Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04Q—SELECTING
  • H04Q11/00—Selecting arrangements for multiplex systems
  • H04Q11/04—Selecting arrangements for multiplex systems for time-division multiplexing
  • H04Q11/0428—Integrated services digital network, i.e. systems for transmission of different types of digitised signals, e.g. speech, data, telecentral, television signals
  • H04Q11/0478—Provisions for broadband connections
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L12/00—Data switching networks
  • H04L12/54—Store-and-forward switching systems 
  • H04L12/56—Packet switching systems
  • H04L12/5601—Transfer mode dependent, e.g. ATM
  • H04L2012/5614—User Network Interface
  • H04L2012/5618—Bridges, gateways [GW] or interworking units [IWU]
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L12/00—Data switching networks
  • H04L12/54—Store-and-forward switching systems 
  • H04L12/56—Packet switching systems
  • H04L12/5601—Transfer mode dependent, e.g. ATM
  • H04L2012/5638—Services, e.g. multimedia, GOS, QOS
  • H04L2012/5646—Cell characteristics, e.g. loss, delay, jitter, sequence integrity
  • H04L2012/5651—Priority, marking, classes
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L12/00—Data switching networks
  • H04L12/54—Store-and-forward switching systems 
  • H04L12/56—Packet switching systems
  • H04L12/5601—Transfer mode dependent, e.g. ATM
  • H04L2012/5638—Services, e.g. multimedia, GOS, QOS
  • H04L2012/5665—Interaction of ATM with other protocols
  • H04L2012/5667—IP over ATM

