WO2016000358A1

WO2016000358A1 - Outer label encoding method, traffic congestion control method and device

Info

Publication number: WO2016000358A1
Application number: PCT/CN2014/089340
Authority: WO
Inventors: 赵玉海
Original assignee: 中兴通讯股份有限公司
Priority date: 2014-06-30
Filing date: 2014-10-23
Publication date: 2016-01-07
Also published as: CN105282029B; CN105282029A

Abstract

An outer label encoding method, traffic congestion control method and device, the outer label encoding method comprising: a PE head node encodes an EXP field of the outer label of a packet according to the class of service of an ingress node PHB of the packet and the packet color information, the EXP field comprising a first field for identifying the class of service of the packet and a second field for identifying whether the packet is a virtual private network service packet within a committed bandwidth when entering a pseudo wire of the PE head node, such that a P node can map PHB according to the EXP field, manage and control congestion of each queue in each tunnel, and conditionally perform two-level QOS scheduling management, thus solving the problem in the relevant art of being unable to achieve end-to-end QoS control due to the failure of a P node to obtain VPN information in the case of multiple VPN service multiplexing a tunnel in an E-LSP based MPLS network.

Description

Outer label encoding method, flow congestion control method and device

Technical field

The present invention relates to the field of outer label coding technology and traffic congestion control technology, and in particular to an outer label coding method, a traffic congestion control method and apparatus.

Background technique

In a Multi-Protocol Label Switching (MPLS) network using an EXP-Label Switch Path (E-LSP), if there is a multi-virtual private network (Virtual Private Network) , referred to as VPN) multiplexing tunnel, the related model can not guarantee the end-to-end bandwidth requirements of the VPN.

For example, in the networking shown in Figure 1, the actual physical link is a Gigabit Ethernet (GE) port. On the first node of the carrier edge router (PE), after the quality of service (QoS) scheduling, the virtual link (Pseudo Wire, PW) 1 with lsp1 actual traffic of 800M is doing its best. (Best-effort, abbreviated as BE) service flow cir=600M, eir=200M (where the agreed information rate, committed bandwidth, Committed Information Rate, referred to as CIR; Excess Information Rate, referred to as EIR), lsp2 complex The traffic of the two VPNs is used, including the BE service flow with the pw2 commitment bandwidth cir=100M, including the pw3 committed bandwidth cir=200M, and the eir=800M Assured Forwarding (AF) 1 service flow. In the P node (that is, the intermediate node except the PE head node and the PE tail node, in the MPLS backbone network, the P node after the initial PE only reads the information of the outer label to determine the next hop, so the backbone network In the simple label switching, lsp1 and lsp2 in Figure 1 will cause congestion on the outbound interface of the P node.

According to the QoS scheduling rules, the tunnel will give priority to the cir part. In this way, lsp1 will preferentially obtain the bandwidth of cir=600M, and lsp2 will preferentially obtain the bandwidth of cir=300M. The sum of the cirs of the two tunnels is 600M+300M=900M, and the link bandwidth has 1G–900M=100M allocated lsp1 and lsp2 according to the scheduling weights. Assuming a weight of 1:1, lsp1 and lsp2 each share a bandwidth of eir=50M. Therefore, the actual forwarding traffic of lsp1 is 600M+50M=650M, and the actual forwarding traffic of lsp2 is 300M+50M=350M. However, the specific VPN cannot be identified during P-node tunnel scheduling. Information, therefore, lsp2 will allocate 350M of bandwidth to the high-priority service according to the Priority Queuing (PQ) relationship, that is, ensure the forwarding of the AF1 service. In fact, all of them are allocated to the VPN service carried by pw3. As a result, VPN traffic carried by pw2 will not be guaranteed. It can be seen that in the case of multi-VPN service multiplexing tunnel, end-to-end QoS cannot be achieved because VPN information is not transmitted to the P node.

In the MPLS network based on the E-LSP related to the related technology, in the case of multi-VPN service multiplexing tunnel, the end-to-end QoS control cannot be realized due to the P-node unable to resolve the VPN information, and no effective solution has been proposed yet. ,

Summary of the invention

The present invention provides an outer label encoding method, a traffic congestion control method and apparatus to solve at least the above problems.

In order to solve the above technical problems, the following technical solutions are adopted:

An outer label encoding method comprising:

The operator's edge router, the PE head node, encodes the EXP field of the outer label of the packet according to the service type of the PHB and the packet color information after the traffic congestion control is performed.

The EXP field includes: a first field that is used to identify the service type of the packet, and a virtual field in the virtual link that is used to identify the packet in the virtual link that enters the PE head node. The second field of the private network service message.

Optionally, before the PE head node encodes the EXP field, the method further includes:

The PE head node performs traffic congestion control according to the PHB of the ingress node;

The step of the PE head node performing traffic congestion control according to the ingress node PHB includes:

The PE head node performs first-level scheduling control according to the PHB of the ingress node, and performs at least port scheduling, tunnel scheduling, virtual link scheduling, and flow level scheduling in the first-level scheduling control process, where the port scheduling is performed. To ensure that the traffic of the multiple ports is not interfered with each other, the tunnel scheduling is used to ensure the traffic control of multiple tunnels in the port, and the virtual link scheduling is used to ensure the traffic control of multiple virtual links in the tunnel. Flow-level scheduling is used to ensure different service types in the virtual link carried by the virtual link. The scheduling forwarding requirements to be performed by the business.

Optionally, the step of the PE head node performing traffic congestion control according to the PHB of the ingress node further includes:

The PE head node performs the second-level scheduling control according to the PHB of the ingress node during the packet header encapsulation process of the packet, and preferentially forwards the packet with the higher forwarding priority of the service type in the second-level scheduling control process.

Optionally, after the step that the PE head node encodes an EXP field of an outer label of the packet, the method further includes:

The intermediate node maps the PHB according to the EXP field of the received message, where the first field is mapped to a service type of the PHB, and the second field is mapped to message color information, where In the second field, the packets in the committed bandwidth are mapped green, and the packets outside the committed bandwidth are mapped to yellow.

