2022-06-14 | Barefoot Networks, Inc. | Expansion of packet data within processing pipeline |
US11700212B2 (en) | 2017-09-28 | 2023-07-11 | Barefoot Networks, Inc. | Expansion of packet data within processing pipeline |
Also Published As
Publication number | Publication date |
---|---|
US20220345405A1 (en) | 2022-10-27 |
US20200099618A1 (en) | 2020-03-26 |
US20200076737A1 (en) | 2020-03-05 |
US11425039B2 (en) | 2022-08-23 |
US11425038B2 (en) | 2022-08-23 |
US20200099617A1 (en) | 2020-03-26 |
US11411870B2 (en) | 2022-08-09 |
US9825862B2 (en) | 2017-11-21 |
US20170063690A1 (en) | 2017-03-02 |
US10432527B1 (en) | 2019-10-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11425039B2 (en) | Packet header field extraction | |
US20230344922A1 (en) | Configuring a switch for extracting packet header fields | |
US10516626B1 (en) | Generating configuration data and API for programming a forwarding element | |
US10805437B2 (en) | Compiler and hardware interactions to remove action dependencies in the data plane of a network forwarding element | |
US20230275835A1 (en) | Apparatus and method of generating lookups and making decisions for packet modifying and forwarding in a software-defined network engine | |
US10694006B1 (en) | Generation of descriptive data for packet fields | |
US9154418B1 (en) | Efficient packet classification in a network device | |
US10454833B1 (en) | Pipeline chaining | |
US11729300B2 (en) | Generating programmatically defined fields of metadata for network packets | |
WO2015006354A1 (en) | Using headerspace analysis to identify flow entry reachability | |
US10645013B2 (en) | Data flow identifiers | |
US10721167B1 (en) | Runtime sharing of unit memories between match tables in a network forwarding element | |
WO2021116770A1 (en) | Hybrid fixed/programmable header parser for network devices | |
US10003676B2 (en) | Method and apparatus for generating parallel lookup requests utilizing a super key | |
US9948756B1 (en) | Automated pipeline generation | |
CN105763296B (en) | Method for scheduling network frames between processing resources |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: FINAL REJECTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: FINAL REJECTION MAILED |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |