TW201929494A - Network communication method, system and controller of PCIe and ethernet hybrid networks - Google Patents

Abstract

A network communication method, a system and a controller of PCIe and Ethernet hybrid networks are provided. In one example of the network communication method, a packet is transmitted or received via a network interface (NIC). The packet is transmitted via an Ethernet over PCIe (EoP) port by adopting a PCIe based EoP architecture based on a PCIe bus for an intra-rack transmission, and the packet is transmitted via an Ethernet port for an inter-rack transmission. If the transmitted packet belongs to a broadcast packet, the network interface receiving the packet discards the packet or transmits the packet to the controller without re-broadcasting the packet again.

Description

Network communication method, system and controller combining fast peripheral component interconnection bus and Ethernet

The disclosure relates to a network communication method, system and controller for combining a fast peripheral component interconnection bus and an Ethernet network.

The Peripheral Component Interconnect Express (PCIe) bus is an industry standard computer extension technology developed by the PCI Special Interest Group (PCI-SIG). PCIe was originally designed as a local bus interconnect technology for connecting CPUs, GPUs, and I/O devices within a machine, and has evolved to become point-to-point (point-to-point) A fully-fledged switched network with features such as links, hop-by-hop flow control, and end-to-end retransmission. PCIe can also be used as an expansion interface for connecting machines to external devices such as storage boxes.

A PCIe network is a switched network with serial point-to-point full duplex lanes. A PCIe device connects to this PCIe network through a link consisting of one or more channels. Recently, an extended PCIe that interconnects multiple servers or virtualized I/O devices using the PCIe interface has become a standard. For example, PCIe applications can be further extended to intra-rack interconnects. PCIe switches can replace standard top of rack (ToR) Ethernet switches, which means that PCIe can connect multiple hosts (such as servers) in the same cabinet. Input and output (I/O) devices connected to the PCIe switch are also allowed to share all servers in the same cabinet. All servers in the cabinet can also communicate with each other through PCIe links.

At present, packet propagation between multiple servers in the same cabinet in a PCIe network cannot be performed across the cabinet, nor does it support Ethernet broadcast behavior.

The disclosure provides a network communication method for combining a fast peripheral component interconnection bus and an Ethernet network, which includes the following steps. Transmitting or receiving packets through a network interface card, wherein if the packets are unicast, the propagation to the same cabinet is propagated via a PCIe based Ethernet over PCIe (EoP) based on a fast peripheral component interconnect bus The packet is transmitted by the Ethernet for transmission to different cabinets. If the packet is broadcast, the packet is discarded, or the packet is transmitted to the controller to respond to the packet.

The present disclosure provides a network communication system that combines a fast peripheral component interconnect bus and an Ethernet network, including a network interface card and a controller. The network interface card is coupled to the controller and configured to perform the following operations: transmitting or receiving a packet, wherein the propagation to the same cabinet is via an EoP-based EoP architecture based on a fast peripheral component interconnection bus. And for the propagation of different cabinets, the packet is transmitted via the Ethernet; and if the packet is broadcasted, the packet is discarded, or the packet is transmitted to the controller in response to the packet.

The present disclosure provides a controller for a network communication system that combines a fast peripheral component interconnect bus and an Ethernet network. This controller includes a memory and a processor. The processor is coupled to the memory for managing propagation between the plurality of hosts in the cabinet and the different cabinets, and the processor is configured to receive the packet, wherein the packet is encapsulated from a network interface card of the external destination host, and according to The rule table determines that the packet does not match all the rules in the rule table, wherein the packet includes the connection information; the connection information is determined according to the connection information, and the information is related to the Ethernet (EoP) or B based on the fast peripheral component interconnection bus. The network generates the judgment information; and generates rules according to the judgment information and the packet and transmits the rules to the external destination host, wherein the external destination host stores the rules in the rule table and responds to the external source host, and the external source host does not re-broadcast the packet. .

Based on the above, hosts in multiple cabinets can use EoP埠 for the propagation of hosts in the cabinet, and Ethernet埠 for the propagation of hosts between the cabinets, while utilizing the maximum bandwidth of the PCIe buss to accelerate the propagation of hosts in the cabinet.