The intermediate node performs the first-level scheduling control and the second-level scheduling control of the traffic according to the mapped PHB, where the first-level scheduling control is used to control the bandwidth, and the second-level scheduling control is used to control the preferential forwarding of the packet.

Optionally, the step of the intermediate node performing the first level scheduling control of the traffic according to the mapped PHB includes:

The intermediate node equates the received packet with an ordered aggregate OA, and adds the received packet to the same ordered scheduling queue according to the mapped packet color information.

The intermediate node preferentially discards the packet whose color information is yellow in the ordered scheduling queue when congestion control is performed on the traffic.

Optionally, the step of the intermediate node preferentially discarding the packet whose color information is yellow in the ordered scheduling queue includes:

The intermediate node preferentially discards the packet whose color information is yellow in the ordered scheduling queue according to the queue congestion discarding policy of the weighted random pre-detection.

Optionally, the step of the intermediate node performing the second-level scheduling control of the traffic according to the mapped PHB includes:

The intermediate node encodes an EXP field of an outer label of the received packet according to the service type and the packet color information of the PHB of the packet, where the EXP field includes: a first field of the service type of the message, and a second field for identifying the color information of the message of the message;

The intermediate node performs forwarding priority control on the packet according to the service type of the packet mapping.

Optionally, after the intermediate node performs the first-level scheduling control and the second-level scheduling control of the traffic according to the mapped PHB, the method further includes:

The PE tail node maps the PHB according to the inner label of the received packet;

The PE tail node performs HQoS traffic congestion control on the packet on the user-side AC interface according to the mapped PHB.

Optionally, the PE head node or the intermediate node is coded or mapped according to the following rules:

The "0" in the second field indicates that the packet is a virtual private network service packet outside the committed bandwidth or the packet color information indicating the packet is yellow in the virtual link that enters the first node of the PE. The value of "1" in the second field indicates that the packet is a virtual private network service packet in the promised bandwidth or the packet color information of the packet is green in the virtual link that enters the first node of the PE. .

An outer label encoding device is located in a first node of a carrier edge router PE, and includes an EXP encoding module, where:

The EXP encoding module is configured to: encode an EXP field of an outer label of the packet according to a service type of the PHB and a packet color information after the traffic congestion control is performed, and the The EXP field includes: a first field for identifying a service type of the packet, and a virtual private network service report for identifying whether the packet is a committed bandwidth in a virtual link entering the PE head node. The second field of the text.

Optionally, the device further includes a congestion control module, wherein:

The congestion control module is configured to: perform traffic congestion control according to the ingress node PHB; the congestion control module includes a first level scheduling control unit, where:

The first-level scheduling control unit is configured to: perform first-level scheduling control according to the ingress node PHB, and perform at least port scheduling, tunnel scheduling, virtual link scheduling, and flow-level scheduling in the first-level scheduling control process, where The port scheduling is used to ensure that traffic of multiple ports does not interfere with each other. The tunnel scheduling is used to ensure traffic control of multiple tunnels in a port, and the virtual link scheduling is used to ensure multiple virtual links. The traffic control of the virtual link is used to ensure the scheduling and forwarding requirements of services of different service types in the tunnel carried by the virtual link.

Optionally, the congestion control module further includes a second level scheduling control unit, where:

The second-level scheduling control unit is configured to perform second-level scheduling control according to the ingress node PHB during the packet header encapsulation process of the packet, and preferentially forward the forwarding priority of the service type in the second-level scheduling control process. Higher message.

A traffic congestion control device is located in an intermediate node, and includes a PHB mapping module and a congestion control module for each jump release, wherein:

The PB mapping module is configured to: map the PHB according to the EXP field of the packet received by the intermediate node, where the EXP field includes: a service type used to identify the packet. a first field, and a second field used to identify the virtual private network service packet in the committed bandwidth in the virtual link of the first node of the operator edge router PE;

The per-hop is a PHB mapping module that maps the first field to a PHB service type, and the second field is mapped to packet color information, and when the second field is mapped, the committed bandwidth is The packets in the packet are mapped in green, and the packets outside the committed bandwidth are mapped to yellow.

The congestion control module is configured to perform two-level scheduling control of traffic according to the mapped PHB, where the first-level scheduling control is used to control bandwidth, and the second-level scheduling control is used to control priority forwarding of packets.

Optionally, the congestion control module includes a first-level scheduling control queue control unit and a first-level scheduling control packet discarding unit, where:

The first level scheduling control queue control unit is configured to: equate the received message with the intermediate node to an ordered aggregate OA, and add the received message to the same ordered scheduling queue according to the mapped message color information. ;

The first-level scheduling control packet discarding unit is configured to: in the case of performing congestion control on the traffic, preferentially discarding the packet whose color information is yellow in the ordered scheduling queue.

Optionally, the first-level scheduling control packet discarding unit is configured to preferentially discard the packet whose color information is yellow in the ordered scheduling queue according to the following manner:

According to the queue congestion discarding policy of the weighted random pre-detection, the packet with the yellow color of the packet in the ordered scheduling queue is discarded.

Optionally, the congestion control module further includes a second level scheduling control EXP coding unit and a second level scheduling control message forwarding unit, where:

The second level scheduling control EXP encoding unit is configured to: encode an EXP field of an outer label of the packet according to the service type and the packet color information of the PHB of the packet mapping, where the EXP field includes: a first field of the service type of the message, and a second field for identifying the message color information of the message;

The second-level scheduling control packet forwarding unit is configured to: perform forwarding priority control on the packet according to the service type of the packet mapping.

A single board includes an EXP encoding chip and a buffer, wherein:

The EXP encoding chip is configured to: encode an EXP field of an outer label of the packet according to a service type of the PHB and a packet color information after the traffic congestion control is performed, wherein the The EXP field includes: a first field for identifying a service type of the packet, and a virtual private network for identifying whether the packet is a committed bandwidth in a virtual link entering a head node of the operator edge router PE The second field of the service message;

The buffer is coupled to the EXP encoding chip and configured to: buffer a message encoded by the EXP encoding chip.

Optionally, the scheduler is coupled to the buffer, configured to perform a first-level scheduling control process according to the PHB of the ingress node, and perform at least port scheduling, tunnel scheduling, and virtual in the first-level scheduling control process. Link scheduling and flow level scheduling, wherein the port scheduling is used to ensure that traffic of multiple ports does not interfere with each other, and the tunnel scheduling is used to ensure traffic control of multiple tunnels in a port, and the virtual link scheduling The flow rate control is used to ensure the traffic forwarding of multiple virtual links in the virtual link. The flow level scheduling is used to ensure the scheduling and forwarding requirements of services of different service types in the tunnel carried by the virtual link.

Optionally, the scheduler is further configured to: during the packet header encapsulation process of the packet, preferentially forward the packet with a higher forwarding priority of the service type in the second-level scheduling control process according to the PHB of the ingress node. .

A board, including a processor and a scheduler, where:

The processor is configured to: according to the EXP field mapping of the packet received by the intermediate node, the per-jump is issued as a PHB, where the EXP field includes: a first field for identifying a service type of the packet, And a second field for identifying a virtual private network service message in the committed bandwidth in the virtual link of the first node of the operator edge router PE;

The processor maps the first field to a service type of the PHB, and maps the second field to packet color information. When mapping the second field, the packet in the committed bandwidth is mapped. Green, packets outside the committed bandwidth are mapped to yellow;

The scheduler, coupled to the processor, is configured to: perform two-level scheduling control of traffic according to the mapped PHB, where the first level scheduling control is used to control bandwidth, and the second level scheduling control is used to Controls the priority forwarding of packets.

A computer program comprising program instructions that, when executed by a PE head node, an intermediate node, and a tail node, cause the PE head node, the intermediate node, and the tail node to perform the outer label encoding method described above.

A carrier carrying the above computer program.

The foregoing technical solution uses the PE head node to encode the EXP field of the outer label of the packet according to the service type and the packet color information of the ingress PHB of the packet, where the EXP field includes the first service type for identifying the packet. The field is used to identify whether the packet is the second field of the virtual private network service packet in the committed bandwidth in the virtual link that enters the PE head node; the PE first node solves the traffic congestion control according to the ingress node PHB. The related art is based on the E-LSP MPLS network. In the case of multi-VPN service multiplexing tunnel, the end-to-end QoS control cannot be realized because the P-node cannot resolve the VPN information, and the multi-VPN service multiplexing tunnel is guaranteed. End-to-end QoS control.

BRIEF abstract

The drawings described herein are intended to provide a further understanding of the invention, and are intended to be a part of the invention. In the drawing:

1 is a schematic flowchart of multiplexing a P node congestion bandwidth by an L2VPN tunnel according to the related art;

2 is a schematic flow chart of a traffic congestion control method according to an embodiment of the present invention;

3 is a schematic structural diagram of a traffic congestion control apparatus according to an embodiment of the present invention;

4 is a schematic structural diagram of another traffic congestion control apparatus according to an embodiment of the present invention;

5 is a schematic diagram of a PE head node HQoS scheduling model according to an alternative embodiment of the present invention;

6 is a schematic diagram of an EGR chip side scheduling and forwarding of an NP chip according to an alternative embodiment of the present invention;

7 is a schematic diagram of a P-node tunnel HQoS scheduling model according to an alternative embodiment of the present invention;

8 is a schematic diagram of a PE tail node HQoS scheduling model in accordance with an alternative embodiment of the present invention.

Preferred embodiment of the invention

It should be noted that the embodiments in the present application and the features in the embodiments may be combined with each other without conflict. The invention will be described in detail below with reference to the drawings in conjunction with the embodiments.

The steps illustrated in the flowchart of the figures may be performed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, may differ from this The steps shown are performed in the order shown or described.

This embodiment provides a traffic congestion control method, which can be applied to an E-LSP based MPLS network. 2 is a schematic flowchart of a traffic congestion control method according to an embodiment of the present invention. As shown in FIG. 2, the process includes the following steps:

Step S202, the PE head node performs traffic congestion control according to the ingress node PHB.

Step S204: The PE head node encodes an EXP field of an outer label of the packet according to the service type and the packet color information of the inbound PHB of the packet, where the EXP field includes a first field of the COS for identifying the packet and It is used to identify whether the packet is the second field of the virtual private network service packet in the CIR in the virtual link that enters the head node of the PE.

Through the above steps, whether the packet is encoded in the EXP field of the outer label in the virtual link of the CIR in the virtual link that enters the head node of the PE, so that the P node can be based on the EXP field in the outer label. The QoS control is implemented to solve the related technology. In the MPLS network based on the E-LSP, in the case of multiplexing the multi-VPN service tunnel, the P-node cannot resolve the VPN information because there is no VPN information in the outer label, and thus the end-to-end cannot be guaranteed. The implementation of the end QoS guarantees the end-to-end QoS control of the multi-VPN service multiplexing tunnel.

Optionally, in step S202, the PE head node performs HQoS control on the packet traffic. In the first-level scheduling control process, at least port scheduling is performed to ensure that the traffic of the multiple ports does not interfere with each other, and the tunnel scheduling is ensured in the port. Traffic control of multiple tunnels, performing virtual link scheduling to ensure traffic control of multiple virtual links in the tunnel, and performing flow-level scheduling to ensure scheduling and forwarding requirements for services of different service types in the virtual link. Through the first level of scheduling control, the outbound node of the PE can be guaranteed Total traffic bandwidth on the node side, and try to ensure the committed bandwidth of each virtual link on the ingress side.

Optionally, in the process of performing the second-level scheduling control on the PE head node, the PE first node preferentially forwards the packet with the higher forwarding priority of the service type according to the PHB of the ingress node in the process of encapsulating the packet header. . The forwarding priorities of the BE, AF, EF, and CS services are increased in turn, that is, the forwarding priority of the BE service is the lowest, and the forwarding priority of the CS service is the highest. The second-level scheduling control ensures that services with higher forwarding priorities are preferentially forwarded and delays are reduced.

Optionally, in the case of performing priority forwarding according to the packet forwarding priority, the priority queue scheduling mode is adopted.

After the PE head node forwards the packet according to the packet forwarding priority, the packet is sent to the P node for processing.

Optionally, the process of processing the packet by the P node includes: the P node mapping the PHB according to the EXP field of the received packet, where the first field is mapped to the service type of the PHB, and the second field is mapped to the report. Text color information, wherein the mapping rule of the second field is: the second field is identified as the message within the committed bandwidth is mapped green, and the second field is identified as the message outside the committed bandwidth is mapped to yellow; The node performs two-level scheduling control of the traffic according to the mapped PHB. The first-level scheduling control is used to control the bandwidth, and the second-level scheduling control is used to control the priority forwarding of the packet. In addition, the service type of the PHB to which the first field is mapped may be different from the service type of the PHB of the inbound node of the packet in the E-LSP. The EXP field is 3 bits in the E-LSP, and the second field occupies at least 1 bit. The first field can only occupy 2 bits at most, and can represent four service types. Therefore, the eight service types of BE, AF1, AF2, AF3, AF4, EF, CS6, and CS7 can be based on different forwarding priority requirements. Mapping to the four types of services, for example, remapping the BE and AF1 to the BE service in the PE head node, and identifying the packet as the BE service in the first field; in the subsequent processing of the P node, according to The first field treats the message as a BE message.

Optionally, the foregoing first-level scheduling control includes: the P-node adds the received packet according to the mapped packet color information, and is equal to an ordered aggregate (OA), and joins the same ordered scheduling queue; When the traffic is congested, the node preferentially discards the packets whose color information is yellow in the ordered scheduling queue. In this mode, the packets received by the P-node are added to the same ordered scheduling queue, and the bandwidth control is ensured on the premise that the green packets are transmitted as much as possible.

Optionally, the P node preferentially discards according to the queue congestion discarding policy of the weighted random prior detection. The packet whose color information is yellow in the ordered scheduling queue.

Optionally, in the foregoing second-level scheduling control, the P-node encodes an EXP field of an outer label of the packet according to the service type and the packet color information of the PHB of the packet mapping, where the EXP field includes: a first field of the service type of the identifier packet, and a second field of the packet color information for identifying the packet, where the first field is encoded according to the PHB and the second field is equal to the second field of the EXP received by the uplink. , equivalent to transparent transmission. . In the EXP encoding process of the P node, the result of the EXP field encoding is consistent with the EXP field in the PE head node. In this way, the color information of the message is transmitted to the next P node or the PE tail node.

Optionally, after the P node performs two-level scheduling control of the traffic according to the mapped PHB, the packet is forwarded to the PE tail node by the P node, and the PE tail node maps the PHB according to the inner label of the received packet; The PE tail node performs HQoS traffic congestion control on the packet on the user-side AC interface according to the mapped PHB.

Optionally, the PE head node or the P node performs coding/mapping according to the rules shown in Table 1 below. For the EXP, the PE head node encodes, the P node performs mapping and coding, and the last node only performs decoding. The "0" in the second field indicates that the packet is a virtual private network service packet outside the committed bandwidth or the packet color information of the packet is yellow in the virtual link that enters the first node of the PE. The value of "1" indicates that the packet is a virtual private network service packet in the promised bandwidth or the packet color information indicating the packet is green in the virtual link that enters the first node of the PE.

Table 1

Among them, the Assured Forwarding (AF) service has four different forwarding levels. AF1, AF2, AF3, and AF4 respectively indicate different forwarding levels to ensure forwarding services. The CS is a class selection code service, which has a higher forwarding level than the fast forwarding EF service, and is generally signaling. The network control message is of this level, and CS7 is higher than CS6. BE means doing its best for business.

In addition, it should also be noted that the encoding/mapping rules shown in Table 1 are exemplary. For example, in some embodiments, the BE, AF1, and AF2 services of the ingress PHB may also be encoded as the same. a field, and is mapped to the BE service in the P node; in other embodiments, the first field may be encoded by using 2 high bits in the EXP field, and the second field may be encoded by 1 low bit, or The first field is encoded by using one high bit of the EXP field, and the first field is encoded by the two lower bits, or the first field and the second field are encoded by other bit allocation forms.

The embodiment also provides a traffic congestion control device, which is located in the PE head node, and is used to implement the foregoing traffic congestion control method. FIG. 3 is a schematic structural diagram of a traffic congestion control apparatus according to an embodiment of the present invention. As shown in FIG. 3, the apparatus includes: an EXP encoding module 32 and a congestion control module 34, where

The EXP encoding module 32 is configured to: encode an EXP field of an outer label of the packet according to the service type and the packet color information of the inbound PHB of the packet, where the EXP field includes: a service type for identifying the packet. a field, and a second field for identifying a virtual private network service message in the committed bandwidth in the virtual link entering the PE head node;

The congestion control module 34 is coupled to the EXP encoding module 32 and is configured to perform traffic congestion control based on the ingress node PHB.

Optionally, the congestion control module 34 includes: a first level scheduling control unit 342, configured to: root According to the ingress node PHB, at least port scheduling, tunnel scheduling, virtual link scheduling, and flow level scheduling are performed in the first-level scheduling control process, wherein port scheduling is used to ensure that traffic of multiple ports does not interfere with each other, and tunnel scheduling is used. To ensure traffic control of multiple tunnels in a port, virtual link scheduling is used to ensure traffic control of multiple virtual links in a virtual link. Flow-level scheduling is used to ensure different service types in a tunnel carried by a virtual link. The scheduling forwarding requirements to be performed by the business.

Optionally, the congestion control module 34 further includes: a second level scheduling control unit 344, the second level scheduling control unit 344 is coupled to the first level scheduling control unit 342, configured to: during the packet header encapsulation process of the message, according to The ingress-based PHB performs preferential forwarding of packets with a higher forwarding priority of the service type in the second-level scheduling control process.

This embodiment further provides another traffic congestion control device, which is located in the P node, and is also used to implement the above traffic congestion control method.

FIG. 4 is a schematic structural diagram of another traffic congestion control apparatus according to an embodiment of the present invention. As shown in FIG. 4, the apparatus includes: a PHB mapping module 42 and a congestion control module 44, where

The PHB mapping module 42 is configured to: map the PHB according to the EXP field of the packet received by the P node, where the EXP field includes: a first field for identifying a service type of the packet, and an identifier for entering the PE head node Whether the virtual link is the second field of the virtual private network service packet in the committed bandwidth; wherein the first field is mapped to the PHB service type, the second field is mapped to the packet color information, and the second field is mapped The message identified as being within the committed bandwidth is mapped green, and the second field is identified as the message outside the committed bandwidth is mapped to yellow;

The congestion control module 44 is coupled to the PHB mapping module 42 and configured to: perform two-level scheduling control of traffic according to the mapped PHB, wherein the first level scheduling control is used to control bandwidth, and the second level scheduling control is used to control packet priority. Forward.

Optionally, the congestion control module 44 includes:

The first-level scheduling control queue control unit 442 is configured to: add the packet received by the P-node to the same ordered scheduling queue according to the mapped packet color information and equal to an ordered aggregate (OA);

The first-level scheduling control packet discarding unit 444 is coupled to the first-level scheduling control queue control unit 442, and configured to preferentially discard the ordered scheduling queue when congestion control is performed on the traffic. The color information of the medium message is yellow.

Optionally, the first-level scheduling control packet discarding unit 444 preferentially discards the packet color information of the ordered scheduling queue as yellow according to the queue congestion discarding policy of the weighted random pre-detection.

Optionally, the congestion control module 44 further includes:

The second level scheduling control EXP encoding unit 446 is configured to: encode an EXP field of an outer label of the packet according to the service type and the packet color information of the PHB of the packet mapping, where the EXP field includes: a first field of the service type of the text, and a second field of the message color information for identifying the message;

The second level scheduling control message forwarding unit 448 is coupled to the second level scheduling control EXP encoding unit 446, and configured to: perform forwarding priority control on the message according to the service type of the message mapping.

It should be noted that the modules and units involved in the embodiments of the present invention may be implemented by software or by hardware. The described modules and units in this embodiment may also be disposed in a processor. For example, it may be described that a traffic congestion control apparatus includes a processor including a PHB mapping module 42 and a congestion control module 44. The name of these modules does not constitute a limitation on the module itself in some cases. For example, the PHB mapping module 42 may also be described as "a module for mapping a PHB according to an EXP field of a message received by a P node" .

The embodiment further provides a board, which is applied to the PE head node, and is used to implement the foregoing traffic congestion control method. The board includes an EXP encoding chip and a buffer, wherein:

The EXP encoding chip is configured to: according to the service type and the packet color information of the inbound PHB of the packet, the EXP field of the outer label of the encoded message, where the EXP field includes: the first service type used to identify the packet. a field, and a second field for identifying a virtual private network service message in the committed bandwidth in the virtual link entering the PE head node;

The buffer is coupled to the EXP encoding chip and configured to: buffer the message encoded by the EXP encoding chip.

Optionally, the board further includes: a scheduler coupled to the buffer, where the scheduler is configured to perform at least port scheduling, tunnel scheduling, virtual link scheduling, and at least in the first level scheduling control process according to the ingress node PHB. Flow-level scheduling, where port scheduling is used to ensure that traffic on multiple ports does not interfere with each other. The tunneling is used to ensure the traffic control of multiple tunnels in the port. The virtual link scheduling is used to ensure the traffic control of multiple virtual links in the virtual link. The flow-level scheduling is used to ensure the tunnels carried by the virtual link. Scheduling forwarding requirements to be performed by different service types of services.

Optionally, the scheduler is further configured to: during the packet header encapsulation process of the packet, preferentially forward the packet with a higher forwarding priority of the service type in the second-level scheduling control process according to the PHB of the ingress node.

This embodiment further provides another board, which is applied to an intermediate node (ie, a P node) for implementing the foregoing traffic congestion control method. The board includes: a processor and a scheduler, where:

The processor is configured to: map the PHB according to the EXP field of the packet received by the intermediate node, where the EXP field includes: a first field for identifying the service type of the packet, and a virtual field for identifying the packet entering the first node of the PE Whether the second field of the virtual private network service packet in the committed bandwidth is in the link; wherein the first field is mapped to the PHB service type, the second field is mapped to the packet color information, and the second field is identified as The message within the committed bandwidth is mapped green, and the second field is identified as the message outside the committed bandwidth is mapped to yellow;

The scheduler is coupled to the processor and configured to: perform two-stage scheduling control of the traffic according to the mapped PHB, where the first-level scheduling control is used to control the bandwidth, and the second-level scheduling control is used to control the preferential forwarding of the packet.

In order to make the technical solutions and implementation methods of the present invention clearer, the implementation process will be described in detail below in conjunction with the optional embodiments.

The present invention provides an MPLS network E-LSP end-to-end QOS method and apparatus, and relates to multi-protocol label switching (MPLS) network tunnel end-to-end quality of service (Quality of Service (referred to as QoS) technical field, especially the end-to-end QoS guarantee technology when using the EXP-Label Switch Path (E-LSP) tunnel model.

In order to overcome the QoS problem and the defect that the end-to-end guarantee VPN is not available in the related art, the method and apparatus for the end-to-end QoS guarantee based on the E-LSP model provided by the alternative embodiment adopt the following technical solutions:

An end-to-end QoS guarantee based on the E-LSP model includes:

Step 1: The PE first node and the P node use a custom code for the EXP field encoding of the E-LSP.

Step 2: The P node ingress node (INGRESS) side identifies the packet discarding level according to the custom encoding rule in step 1.

Step 3: The P-node tunnel service flow is treated according to an ordered aggregate (Ordered Aggregate, OA for short), and the queue congestion policy is configured to partially/all discard yellow packets when they are congested.

The color of the P-node packet is transparently transmitted, so that the execution of the P-node congestion management policy is consistent with the coloring of the first node, and the packet is not out of order;

Among them, the PE first and last nodes, the P node strictly performs HQOS scheduling management. Conditionally implement a two-level scheduling mechanism. The first and last nodes perform the first level of scheduling, the purpose of which is to ensure the virtual link bandwidth requirement and control the total traffic, and to ensure that the scheduling behavior of the internal flow of the same virtual link satisfies the superior forwarding relationship of the service service; It is guaranteed to preferentially forward high-priority services and ensure service delay requirements. The P-node performs the first-level scheduling, and the purpose is to ensure the tunnel bandwidth requirement and control the priority to forward the green packet. The second-level scheduling is consistent with the first-to-last node.