Definitions

• microfiche appendix containing, inter alia, a listing of pseudo-code for implementing certain aspects of the invention.
• the microfiche appendix consists of 14 microfiche and a total of 828 frames.
• the microfiche appendix also includes copies of certain documents, including draft versions of standards, which, although well-known in the art at the present time, are subject to ongoing change and revision. These documents are included to capture the state of these evolving standards at this point in time.
• the invention relates generally to the field of interconnection of networks, and, more particularly, to subscriber end point equipment capable of selectively allocating bandwidth on demand to communications originating from a network or sub-network at a subscriber's premises for transmission across another network, such as a wide area network.
• Local area networks are well known in the art in which a plurality of subscriber terminals or workstations are interconnected over a network.
• local area networks are confined to a collection of devices that are located in reasonably proximity to each other.
• Wide area networks are similarly known in which stations which are relatively widely separated in geography can be interconnected.
• a local area network is connected to a wide area network at a node of the wide area network which is commonly referred to as an edge-switch.
• Permanent Virtual Circuits are communication paths that are set up in advance and remain established so that they are available for use at any time by a subscriber desiring communications.
• a Switched Virtual Circuit (SVC), on the other hand, is selectively established at the beginning of a communications session and torn down at the end of the session.
• a Permanent Virtual Circuit has the advantage that the signalling overhead associated with establishment and tearing down of a connection between end points is not needed since the PVC is always available.
• Some types of networks deliver packets of a communications session only on a "best efforts" basis. That means that no special precautions are taken to ensure a given Quality Of Service (QOS) in terms of network metrics like bandwidth available and end to end delay.
• QOS Quality Of Service
• Modern wide area networks utilize ATM and frame relay switches and protocols. Other types of switches and protocols are known in the art for wide area networks. A number of protocols are also commonly used for local networks. Increasingly, local area network protocols utilize TCP/IP (Transmission Control Protocols/Internet Protocols) for communications. This is particularly convenient because TCP/IP is the communications standard for the Internet.
• TCP/IP Transmission Control Protocols/Internet Protocols
• a number of problems result from attempting to interconnect a local area network with another local area network over a wide area network. Differences in protocol between that utilized for the local area network (e.g., TCP/IP) and that use for the wide area network (e.g., ATM) need to be accommodated. Certain types of local area networks utilize signalling, such as out of band signalling, to establish connections among users where as other networks do not. Typically, local area networks using TCP/IP do not use signalling. However, the network to which the local area networks connect may in fact be signaled.
• the present invention provides apparatus, methods, systems, techniques and computer program products which overcome the problems of the prior art and provide the advantages identified as desirable above.
• the invention permits flexible interconnection between local and wide area networks by detecting a quality of service (QOS) needed for an application invoking a communication session and by creating or selecting a virtual circuit having that QOS. Singnalled and unsignalled networks can be accommodated.
• QOS quality of service
• Singnalled and unsignalled networks can be accommodated.
• Figure 1 illustrates a network arrangement for interconnecting LANs over a WAN in accordance with one aspect of the invention.
• FIG 2 is a description of the hardware architecture of subscriber and point equipment 220, shown in Figure 1, in accordance with one aspect of the invention
• Figure 3 is a network diagram illustrating standard IP PVC routing as used in the prior art.
• FIG. 4 is a network diagram illustrating set-up of a simple switched virtual circuit (SVC) across an ATM network as shown in the prior art.
• SVC simple switched virtual circuit
• Figure 5 is a block diagram showing the software architecture of the service end point equipment 220.
• Figure 6 is a block diagram of software used in the fast routing layer
• Figure 7 is a flow chart used to explain operation of the fast routing layer.
• FIG. 8 is a network diagram illustrating set-up of a switched virtual circuit (SVC) across a wide area network in accordance with one aspect of the invention.
• SVC switched virtual circuit
• Figure 9 is a block diagram illustrating basic LAN- WAN Bandwidth on Demand features with an unsignaled LAN.
• Figure 10 is a block diagram illustrating basic LAN- WAN Bandwidth on Demand features when a signalled LAN is utilized.
• Figure 11 is a flow chart of a process for detecting applications bandwidth requirements.
• Figure 12 is a block diagram used to illustrate detection of a mulitple flows for a connection.
• Figure 13 is a data structure used to represent information about a flow.
• Figure 14 is an exemplary drop down menu for selecting quality of service for a particular connection.
• Figure 15 is a block diagram showing selection from among existing PVCs of differing quality levels when establishing a connection across a network.
• Figure 16 is a block diagram showing selection from among existing types of connections (TOC) when establishing a connection across a network.
• TOC existing types of connections
• Figure 17 is a block diagram illustrating reservation of bandwidth.
• Figure 18 is a pseudocode representation of an exemplary technique for evaluating new reservation requests.
• Figure 19 illustrates exemplary pseudo data structures used with the fast routing layer.
• FIG 20 PCT is a pseudocode representation for the upper IP drivers 206 shown in figure 5.
• Figure 21 PCT is a pseudocode representation of implementation of the fast routing layer 205 shown in figure 5.
• FIG 22 PCT is a pseudocode representation for the lower IP drivers 204 shown in figure 5.
• Figure 23 PCT is a pseudocode representation for the rules engine 301 shown in figure 5.
• Figure 24 is a pseudocode representation for the flow group monitor 306 shown in figure 5.
• FIG. 1 illustrates a network arrangement for interconnecting LANs over a WAN in accordance with one aspect of the invention.
• Wide area network 200 serves to interconnect a plurality of local area networks 110A and HO C.
• the interface between the local area networks 110 and the WAN 160 is subscriber end point equipment (SEE) 120.
• SEE 120 connects to the WAN 160 through an edge computer which constitutes the entry point into the WAN 160.
• An internet service provider, 130 is also shown connected to the WAN.
• each LAN will be presumed to utilize a TCP/IP communications protocol and, for purposes of illustration, arranged in a 10 base T configuration for communications.
• each SEE 120 contains memory and routing tables such that a virtual private network can be defined in the memory of SEE 220 devices.
• FIG. 2 is a description of the hardware architecture of subscriber end point equipment 120, shown in Figure 1, in accordance with one aspect of the invention.
• the main controller for the hardware shown in Figure 2 is CPU 114. It is preferably a Motorola MPC860SAR processing element.
• the CPU 114 has a plurality of input-output ports which are utilized as shown in Figure 2.
• Nonvolatile memory 116 Serial EEPROM
• LED indicators 111 permit certain output information concerning the state of the device to be displayed to an external user.
• RS 232 serial console port 113 connects to the CPU over line SMC1. This permits a PC class device to be connected to the SEE 120 to serve as a console for troubleshooting and other purposes.
• the CPU 114 has integral protocol processing capabilities for Ethernet, HDLC, and ATM/SAR.
• Network input/output is provided by items 101, 102, 103, 104, 105, 106, and 107 which include Ethernet, ATM, POTS, and frame relay capable bi-directional ports.
• a programmable logic device 112 is provided to support multiple Utopia capable I/O devices to the CPU 114. This allocates the Utopia bus as needed to the connected network interfaces.
• a routing field programmable gate array 110 provides data path routing to signals between the main CPU/Protocol processor 114 signals and the V.35 port 104 and the second Ethernet port 105.
• This routing FPGA also routes TDM data between the CPU 114, DSPs (e.g. 2 DSPs) 117, Tl/El ports 106 and POTS port 107.
• the POTS port 107 includes necessary analog to digital conversion circuitry for interfacing standard POTS telephone connections to the networks.
• the DSPs 117 provide high computation capabilities for any signals delivered to them via the TDM bus as routed by FPGA 110. Examples of the types of computing utilizing the DSPs include voice or data compression, DTMF tone detection and generation. Two separate content addressable memories (CAM) hang off the memory bus 250. CAM1 (108) is used for layer 3 fast look ups and CAM2 (109) is used for layer 4 fast look ups.
• CAM1 (108) is used for layer 3 fast look ups
• CAM2 (109) is used for layer 4 fast look ups.
• FIG. 3 is a network diagram illustrating standard IP PVC routing as used in the prior art.
• Network 300 is an IP network having nodes 310 which each connect to one or more sub-networks 110 which are to be linked over network 300. The simple illustration shown, with only three nodes, each node is linked to all the other nodes over respective permanent virtual circuits. Thus, node 310a is linked to node 310c over permanent virtual circuit PVC 1. Node 310a is linked to node 310b over PVC 2 and node 310b linked to node 310c over PVC 3.
• a routing table and a linked table are maintained at each node so that packets from a user such as WS 1 can be directed over sub network 110 to the appropriate permanent virtual circuit to reach a node associated with a sub network of a destination user, such as WS 129 as shown in the illustration.
• the routing table contains entries showing each destination net (e.g., sub network) and the "next hop" to be taken to get to that destination network.
• the link table identifies in one field each destination node and in a related field the particular permanent virtual circuit to be taken to reach that node.
• a packet from user WS1 is routed over sub network 110a to node 310a (IP 1) where the routing table is referenced to determine that the next hop should be to IP3 in order to reach destination net 210c on which WS29 is located, to which the packet is to be directed.
• IP 1 node 310a
• the link table is referenced to determine that the link to node IP3 should be over PVC 1.
• FIG. 4 illustrates a set-up of a simple switch virtual circuit (SVC) across an ATM network as shown in the prior art.
• IP1, IP2, and IP3 in this instance are gateways between a TCP/IP network 110a and 110c, respectively in an ATM network 400.
• Gateway IP1 connects to edge switch 410a over a well known address and port.
• IP1 initiates a request for a connection to be established to work station 129 of sub network 110c.
• the edge switch has an ATM address table as shown which relates the various destination nodes to their respective ATM addresses.
• Gateway IP1 requests then a virtual circuit be established for use by its user WS1 for communicating with user WS129.
• the routing table identifies the net portion of the address WS129 and identifies that it is served by gateway IP3. However, there is no virtual circuit identified at this time in the link table.
• Gateway IP1 requests the establishment of an SVC for use during this communication from edge switch 410a.
• Edge switch 410a has the ATM address for gateway IP3 which is serviced by edge switch 410b. Accordingly, edge switch 410a requests the establishment of a virtual circuit between itself and 410b for use during these communications.
• the edge switches establish the virtual circuit and make the assignment to the gateway, notifying it of the virtual circuit to be utilized for the communications requested. At this time, the identity of the virtual circuit is inserted into the link table for reference and routing. With the tables populated, communications may begin between WSl and WS129.
• Figure 5 is a block diagram showing the software architecture of the service end point equipment 120. A rough comparison to the ISO/OSI model will be useful in explaining the software architecture.
• Block 217 corresponds roughly to layer 1 or the physical layer of the ISO/OSI model.
• Blocks 201, 202, 203 and 204 correspond generally to layer 2 of the ISO/OSI model (the link layer) and block 205, 206 and 209 correspond generally to layer 3 of the ISO OSI model (the network layer).
• the group of functions labeled SmartAgent handle the top level management of the software.
• the Web/SNMP/XML controls 210 permit the overall box configuration to be managed.
• Initial loading of the executable image for program memory or network is handled by the bootloader 211.
• Routing (both basic and items like routing table and parametrics exchange) is handled by Routing Protocols 213.
• the Network I/O ports in Figure 2 (201, 202, 203, 204, 205, 206, 207) are logically presented in Figure 3 as a group 217.
• the I/O configuration manager 216 configures these ports according to management commands from the SmartAgent Controls 210.
• the result of this configuration is the activation of one or more ports towards the enterprise side and one or more ports towards the service provider.
• This configuration information is then made available to the Fast Routing Layer (FRL) 205.
• FTL Fast Routing Layer
• the FRL 505 is the central software entity for data plane activities in the SEE 120. It performs the functions necessary to move data at maximum bandwidth when the data packet is for an existing flow.
• a typical configuration would have the Ethernet port 517 connected to an Ethernet driver 510 as the enterprise side and a Tl port in 517 connected to an ATM driver 503.
• a packet arrives at 501 and is passed up to 504 after it is confirmed as valid. 504 would strip any Layer 2 information/headers (the 14 byte MAC header in this case) and send the IP packet up to the FRL 205.
• the FRL 205 would see if this is an existing flow (i.e., does it match source and destination for the IP/Layer 3 information, the source and destination TCP/UDP port, and protocol type) and if it does, immediately sends it to appropriate transmit point in 204. After encapsulating the IP packets (e.g., RFC 1483) it would be transmitted via the ATM driver 203 out the configured service provider port in 217.
• IP packets e.g., RFC 1483
• the first time a packet travels between two new layer 3 entities i.e., IP addresses
• the final entity in Figure 5 is the Queue Manager 515. To the Fast Routing Layer 505 it appears as another formatting module in 504. The Queue Manager, however, holds and reorders the transmission of packets according to priorities given it by the Controls 210. When a packet is ready for actual transmission, the appropriate entry point in the formatting module 204 is called. This allows for a 'shaping' (e.g., prioritization) of packets traveling on the same virtual circuit.
• a 'shaping' e.g., prioritization
• Figure 6 is a block diagram of software used in the fast routing layer. It will be used to describe the operation of the fast routing layer in more detail in conjunction with the flow chart of figure 7..
• a flow match 307 is performed by taking the source and destination IP address from the packet and comparing it to all known pairs via CAM 1 108. If there is no match, the packet is sent 'up' the stack to 206. If there is a match, then the match number from CAM 1 108, type of packet, source port number and destination port number are used to try for a match on CAM 2 109. If there is no match, then the packet is sent to the rules engine 301.
• Packets that have matched both CAMs are existing flows and can be completely handled by FRL 205.
• the match information gives an index into the flow context table 312. This table contains the information for the remaining processing that needs to take place.
• block 308 takes the TOS value in 312 and applies to the packet.
• block 309 decrements the TTL and recalculates the IP header checksum (including TOS change if applicable.)
• the transmit bytes and transmit packets counters are then updated by 310.
• block 311 using the transmit function pointer, transmit arg, and the new packet (via the information in 312), calls the proper entry point in 204 to send the packet.
• the flow info 312 also contains a flag for control flows which need to be sent to the rules engine 301 so that control flows can be continously monitored.
• Packets that were sent 'up' the IP stack will come back to the FRL 205 through the IP adaptation layer 206 after a next hop address has been determined by the IP layer 209. These packets are then processed by the rules engine 301. As noted above, the rules engine is also called if CAM 2 109 has no match (a layer 4 miss.)