The above described features and advantages of the invention will be apparent from the following description.

Exemplary embodiments of the present disclosure may include any one or more of the novel features described herein (including those described in the Detailed Description) and/or illustrated in the drawings. As used herein, "at least one", "one or more" and "and/or" are used to mean both connected and separated in operation. For example, "at least one of A, B, and C", "at least one of A, B, or C", "one or more of A, B, and C", "A, B, or C" Each of one or more of "and A, B and/or C" means alone A, alone B, alone C, A and B together, A and C together, B and C together or A, B and C together.

It should be noted that the term "a" entity refers to one or more of the entities. Thus, the terms "a", "an" or "an"

The exemplary embodiment of the present disclosure proposes a hybrid network combining a PCIe based Ethernet over PCIe (EoP) technology and an Ethernet (hereinafter referred to as Ethernet) based on a fast peripheral component interconnect bus. The communication system and method enable the EoP technology that originally only supports the transmission of the cabinet package, and also supports the cross-cabinet data transmission after combining the hybrid network communication system and method proposed by the exemplary embodiment of the present disclosure.

In the hybrid network communication system and method proposed by the exemplary embodiments of the present disclosure, at least one or part of the multiple embodiments may include a controller and a network interface card. Card, hereinafter referred to as NIC), this network interface card is built-in to support Ethernet (EoP) and Ethernet (Ethernet) based on fast peripheral component interconnection bus, and set rules on the NIC through the controller (Flow) Rule), so that the packet can be transmitted between machines in different cabinets. Therefore, the above network interface card (NIC) is also a Super Network Interface Card. In addition, this system also provides broadcast services for Ethernet behavior while avoiding broadcast storms. In one embodiment, the controller may be built in a management host of a network system, or in another embodiment, may be built in any host in the same cabinet, depending on the environment to be applied. , not as a limit.

In the fast peripheral component interconnection (PCIe) and Ethernet hybrid network communication system and method proposed by the exemplary embodiments of the present disclosure, the network that originally only supports EoP technology can also be used in the exemplary embodiment disclosed by the present disclosure. In a mixed or compatible Ethernet network environment, packets are transmitted in the same cabinet and in different cabinets.

The strategy for enabling EoP-enabled networks to spread packets in the same cabinet and in different cabinets in a mixed or compatible Ethernet network environment includes, for example, allowing users (network management) to issue rules through the controller. To achieve this goal, the controller first performs a Medium Access Control (MAC) learning behavior, and then the controller issues a rule to the corresponding NIC. After the host (Host) performs data transmission with the cabinet or different cabinets, the data can be transmitted according to the Unicast propagation rules issued by the user (network management). In order to avoid broadcast storms, the hybrid network communication system and method proposed by the exemplary embodiment of the present disclosure also formulates a broadcast rule, so that the NIC drops the packet according to the rule or sends the packet to the host and the packet. After the controller, the broadcast action of the packet is no longer performed.

In the present disclosure, an Ethernet (EoP) based on a fast peripheral component interconnection bus on the fast peripheral component interconnection bus is used to speed up the propagation of the host in the cabinet, and combined with the Ethernet port to perform inter-cabinet propagation. Further, EoP埠 is a PCIe埠 method using a memory mapping method, and can be implemented by a software driver; in another embodiment, the EoP埠 can also be a physical/hardware port.

Exemplary embodiments of the present invention are described in detail below with reference to the drawings.

FIG. 1 is a block diagram illustrating a memory mapping system in accordance with an exemplary embodiment of the present disclosure.

Referring to FIG. 1, a memory mapping system 100 includes a management host 110, a PCIe bus 130, and a plurality of servers 120a through 120n. The servers 120a to 120n may be coupled to the management host 110 via the PCIe bus 130, and the Non-Transparent Bridge (hereinafter referred to as NTB) (ie, 140a-140n) on the PCIe bus 130 performs management. Memory address mapping between host 110 and servers 120a through 120n. Specifically, each NTB 140a-140n has two endpoints, and each endpoint is responsible for memory address mapping in one direction. The management host 110 includes a central processing unit (Central Processing Unit, CPU) 112, a memory 114, a memory mapping controller 116, and a single root I/O virtualization (Single Root I/O Virtualization, hereinafter referred to as SR). -IOV) devices 118a~118n. The memory map controller 116 can control the memory address mapping between the management host 110 and the servers 120a through 120n. SR-IOV devices 118a-118n (e.g., 10G Ethernet cards) can be plugged into management host 110, and the virtual functions of SR-IOV devices 118a-118n can be shared to servers 120a through 120n via PCIe bus 130 .

In an exemplary embodiment, the endpoint of NTB 140a may map the address space of management host 110 to server 120a via PCIe bus 130, while the other endpoint of NTB 140a may place the server via PCIe bus 130 The address space of 120a is mapped to the management host 110. In this method, access to the memory of the management host 110 can be directed to the mapped memory address of the server 120a, and access to the memory of the server 120a can be directed to the mapped map of the management host 110. Memory address.

In various embodiments, the memory map controller 116 can be implemented as a software module or code. For example, when a new server is detected on the PCIe bus, the code can be loaded into the memory 114 and executed by the CPU 112 to update the mapping information between the management host 110 and the server. . However, the disclosure is not limited thereto. In some implementations, memory mapping controller 116 may also be implemented as a hardware circuit that controls memory address mapping between management host 110 and servers 120a through 120n.

2 is a flow chart illustrating a memory mapping method in accordance with an exemplary embodiment of the present disclosure.

Referring to FIG. 2, in step S201, the management host having the memory address configures the extended memory address.

In step S203, the NTB of the PCIe bus bar maps the extended memory addresses corresponding to the management host of each server to the memory addresses of each of the servers, respectively.

In step S205, each server configures an extended memory address.

In step S207, the NTB maps the extended memory address of each server to the memory address and the extended memory address of the management host, wherein the extended memory address of each server corresponds to the server and the management host.

As described above, the memory mapping method constructs a global memory mapping structure in the server and the management host, so that the server can share the virtual functions of the SR-IOV device through the global memory address mapping using the BAR of each NTB and mutual Communication while taking advantage of the maximum bandwidth of the PCIe bus. Therefore, the present disclosure implements EoP埠 by the memory mapping method, thereby increasing the propagation bandwidth of the data and reducing the propagation delay. In other words, the disclosure may combine PCIe埠 with the management host into an EoP埠 in the form of a software module or a code.

3 is a block diagram illustrating a network communication system within a cabinet in accordance with an exemplary embodiment of the present disclosure.

Referring to FIG. 3, network system 300 includes a controller 310, a plurality of hosts 320a through 320n, a PCIe switch 330, and an Ethernet switch 340. Each of the hosts 320a to 320n includes Ethernet埠Eth and EoP埠, and Ethernet埠Eth and EoP埠 are coupled to the PCIe switch 330. The host 320a to 320n may be a server or a virtual server, etc., and the EoP may be a connection or a physical port of the memory mapping system implemented by the software as described in FIG. 1 and FIG. It is worth noting that Ethernet 埠Eth and EoP埠 can be combined into a network interface card (NIC) by software module or code. Based on this, the NIC can be a virtual switch, a driver module, or other various virtual modules, etc., and the NIC can have at least one rule table to transmit or receive packets according to the rules of the rules table. The controller 310 is coupled to the NICs of the plurality of hosts 320a to 320n and the NICs of the hosts of different cabinets, thereby monitoring the hosts in the cabinet and outside the cabinet in real time. Specifically, a management network is formed between the controller 310 and the host of each cabinet, and the network is completely independent of the network between the hosts. It should be noted that the controller 310 may be disposed in the cabinet, in the host, or outside the cabinet, but the disclosure is not limited thereto. The PCIe switch 330 is coupled to the Ethernet switch 340 for handling propagation in the cabinet and can transmit or receive packets to and from different cabinets via the Ethernet switch 340. In various embodiments, the PCIe switch 330 may be disposed in the cabinet, in the host, or outside the cabinet, but the disclosure is not limited thereto.