The end-to-end QoS of the E-LSP in the optional embodiment indicates the discarding level of the VPN service packet transmitted in the tunnel through a certain bit of the EXP field of the E-LSP to ensure that the P node preferentially forwards the VPN service to ensure the bandwidth portion of the traffic. .

Optionally, the method for ensuring end-to-end QoS guarantee based on the E-LSP model provided by the optional embodiment includes the following steps:

Step 1: Specify the EXP field encoding (ENCODE) and decoding (DECODE) table rules for the tunnel, for example:

EXP ENCODE table: any 2 bit indicates the service category, which can represent 4 categories: BE, AF, Expedited Forwarding (EF), Class Selector (CS); EX remaining The bit bit indicates the discard level, which can represent two discard levels: low discard level, high discard level.

The DECODE table of the EXP: the service type (Class of Service, referred to as CoS) of the 2-bit mapping per-hop behavior (PHB) of the service level in the EXP, and the bit indicating the discarding level in the EXP indicates the discarding priority of the PHB. level. DECODE and ENCODE are reversible.

The second step: the first node processing rule of the INGRESS side PE:

The first-stage scheduler of the PE head node performs at least port, tunnel, pseudowire, and flow level 4-level HQoS scheduling. The port scheduling is to ensure that the traffic of multiple ports does not interfere with each other; the tunnel scheduling is to ensure the flow control of each tunnel in the port; the pseudowire scheduling is to ensure the flow control of the pseudowire in the tunnel, and the bandwidth of the pseudowire is the VPN service in the MPLS network. The bandwidth to be transmitted; the flow level scheduling is to meet the scheduling and forwarding requirements of different service level services in the VPN carried by the pseudowire.

One of the purposes of the second-level scheduling of the PE first-node is to preferentially forward high-priority services after the packet is encapsulated on the ENGRESS side.

The two-stage scheduling mechanism of the PE first node ensures the bandwidth, delay, and jitter requirements of the VPN service.

The mapping from PHB to EXP is mapped according to the encoding method of the above ENCODE, and can carry the above four service levels; the VPN service packet whose traffic is in the pseudowire (ie virtual link) committed bandwidth (CIR) is mapped to the tunnel EXP. Low drop level, otherwise, mapped to the high drop level of the tunnel EXP.

The third step: the P node decodes according to the encoding/decoding table shown in the first step. A total of four service levels are supported: BE, AF, EF, CS, and 2 drop levels. In the case of hierarchical QoS (HQoS), the first-level scheduling controls the bandwidth, and the tunnel bearer service is equivalent to one OA, and enters an ordered queue. The congestion and discarding policy is configured to cause the yellow packet to be discarded immediately after congestion. The secondary scheduling controls the forwarding priority to ensure that the high-priority services are preferentially forwarded. Through HQoS, not only the bandwidth of different tunnels but also the QoS of different services in each tunnel can be realized.

The color-coded bit in the P-block EXP code is transparently transmitted and is not re-encoded, so that the next P-node performs congestion policy control to ensure that the yellow packet is a traffic packet in the EIR bandwidth of the first node.

The fourth step: PE tail node processing rule: PHB mapping no longer pays attention to the tunnel pipeline mode, and uniformly maps the PHB with the EXP in the innermost label. To ensure end-to-end QoS, QoS control always acts on the downstream AC (VPN Access Control Interface) exit.

Compared with the related technologies, the method of the alternative embodiment or the device and system based on the method has wide adaptability, and can ensure the end-to-end QoS capability of each VPN for the MPLS network adopting the E-LSP technology, and improves the QoS capability of the VPN. System suitability and usability.

The above alternative embodiments will be further described below with reference to the accompanying drawings.

In the first step and the second step, the PE first node performs HQoS processing, including: enabling the HQoS (hierarchical QoS) function of the first level scheduling at the PE head node, and setting the tunnel EXP coding mode to 4P4D (4P indicates 4 types of priority) Level service class, 4D indicates the discarding category of the four priority service classes, as shown in Table 1. Its scheduling tree structure is shown in Figure 5. The pseudowire, tunnel, and port perform strict CAC check. The flows of each service level in the pseudowire are scheduled according to the flow scheduling policy to meet the requirements of the traffic flow characteristic QOS. The pseudowire scheduling is to meet the requirements of the pseudowire pair bandwidth resources. The scheduling of the tunnel is to meet the bandwidth requirements of the tunnel. Port-level scheduling is to ensure fairness and fairness of flow control between ports. According to the above hierarchical model, the flow forwarded from the node, the flow in the pseudowire is preferentially forwarded according to the service level relationship according to the PQ policy/WFQ policy/DWRR policy, and then the overall traffic of the pipeline is pseudo-wired. The control method is to prioritize its guaranteed bandwidth and then try to forward its excess bandwidth traffic. After the flow carried by the pseudowire reaches the tunnel, the tunnel is not aware of the customer service. The tunnel primary scheduling ensures that the promised bandwidth of the tunnel is guaranteed and tries to forward the excess bandwidth traffic of the tunnel. And so on.

At the SES Level 1 scheduler, recoloring is turned on and the color is carried to the Network Processor (NP). The color of the PIR part of the pseudowire is green, and the color of the EIR part of the pseudowire is yellow. When the NP encapsulates the MPLS E-LSP tunnel label, the color is changed to the low bit of the EXP. For example, the low bit 0 indicates that the packet color is yellow, 1 indicates the packet color is green, and the high 2 bit is encoded according to the first field of Table 1. .

The second-level scheduling is based on the NP chip scheduling tree configuration to ensure that high-priority services are forwarded without blocking priority. Schematic diagram of NP chip EGRESS side scheduling and forwarding is shown in Figure 6. There is only one SMS module in the chip. In this embodiment, the SMS module is divided into three types of forwarding directions: A, B, and C.