• the rules engine 301 must make a determination as to whether this is an ordinary/low priority application or a special/high priority application. This is done by comparing both static and dynamic TCP/UDP port information (i.e. Layer 4.) Static rules are constant port number assignments that exist for many applications. Dynamic rules are determined by monitoring the packets in the control channel flows for each of the different applications 303 and determining what the applications have negotiated for dynamic port numbers.
• the rules engine 301 determined a new VC is required (i.e. this is a new application session with special handling needs) by checking the application policy database 315. This activity is run by the Flow Group Monitor 306. First, if the policy indicated multiple selection option, the Customer Contact Manager 218 would be directed to select from these options. At this point, a new flow group 313 would be created, flow group activites in 304, 305, 306 would be started and the new VC creation would be started.
• the rules engine 301 would tell block 214 to initiate SVC signalling.
• an entry into the system flow table in 312 would be added, indicating either transmit onto a default channel or queue packets as a starting state.
• this flow table entry 312 is updated to reflect the new SVC information.
• the flow management monitor also 208 creates the information necessary for billing and Service Level Agreement (SLA) monitoring based on and contained in 313.
• Billing information includes start time, end time, application type, source IP, destination IP, statistics (transmitted bytes and packets, received bytes and packets), and QOS related parameters.
• the flow monitor 208 When the primary control channel monitoring by 301 and 303 indicates the session is ending, the flow monitor 208 will tell 214 to terminate (teardown) the VC. Next it will gather all of the statistics and SLA information from 312, 313, and 305 and a session or call detail record will be sent to the billing host by 304.
• the final role for the flow group monitor 306 is to perform periodic (minute granularity) housekeeping on the flows that are in the default group (for timeouts) so that billing can also be performed on this class of information. This way, applications in 315 can be marked to go onto the default channel but still be billed, while the bulk data (non- application specific) traffic information is also periodically reported.
• FIG. 8 is a network diagram illustrating set-up of a switched virtual circuit (SVC) across a wide area network in accordance with one aspect of the invention.
• SVC switched virtual circuit
• Figure 8 is similar to the drawing similar to figure 4.
• the gateway of Figure 4 lack the ability to set up a switched virtual circuit with the quality of service specified by the gateway.
• the SEE 120 detects the quality of service required for a connection and establishes a switched virtual circuit having that quality of service across the WAN to the destination.
• each new connection utilizing the services of the WAN gets a level of service commensurate with either the application utilizing the communication services, or a level of service specified by a user as described more and hereinafter.
• FIG 9 is a block diagram illustrating basic LAN- WAN Bandwidth on Demand features with an unsignaled LAN.
• a packet arrives at LAN (typically Ethernet) port 1.
• LAN typically Ethernet
• a packet may go to either: a) up through to local IP stack, b) directly to the rules engine 4.
• Packets for a) are sent 'up' the IP stack for other local clients and require no additional special handling.
• Packets which are destined for the WAN side (as determined by 2) but which are not part of an existing Layer 3 (IP) connection need to be routed first by 7 (i.e. a next hop IP address is required) which is part of the system's IP stack.
• Packets for b) go to the flow rules engine 4.
• the packets are first encapsulated in block 3 (e.g. via an RFC 1483 standard header addition) and then sent along to be transmitted out the proper VC by 6 and then out the WAN port 8.
• the rules engine 4 finds this is a new flow, a determination is made as to whether this is an ordinary/low priority application or a special/high priority application. This is done in block 4 according to a rules comparison engine based on both static and dynamic TCP/UDP port information (i.e. Layer 4) and IP addresses (i.e. Layer 3.) Static rules are constant port number assignments that exist for many applications. Dynamic rules are determined by the system in block 4 monitoring the packets in the control channel flows for each of the different applications and determining what the applications have negotiated for dynamic port numbers. If the rules engine 4 determined a new VC is required (i.e. this is a new application session with special handling needs), the system would initiate SVC signalling out the WAN port via block 5.
• a rules comparison engine based on both static and dynamic TCP/UDP port information (i.e. Layer 4) and IP addresses (i.e. Layer 3.) Static rules are constant port number assignments that exist for many applications. Dynamic rules are determined by the system in block 4 monitoring the packets in the
• This SVC to flow information provides the basis for block 6 to send packets on the VC in 8 associated with a given flow.
• Figure 10 is a block diagram illustrating basic LAN- WAN Bandwidth on Demand features when a signalled LAN is utilized.
• Packet arrives at LAN (typically Ethernet) port 11. Packet may go to one of several places: a) proxy for LAN signalling 15, b) up through to local IP stack, c) directly routed to appropriate output port based on the determination made by block 12.