4 is a block diagram illustrating a network communication system between cabinets in accordance with an exemplary embodiment of the present disclosure.

Referring to FIG. 4, the network system 400 includes a controller 410, cabinets 420a through 420n, and an Ethernet switch 430, and the controller 410 and the Ethernet switch 430 are independent of the cabinets 420a through 420n, and each of the cabinets 420a through 420n has at least one host. The controller 410 is coupled to the cabinets 420a through 420n to monitor and manage the hosts within each cabinet. In other words, controller 410 and cabinets 420a through 420n have formed a management network. The cabinets 420a through 420n are coupled to the Ethernet switch 430 to perform propagation between the cabinets. In other words, when hosts in different cabinets transmit data to each other, data is transmitted or received through the Ethernet switch 430, and the Ethernet switch 430 and the cabinets 420a to 420n form a data network independent of the management network.

Referring to Figures 3 and 4, the NICs of the hosts of the cabinets 420a through 420n can transmit or receive packets, wherein the propagation of the same cabinets is propagated via the EoP(R) PCIe-based EoP architecture, and for different cabinets via Ethernet. Eth spreads the packet. If the packet is broadcast, the packet is discarded, or the packet is transmitted to the controller.

FIG. 5 is a block diagram illustrating a controller in accordance with an exemplary embodiment of the present disclosure.

Referring to FIG. 5, the controller 500 includes a processor 510, a memory 520, and an input/output interface 530. The processor 510 can be coupled to the memory 520 and the input/output interface 530, and the processor 510 pre-configures the EoP and the Ethernet port to correspond to the first port and the hosts in the same cabinet in the memory 520. In other words, the memory 520 stores a cabinet correspondence table and a connection port correspondence table, wherein the connection port corresponding table has a correspondence between each host port and EoP埠 and Ethernet埠, and the cabinet correspondence table records each host setting. Which cabinet. Whenever the connection between each host and the relationship between EoP and Ethernet is changed, the processor 510 modifies the connection table in real time. The processor 510 also modifies the cabinet correspondence table in real time whenever the host is added or removed from the cabinet. When the controller 500 receives the packet via the output interface 530, the processor 510 analyzes the packet to obtain a Media Access Control (MAC) address and port information, wherein the output interface 530 can be a wired network interface. The card or the wireless transceiver interface, and the connection information can be the host receiving the packet. In addition, the processor 510 reads the connection/correspondence table of the memory 520, and determines the connection information according to the connection/correspondence table to be related to the Ethernet埠 or EoP埠 to generate the judgment information, and then controls the address and the judgment information through the media access control. Generate a rule to further pass back this rule. For example, when the connection information is the 15th pin of the host, the 15th pin is determined as EoP埠 according to the connection table, and the processor 510 sets the EoP埠 as the judgment information.

Further, when the external source host broadcasts the packet to the external destination host, the processor 510 can receive the packet via the output interface 530, wherein the packet is from the NIC of the external destination host, and the NIC determines the packet mismatch rule table according to the rule table. All rules within, including the MAC address of the source host and the connection information. Then, the processor 510 determines, according to the connection information, the connection information related to the EoP or Ethernet of the NIC of the external destination host to generate the judgment information, and then generates another rule according to the judgment information and the packet and passes through the output interface 530. Transferred to an external destination host, where the external destination host stores another rule in the rules table and responds to the external source host, while the external source host no longer rebroadcasts the packet.

FIG. 6 is a schematic diagram illustrating a rules table in accordance with an exemplary embodiment of the present disclosure.

Referring to Figure 6, the rules table 600 includes at least one rule, and each rule has at least a host MAC address field 610a through 610n and a corresponding port field 620a through 620n. In addition, each rule may have a priority field. Specifically, when each host wants to transmit a packet, the rule table is searched according to the MAC address of the destination host to find a corresponding connection field, and the packet is transmitted to the destination host through the corresponding connection port, wherein the connection is The 埠 field is the judgment information generated by the controller. If the MAC address of the destination host is found to correspond to two rules, the priority fields of the two rules are compared to select a higher priority rule, and the connection field can store Ethernet埠 or EoP埠. When each host receives the packet, it searches the rule table according to the MAC address of the source host to determine whether the corresponding rule already exists. If the corresponding rule already exists and the packet is broadcast broadcast, the packet will be discarded or returned to the controller according to the rules. If there is no corresponding rule, the connection information of the receiving packet is added to the packet for transmission to the controller, wherein the connection information may be the pin of the host receiving the packet.