The SMS Class A module corresponds to the forwarding direction of the TI to SMS module.

After the SMS B-type module corresponds to the Pipeline to SMS module, the SMS implements the multicast replication direction.

The SMS C class module is sent to the TI (Traffic Interface) interface direction after the Pipeline to the SMS module.

The PT0 queue management adopts the following configurations:

1) 512 queues, divided into 64 queue groups, each group of 8 queues.

2) Each queue group corresponds to one Node A, a total of 64 Node A.

3) 16 Node Bs, and the mapping relationship between Node A and Node B can be arbitrarily configured as needed.

4) 2 trees, Node B to tree mapping can be arbitrarily configured.

5) The Level A Node corresponds to the TI subport.

6) 8 queues are connected under the Level A Node to distinguish 8 priority levels.

The PT1 queue management adopts the following configurations:

1) 128 queues, divided into 64 queue groups, each group of 2 queues.

2) Each queue group corresponds to one node A, 64 Node A.

4) 2 trees, Node B to tree mapping can be arbitrarily configured.

5) The Level B node corresponds to the PP subport.

6) In addition to the TI access, the Level A node also includes logical multicast and loopback access.

7) Each Level A node is connected to two queues and is divided into high and low priorities.

In the third step, the P-node performs the HQoS processing process, including: performing uplink aggregation on the P node, indicating behavior aggregation (Behavior Aggregate, abbreviated as BA) according to the high 2 bit of the EXP, and discarding the low-bit bitmap mapping level according to Table 1. An ordered queue is assigned based on the uniqueness of the tunnel. The queue congestion discarding policy is Weighted Random Early Detection (WRED).

The WRED parameter can be configured as follows: high threshold = maximum allowable length of the queue, low threshold of the yellow service = high threshold/10, high threshold of the yellow service = high threshold, and maximum discard rate = 100%. The lower threshold of the green service = the high threshold * 0.75, the high threshold of the green service = the high threshold, and the maximum discard rate = 10%.

In the P node, the primary scheduling is performed by the scheduling management chip (SA), and its scheduling tree is as shown in FIG. The secondary scheduling is done by the NP, which performs strict priority (SP) scheduling in accordance with 8 service levels. The P-node downlink encapsulates the high 2-bit and the PE first node, and the low-bit replicates the uplink EXP low-bit.

In the fourth step, the PE tail processing process of the PE tail node includes: at the PE tail node, if the layer 2 (L2) VPN, the PHB is mapped according to the EXP of the PW label; if the layer 3 (L3) VPN, according to the VPN routing Forwarding (VRF) tag EXP mapping PHB. In order to prevent multiple VPNs After the aggregation to a port, the traffic of the local VPN in different routing directions may cause other VPN bandwidths to be unguaranteed. Therefore, the HQoS control of the AC interface on the user side (U side) needs to be started. The scheduling tree structure is shown in Figure 8. Show. Since an AC interface can only be bound to one VPN, controlling the total traffic of the AC interface is equivalent to controlling the total traffic of the VPN service termination packet. The sub-port scheduler in Figure 8 represents the AC interface level 1 scheduling. The flow-level scheduler is connected to the scheduler. The flow-level scheduler is configured to implement the user scheduling policy and preferentially forward the high-priority service. The sub-port scheduler forwards the CIR part traffic preferentially, and forwards the EIR part of the traffic if the port bandwidth is sufficient. Each AC occupies a sub-port scheduler to ensure that each VPN service burst does not preempt other VPN guaranteed bandwidth. For the L2LINE service, the sub-port scheduler CIR is converted according to the pseudowire bandwidth (for example, if the sub-port link is an Ethernet link, the total bandwidth is calculated according to the pseudo-line bandwidth minus the N-side encapsulation plus the U-side encapsulation); For LAN/TREE services, you need to configure the AC interface bandwidth. For the L3VPN, the upstream and downstream routes are dynamic and static. Therefore, the traffic on the AC cannot be accurately estimated. This requires the user to manually set the bandwidth of the AC interface according to the networking plan. By default, the traffic is forwarded at full speed.

In summary, the foregoing embodiment of the present invention encapsulates the corresponding VPN information in the outer label of the MPLS packet, thereby solving the P-node in the case of multiplexing the multi-virtual private network service in the related art. The problem that the VPN information cannot be resolved and the end-to-end quality of service cannot be guaranteed ensures the end-to-end QoS control of the multi-VPN service multiplexing tunnel.

It will be apparent to those skilled in the art that the various modules or steps of the present invention described above can be implemented by a general-purpose computing device that can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in a storage device by a computing device, or they may be fabricated into individual integrated circuit modules, or Multiple modules or steps are made into a single integrated circuit module. Thus, the invention is not limited to any specific combination of hardware and software.

The above is only an alternative embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Industrial applicability

The foregoing technical solution uses the PE head node to encode the EXP field of the outer label of the packet according to the service type and the packet color information of the ingress PHB of the packet, where the EXP field includes the first service type for identifying the packet. The field is used to identify whether the packet is the second field of the virtual private network service packet in the committed bandwidth in the virtual link that enters the PE head node; the PE first node solves the traffic congestion control according to the ingress node PHB. The related art is based on the E-LSP MPLS network. In the case of multi-VPN service multiplexing tunnel, the end-to-end QoS control cannot be realized because the P-node cannot resolve the VPN information, and the multi-VPN service multiplexing tunnel is guaranteed. End-to-end QoS control. Therefore, the present invention has strong industrial applicability.

Claims

An outer label encoding method comprising:

The operator's edge router, the PE head node, encodes the EXP field of the outer label of the packet according to the service type of the PHB and the packet color information after the traffic congestion control is performed.