• Packets for a) go to the local LAN signalling agent 15. Upon arrival at 15 several actions would take place. First, the appropriate response to the signalling would be sent back out the LAN port. Second, if the signalling indicated a new flow was going to be taking place, the system would determine is a new WAN SVC was appropriate or if the new flow should go onto an existing VC in which case the flow information would be added in 14 for this.
• the system would initiate SVC signalling out the WAN port via block 16. Next, an entry into the system flow table in 14 would be added. When the WAN signalling completed this flow table entry would be updated by 16 to reflect the new SVC information.
• This SVC to flow information provides the basis for block 18 to send packets on the VC in 19 associated with a given flow.
• Packets for b) are sent 'up' the IP stack for local clients and require no additional special handling. Packets which are destined for the WAN side (as determined by 12) but which are not part of an existing Layer 3 (IP) connection need to be routed first by 17 (i.e. a next hop IP address is required) which is part of the system's IP stack. Packets for c) (i.e. they are part of a known Layer 3/IP connection) are first encapsulated in block 13 (e.g. an RFC 1483 standard header addition) and then sent along to be transmitted out the proper VC by 18 and then out the WAN port 19.
• block 13 e.g. an RFC 1483 standard header addition
• FIG 11 is a flowchart of a process for establishing SVC's with selectable QOS.
• a check is made at the SEE 120 to determine whether an existing virtual circuit exists between the SEE 120 and the desired destination (1110). If such a virtual circuit exists, (1110-Y) the packet flow will be directed to the existing virtual circuit for transmission to the destination (1120). If there is no existing virtual circuit (1110-N), the flow is routed to a queue (1130) and simultaneously to an application specific analyzer. If the connection is a multi-flow connection, such as found in ITU standard H.323, one will add table entries as additional flows are detected that are associated with the H.323 connection (1140).
• Figure 12 is a block diagram used to illustrate detection of a multiple flows for a connection.
• a workstation WS 1 desires to communicate in a mode which requires multiple flows, such as when device WS1 desires to communicate using the ITU H.323 standards (multimedia applications)
• a connection is requested to a destination device addressed to a particular well known port which is utilized for establishment of connection with H.323 devices.
• an initial flow is established carrying a request for a connection.
• the request specifies a connection going to port 86, in this example, and results in an assignment received from the connected device of a second port number, such as 122 in the example shown over which control information regarding the set-up of data flows needed for the H.323 connection can be transmitted.
• one or more data flows such as 193 shown can be established to handle respective streams or flows of multimedia information.
• the packet flow is teed off the normal routing path to permit an application specific analyzer to monitor the flows at the same time packet information is flowing across the communications connection.
• the plurality of flows associated with a service such as the H.323 service constitute part of a flow group.
• the disconnecting device will send a tear-down message across control channel 2 (port 122 in this example), directing tear-down of the SVC used to handle the flow group.
• Figure 13 is a data structure used to represent information about a flow.
• the data structure also includes SVC information, such as port number, virtual path indicator (VPI) and virtual connection indicator (VCI) and an encapsulation header.
• VPI virtual path indicator
• VCI virtual connection indicator
• Figure 14 is an exemplary drop down menu for selecting quality of service for a particular connection.
• a drop down menu can also be utilzied to specify a quality of service for a particular application type.
• This particular drop down menu provides the user with the opportunity to select one of several options as to network quality and as to network security.
• This particular form of drop down menu has additional advantages in that it provides the opportunity for advertising to be displayed to a user as part of the network connection quality selection process.
• the look up time and resources required to access routing information increases dynamically. If the look up is done in the software, as conventionally the case, the amount of time required for a particular routing process is a function of the number of flows underway.
• the SEE 120 utilizes content addressable memories (CAM) for routing determinations.
• CAM content addressable memories
• one content addressable memory is utilized for level 3 routing and one is utilized for level 4 routing. This particular configuration is flexible and provides a richer set of capabilities than would otherwise be the case.
• Content addressable memories have the advantage that their time per calculation is deterministic.
• Figure 15 is a block diagram showing selection from among existing PVCs of differing quality levels when establishing a connection across a network.
• the SEE 120 can also be utilized to select among a plurality of existing PVCs, each having a different QOS.