FIG. 7 is a flowchart illustrating a transmitting end of a unicast transmission according to an exemplary embodiment of the present disclosure.

Referring to FIG. 7, in step S701, the source host compares the MAC address of the packet of the destination host with the rule table. In other words, the MAC address of the destination host in the packet is directly raised, and the MAC address is compared with all the rules of the rule table.

In step S703, a connection port corresponding to the destination host is obtained according to the rule matching the packet. Specifically, when the MAC address of the destination host in the packet matches the rule of the rule table, the connection field of the rule is read to obtain a connection port corresponding to the MAC address of the destination host, where the connection may be It is an EoP埠 or Ethernet埠. In multiple implementations, if the MAC address of the destination host matches the two rules, the port corresponding to the MAC address of the destination host is determined according to the priority field in the rule.

In step S705, the packet is transmitted by the corresponding port. Specifically, if the corresponding port is EoP, the packet will be transmitted to the destination host in the same cabinet by EoP. If the corresponding port is Ethernet, the packet will be sent to the destination host in different cabinets by Ethernet.

FIG. 8 is a flowchart illustrating a receiving end of a unicast according to an exemplary embodiment of the present disclosure.

Referring to FIG. 8, in step S801, the destination host determines whether there is a rule for matching the packet according to the rule table and the packet. Specifically, the destination host reads the MAC address in the packet and reads the MAC address field of the rule of the rule table to match whether the MAC address in the packet matches.

In step S803, if there is no match, the destination host adds the MAC address of the source host receiving the packet and the port information of the port to the packet for transmission to the controller. Specifically, the destination host's NIC determines the pin of the received packet to generate the port information with the pin information, and adds the port information and the MAC address of the source host to the packet for transmission to the controller.

In step S805, the controller generates and returns a rule to the destination host. Specifically, the controller determines, according to the connection/correspondence table, that the connection information belongs to EoP埠 or Ethernet埠 to generate judgment information (at least information with EoP埠 or Ethernet埠), and uses the MAC address and judgment information of the source host of the packet. Generate a rule and pass it back to the destination host.

In step S807, the destination host stores the rule in the rule table. Further, the NIC of the destination host has a preset rule table, and the NIC adds the rule to the rule table whenever the NIC receives the new rule.

In step S809, the unicast propagation of the destination is ended.

9 is a flow diagram illustrating a destination host for broadcast propagation in accordance with an exemplary embodiment of the present disclosure.

Referring to FIG. 9, in step S901, the destination host receives the broadcasted packet. In other words, at least one destination host receives an Address Resolution Protocol (ARP) packet broadcast by the source host.

In step S903, the destination host determines, according to the rule table, whether there is a rule for matching the packet. Specifically, the destination host will propose a MAC address in the packet to compare the MAC address with the rules in the rule table to determine whether the matching is completed.

In step S905, if the matching is completed, the packet is discarded.

In step S907, if the matching is not completed, the MAC address of the source host that receives the packet and the port information of the port are added to the packet for transmission to the controller. Specifically, the destination host's NIC determines the pin of the received packet to generate the port information with the pin information, and adds the port information and the MAC address of the source host to the packet for transmission to the controller.

In step S909, the controller generates and returns a rule to the destination host. Specifically, the controller determines, according to the connection/correspondence table, that the connection information belongs to EoP埠 or Ethernet埠 to generate the judgment information, and generates a rule by using the MAC address of the source host of the packet and the judgment information, and then returns the rule to the destination. Host.