• multiple PVCs In certain Internet applications, for example, when schemes like Diffserv are utilized, it is possible for multiple PVCs to be established at different levels. In the case of Diffserve, for example, multiple levels (e.g. bronze, silver, gold and platinum) service are defined. If PVCs are established across a network at different levels of service, the level of bronze might be utilized for "best efforts" communications and the others dedicated to particular QOS specific connections.
• levels e.g. bronze, silver, gold and platinum
• the user has no say about changing the quality of service because the establishment of the connection is associated relatively permanently with either the user or the application. Therefore, the user may be paying for more capability than needed for a particular session or may be unable to guarantee the level of service needed for a particular communications.
• the selection capabilities of the SEE 120 the user can specify, on a connection by connection basis, or by class of application, the quality of service needed for a particular session.
• the SEE 120 can also be used in a topology like figure 15 to select the PVC based on the "type of service" (TOS) bits in the IP header of incoming packets, resulting in QOS across the network for appropriately marked packets.
• TOS type of service
• Figure 16 illustrates a network arrangement for interconnecting LAN over a WAN.
• the QOS for a given packet is based on the TOS bits settings in the IP header of the packet.
• the SEE 120 can be utilized to set the TOS bits appropriate for the particular session or appropriate for the user selection, so that the WAN network will transport these packets with the QOS desired.
• bandwidth reservation system involves a user 'talking' to a network box to reserve a certain amount and quality of bandwidth at a time in the future. For instance, a user wants to have a teleconference across the WAN to a remote office 'next Thursday from 10:00AM to 11 :30AM'.
• the teleconference software requires 384Kbit/second with less than 30 milliseconds of latency.
• FIG 17 is a block diagram illustrating a Bandwidth Reservation System (BRS).
• the reservation request is received by the reservation server and is first checked for a valid format. Bad format results in rejected request.
• a valid request includes sufficient information to completely quantify the network based resources required and the time, date and duration of the requirement.
• the request is then compared for validity against the basic capabilities of the affected network elements (i.e. is it within the basic bandwidth, processing resources, etc. that the transport equipment can support). Inadequate capabilities result in a failed reservation request.
• the request enters the conflict/priority resolution process. This process will determine whether the required resources can be reserved according to initiator's criteria. For any of the resources being requested, the time, date and duration of the reservation request is compared against a database of existing reservations for network resources.
• the new reservation function is initiated . If there is a conflicting resource reservation, then a check of the user priority in the policy database is performed. If the user making the new request has been defined to have a higher priority than the holder of the existing reservation, the lower priority reservation is released and the new reservation function is initiated. Notification of released reservation to its holder also takes place. If a pre-existing reservation holder has higher priority than the new request, the new request returns a "failed" indication. The new reservation function commits a transaction with the reservation database resulting in a new reservation for specified time, date duration, and resources.
• the BRS includes a user policy database. This database would include a field for every system user with an eight bit reservation priority value. A value of zero would indicate the lowest priority for reservations and a value of 255 would indicate the highest reservation priority
• the BRS also includes a reservation database.
• This database would include, for . each network resource which can potentially be reserved, a variable number of data structures describing reservations for the resource. Each such data structure would include the following information:
• Figure 18 is a pseudocode representation of an exemplary implementation for evaluating new reservation requests.
• Pseudo code and Pseudo data structures and standards utilized for implementing the SEE 120 are set forth in detail in the attached microfiche appendix and are incorporated in their entirety into the disclosure hereof by reference.
• Figure 19 illustrates exemplary pseudo data structures used with the fast routing layer.
• FIG 20 PCT is a pseudocode representation for the upper IP drivers 206 shown in figure 5.
• Figure 21 PCT is a pseudocode representation of implementation of the fast routing layer 205 shown in figure 5.
• FIG 22 PCT is a pseudocode representation for the lower IP drivers 204 shown in figure 5.
• Figure 23 PCT is a pseudocode representation for the rules engine 301 shown in figure 5.
• Figure 24 is a pseudocode representation for the flow group monitor 306 shown in figure 5.

Landscapes

• Engineering & Computer Science (AREA)
• Computer Networks & Wireless Communication (AREA)
• Data Exchanges In Wide-Area Networks (AREA)

Priority Applications (3)

Applications Claiming Priority (2)

Publications (2)

Family

ID=22290982

Family Applications (1)

Country Status (4)

Cited By (4)

Citations (2)

• 1999
  • 1999-10-01 JP JP2000573044A patent/JP2002526979A/ja active Pending
  • 1999-10-01 AU AU65044/99A patent/AU6504499A/en not_active Abandoned
  • 1999-10-01 WO PCT/US1999/022773 patent/WO2000019665A1/en not_active Application Discontinuation
  • 1999-10-01 EP EP99953006A patent/EP1118186A1/en not_active Withdrawn

Patent Citations (2)

Also Published As

Similar Documents

Legal Events