In step S911, the destination host stores the rule to return the response packet to the source host according to the rule. Specifically, the NIC of the destination host stores the rule in the rule table, and generates a response packet according to the rule and the MAC address of the destination host, and the destination host selects the connection according to the connection bar of the corresponding rule to return the response packet. . It should be noted that if the two rules of the rule table have the same MAC address, the rule that connects the EoP is set to a higher priority, and the destination host preferentially transmits the response packet according to the connection with higher priority. . If the connection with higher priority is damaged, the destination host will transmit the response packet by the lower priority connection.

In step S913, the destination host will end the address resolution.

FIG. 10 is a flowchart illustrating a source host of broadcast propagation in accordance with an exemplary embodiment of the present disclosure.

Referring to FIG. 10, in step S1001, in order to complete the address resolution, the source host broadcasts the packet. Specifically, when the source host is in an initial state, the controller will transmit a broadcast rule to the source host to generate a rules table with broadcast rules. The source host will generate an ARP packet according to the rule table and the MAC address of the source host to transmit the packet from EoP埠 and Ethernet埠. Therefore, the destination host in the same cabinet will receive packets from EoP埠 and Ethernet埠, and the destination hosts in different cabinets will only receive packets from Ethernet. The destination host discards or transmits the packet to the controller according to its NIC rule table.

In step S1003, the source host receives the response packet of the destination host. Specifically, the source host will receive response packets from the destination host of the same cabinet from the EoP port, and receive response packets from the destination host of the different cabinets from the Ethernet port.

In step S1005, the source host adds the MAC address of the destination host that receives the response packet and the port information of the port to the response packet for transmission to the controller. Specifically, the source host's NIC will detect the connection to receive the response packet to generate the connection information, and add the connection information and the destination host's MAC address to the response packet for transmission to the controller.

In step S1007, the controller generates and returns a rule. Specifically, the controller determines, according to the connection/correspondence table, the connection of the response packet, the information belongs to EoP埠 or Ethernet埠 to generate the judgment information, and returns the message according to the MAC address of the destination host of the response packet and the judgment information generation rule. Source host.

In step S1009, the source host stores the rules to the rules table. Specifically, the NIC of the source host receives the rule to store the rule in the rule table, thereby completing the address resolution.

The practical application of the exemplary embodiments of the present disclosure will be described below with reference to FIGS. 11 and 12, wherein the same element symbols are used in the drawings and the description to represent the same or similar parts. 11 is a schematic diagram illustrating a network communication system within a cabinet in accordance with an exemplary embodiment of the present disclosure. FIG. 12 is a schematic diagram illustrating a network communication system between cabinets according to an exemplary embodiment of the present disclosure.

Referring to Figure 11, the network system 1100 in the cabinet can have a plurality of hosts 1110a through 1110d, a controller controller, and a PCIe switch, wherein each of the hosts 1110a through 1110d has a NIC. When the propagation within the cabinet is required, the packet can be transmitted by the EoP port of the host's NIC and transmitted to the EoP port of the NIC of the other host via the PCIe switch. For example, when the host 1110a transmits a packet to the host 1110b, the host 1110c or the host 1110d, the packet can be transmitted by the EoP port of the NIC of the host 1110a, and the packet is received by the host 1110b, the host 1110c, or the EoP port of the NIC of the host 1110d.

Referring to Figure 12, the inter-cabinet network system 1200 can have multiple network systems within the cabinet as well as PCIe switches. Referring to Figures 11 and 12, for example, hosts 1210a through 1210c belong to the same cabinet, while hosts 1220a through 1220c belong to another cabinet. When the inter-rack propagation is required, the packet can be transmitted by the Ethernet NIC of the host NIC and transmitted to the Ethernet port of the NIC of the host of the other cabinet via the Ethernet switch. For example, when the host 1210a transmits the packet to the host 1220a, the host 1220b or the host 1220c of another cabinet, the packet can be transmitted by the Ethernet port of the NIC of the host 1210a, and the Ethernet of the NIC of the host 1220a, the host 1220b or the host 1220c is used.埠 Receive packets. In addition, the host's NIC determines whether the rules related to the received packet have been stored, whether in the cabinet or between the cabinets. If it is determined that there is no rule associated with the received packet, the NIC will inform the controller to generate a new rule and store the new rule in the NIC.

FIG. 13 is a flowchart illustrating a method of network communication combining PCIe and Ethernet according to an exemplary embodiment of the present disclosure.

Referring to FIG. 13, in step S1301, a packet is transmitted or received through a network interface card, wherein the propagation for the same cabinet propagates the packet via the EoP, and the propagation for the different cabinet propagates the packet via the Ethernet.

In step S1303, if the packet is broadcast broadcast, the packet is discarded, or the packet is transmitted to the controller in response to the packet.

In summary, the network communication method disclosed in the present invention constructs a propagation structure combining the fast peripheral component interconnection bus and the Ethernet between the cabinets or the hosts between the cabinets, so that the hosts in the cabinet are adopted between the hosts. EoP埠 communicate with each other, and the Ethernet between the cabinets uses an Ethernet port. The disclosure can integrate the above-mentioned universal energy into one network card, and utilizes the maximum bandwidth of the PCIe bus, and does not need to set an extra network card to accelerate the transmission between the cabinets. The previously described exemplary implementation of the present disclosure has the aforementioned advantages, wherein the foregoing advantages are not required in all forms of the present disclosure.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention, and any one of ordinary skill in the art can make some changes and refinements without departing from the spirit and scope of the present invention. The scope of the invention is defined by the scope of the appended claims.

100‧‧‧Memory mapping system

110‧‧‧Management host

112‧‧‧CPU

114‧‧‧ memory

116‧‧‧Memory Mapping Controller

118a~118n‧‧‧SR-IOV device

120a~120n‧‧‧Server

130‧‧‧PCIe bus

140a~140n‧‧‧Non-transparent bridge (NTB)

S201~S207‧‧‧Steps

300‧‧‧Network System

310‧‧‧ Controller

320a~320n‧‧‧Host

330‧‧‧PCIe exchanger

340‧‧‧Ethernet switch

400‧‧‧Network System

410‧‧‧ Controller

420a~420n‧‧‧ cabinet

430‧‧‧Ethernet switch

500‧‧‧ controller

510‧‧‧ processor

520‧‧‧ memory

530‧‧‧Import interface

600‧‧‧Rules

610a~610n‧‧‧MAC address field

620a~620n‧‧‧Connected field

S701~S705‧‧‧Steps

S801~S809‧‧‧Steps

S901~S913‧‧‧Steps

S1001~S1009‧‧‧Steps

1100‧‧‧Network System

1110a~1110d‧‧‧Host

1200‧‧‧Network System

1210a~1210c‧‧‧Host

1220a~1220c‧‧‧Host

S1301~S1303‧‧‧Steps

FIG. 1 is a block diagram illustrating a memory mapping system in accordance with an exemplary embodiment of the present disclosure. 2 is a flow chart illustrating a memory mapping method in accordance with an exemplary embodiment of the present disclosure. 3 is a block diagram illustrating a network communication system within a cabinet in accordance with an exemplary embodiment of the present disclosure. 4 is a block diagram illustrating a network communication system between cabinets in accordance with an exemplary embodiment of the present disclosure. FIG. 5 is a block diagram illustrating a controller in accordance with an exemplary embodiment of the present disclosure. FIG. 6 is a schematic diagram illustrating a rules table in accordance with an exemplary embodiment of the present disclosure. FIG. 7 is a flowchart illustrating a transmitting end of a unicast according to an exemplary embodiment of the present disclosure. FIG. 8 is a flowchart illustrating a receiving end of a unicast according to an exemplary embodiment of the present disclosure. 9 is a flow diagram illustrating a destination host for broadcast propagation in accordance with an exemplary embodiment of the present disclosure. 10 is a flow diagram illustrating a source host for broadcast point propagation in accordance with an exemplary embodiment of the present disclosure. 11 is a schematic diagram illustrating a network communication system within a cabinet in accordance with an exemplary embodiment of the present disclosure. FIG. 12 is a schematic diagram illustrating a network communication system between cabinets according to an exemplary embodiment of the present disclosure. FIG. 13 is a flowchart illustrating a method of network communication combining PCIe and Ethernet according to an exemplary embodiment of the present disclosure.

Claims (13)

A network communication method combining a fast peripheral component interconnection bus and an Ethernet network, comprising: transmitting or receiving a packet through a network interface card, wherein the communication for the same cabinet is via a quick peripheral component-based interconnection bus The Ethernet (EoP) uses the EoP architecture to propagate the packet, and the propagation of the different cabinets is propagated via an Ethernet network; and if the packet is broadcast, the packet is discarded or the packet is transmitted. To a controller to return the packet. The network communication method according to claim 1, wherein if the packet is broadcast, the packet is discarded, or the packet is transmitted to the controller to respond to the packet, including: not re-broadcasting the packet Decapsulating; and discarding or transmitting the packet according to a rule table of the network interface card. The network communication method of claim 2, wherein the step of storing the rule table on the network interface card and discarding or transmitting the packet according to the rule table comprises: determining, by using the network interface card Whether the packet matches the rule table; and if the packet matches the rule table, the packet is discarded, and if the packet does not match the rule table, the media access control address of the source host and a link information are added to the packet. Transmitting to the controller, wherein the controller controls the destination host according to the packet to respond to the source host. The network communication method of claim 3, wherein the controlling the host according to the packet to return the source host according to the packet comprises: determining the connection information according to a connection/correspondence table of the controller Corresponding to the EoP or the Ethernet network to generate a judgment information, and storing the judgment information in the controller; and the media access control address and the connection according to the judgment information and the source host The information generates a rule and transmits it to the destination host, wherein the destination host stores the rule in the rule table to respond to the source host according to the rule. A network communication system combining a fast peripheral component interconnection bus and an Ethernet network, comprising: a controller; and a network interface card coupled to the controller and configured to perform the following operations: Receiving a packet in which the propagation to the same cabinet is transmitted via an EoP (EoP) based on a fast peripheral component interconnection bus, and the packet is transmitted through the EoP architecture for the different cabinets. Propagating the packet; and if the packet is broadcast, discard the packet, or transmit the packet to a controller to respond to the packet. The network communication system of claim 5, wherein the network interface card is one of a virtual switch, a driver module or a virtual module. The network communication system of claim 5, wherein if the packet is broadcast, the packet is discarded, or the packet is transmitted to the controller to respond to the packet: the packet is not re-broadcasted And discarding or transmitting the packet to the controller according to a rule table of the network interface card. The network communication system of claim 7, wherein the network interface card determines whether the packet matches the rule table, and if the packet matches the rule table, the network interface card discards the packet, and if the packet is discarded, The packet does not match the rule table, and the network interface card adds the media access control address of the source host and a link information to the controller, where the controller controls the destination host according to the packet. The source should be back to the source. The network communication system of claim 8, wherein the controller determines, according to a connection/correspondence table of the controller, that the connection information is related to the EoP port or the Ethernet port to generate a Determining the information, and storing the judgment information to generate a rule according to the judgment information and the media access control address of the source host and the connection information, and transmitting the rule to the destination host, wherein the destination host stores the rule in the rule The table responds to the source host according to the rule. The network communication system of claim 9, wherein the rule has at least a media access control address of the source host and the determination information. A controller for a network communication system combining a fast peripheral component interconnection bus and an Ethernet network, comprising: a memory; and a processor coupled to the memory for managing the same cabinet Propagating between a plurality of hosts in the cabinet, the processor is configured to: receive a packet, wherein the packet is from a network interface card of an external destination host, and the network interface card determines the rule according to a rule table The packet does not match all the rules in the rule table, wherein the packet includes a media access control address and an interface information of an external source host; and the connection information is determined according to a connection/relationship table. a component interconnecting bus (EoP) or an Ethernet network to generate a judgment message; and generating a rule with the packet according to the judgment information and transmitting the rule to the external destination host, wherein the external The destination host stores the rule in the rule table and responds to the external source host, and the external source host does not rebroadcast the packet. The controller of claim 11, wherein the network interface card discards or transmits the packet according to the rule table. The controller of claim 12, wherein the rule has at least a media access control address of the external source host and the determination information.