The EXP field includes: a first field that is used to identify the service type of the packet, and a virtual field in the virtual link that is used to identify the packet in the virtual link that enters the PE head node. The second field of the private network service message.
The outer label encoding method according to claim 1, wherein before the PE head node encodes the EXP field, the method further comprises:

The PE head node performs traffic congestion control according to the PHB of the ingress node;

The step of the PE head node performing traffic congestion control according to the ingress node PHB includes:

The PE head node performs first-level scheduling control according to the PHB of the ingress node, and performs at least port scheduling, tunnel scheduling, virtual link scheduling, and flow level scheduling in the first-level scheduling control process, where the port scheduling is performed. To ensure that the traffic of the multiple ports is not interfered with each other, the tunnel scheduling is used to ensure the traffic control of multiple tunnels in the port, and the virtual link scheduling is used to ensure the traffic control of multiple virtual links in the tunnel. The flow-level scheduling is used to ensure the scheduling and forwarding requirements of services of different service types in the virtual link carried by the virtual link.
The outer label encoding method according to claim 2, wherein the step of the PE head node performing traffic congestion control according to the ingress node PHB further comprises:

The PE head node performs the second-level scheduling control according to the PHB of the ingress node during the packet header encapsulation process of the packet, and preferentially forwards the packet with the higher forwarding priority of the service type in the second-level scheduling control process.
The outer tag encoding method according to claim 2, wherein after the step of the PE header encoding the EXP field of the outer label of the packet, the method further comprises:

The intermediate node maps the PHB according to the EXP field of the received message, where the first field is mapped to a service type of the PHB, and the second field is mapped to message color information, where In the second field, the packets in the committed bandwidth are mapped green, and the packets outside the committed bandwidth are mapped to yellow.

The intermediate node performs the first-level scheduling control and the second-level scheduling control of the traffic according to the mapped PHB, where the first-level scheduling control is used to control the bandwidth, and the second-level scheduling control is used to control the preferential forwarding of the packet.
The outer label encoding method according to claim 4, wherein the step of the intermediate node performing the first level scheduling control of the traffic according to the mapped PHB comprises:

The intermediate node equates the received packet with an ordered aggregate OA, and adds the received packet to the same ordered scheduling queue according to the mapped packet color information.

The intermediate node preferentially discards the packet whose color information is yellow in the ordered scheduling queue when congestion control is performed on the traffic.
The outer label encoding method according to claim 5, wherein the step of the intermediate node preferentially discarding the packet color information in the ordered scheduling queue is yellow:

The intermediate node preferentially discards the packet whose color information is yellow in the ordered scheduling queue according to the queue congestion discarding policy of the weighted random pre-detection.
The outer tag encoding method according to claim 4, wherein the step of the intermediate node performing the second-level scheduling control of the traffic according to the mapped PHB comprises:

The intermediate node encodes an EXP field of the outer label of the received packet according to the service type and the packet color information of the PHB of the packet, where the EXP field includes: a service type for identifying the packet. a first field, and a second field for identifying message color information of the message;

The intermediate node performs forwarding priority control on the packet according to the service type of the packet mapping.
The outer label encoding method according to claim 4, wherein after the intermediate node performs the first level scheduling control and the second level scheduling control of the traffic according to the mapped PHB, the method further includes:

The PE tail node maps the PHB according to the inner label of the received packet;

The PE tail node performs HQoS traffic congestion control on the packet on the user-side AC interface according to the mapped PHB.
The outer label encoding method according to any one of claims 1 to 8, wherein

The PE head node or intermediate node is coded or mapped according to the following rules:

The "0" in the second field indicates that the packet is a virtual private network service packet outside the committed bandwidth or the packet color information of the packet is yellow in the virtual link that enters the first node of the PE. The "1" in the second field indicates that the packet is a virtual private network service packet in the promised bandwidth or the packet color information indicating the packet is green in the virtual link entering the first node of the PE.
An outer label encoding device is located in a first node of a carrier edge router PE, and includes an EXP encoding module, where:

The EXP encoding module is configured to: encode an EXP field of an outer label of the packet according to a service type of the PHB and a packet color information after the traffic congestion control is performed, and the The EXP field includes: a first field for identifying a service type of the packet, and a virtual private network service report for identifying whether the packet is a committed bandwidth in a virtual link entering the PE head node. The second field of the text.
The outer tag encoding apparatus according to claim 10, said apparatus further comprising a congestion control module, wherein:

The congestion control module is configured to: perform traffic congestion control according to the ingress node PHB; the congestion control module includes a first level scheduling control unit, where:

The first-level scheduling control unit is configured to: perform first-level scheduling control according to the ingress node PHB, and perform at least port scheduling, tunnel scheduling, virtual link scheduling, and flow-level scheduling in the first-level scheduling control process, where The port scheduling is used to ensure that traffic of multiple ports does not interfere with each other. The tunnel scheduling is used to ensure traffic control of multiple tunnels in a port, and the virtual link scheduling is used to ensure multiple virtual links. The traffic control of the virtual link is used to ensure the scheduling and forwarding requirements of services of different service types in the tunnel carried by the virtual link.
The outer tag encoding apparatus according to claim 11, wherein said congestion control module further comprises a second level scheduling control unit, wherein:

The second-level scheduling control unit is configured to perform second-level scheduling control according to the ingress node PHB during the packet header encapsulation process of the packet, and preferentially forward the forwarding priority of the service type in the second-level scheduling control process. Higher message.
A traffic congestion control device is located in an intermediate node, and includes a PHB mapping module and a congestion control module for each jump release, wherein:

The PB mapping module is configured to: map the PHB according to the EXP field of the packet received by the intermediate node, where the EXP field includes: a first field for identifying a service type of the packet, and The second field of the virtual private network service packet in the committed bandwidth is determined in the virtual link of the first node of the operator edge router PE;

The per-hop is a PHB mapping module that maps the first field to a PHB service type, and the second field is mapped to packet color information, and when the second field is mapped, the committed bandwidth is The packets in the packet are mapped in green, and the packets outside the committed bandwidth are mapped to yellow.

The congestion control module is configured to perform two-level scheduling control of traffic according to the mapped PHB, where the first-level scheduling control is used to control bandwidth, and the second-level scheduling control is used to control priority forwarding of packets.
The traffic congestion control apparatus according to claim 13, wherein said congestion control mode The block includes a first-level scheduling control queue control unit and a first-level scheduling control packet discarding unit, where:

The first level scheduling control queue control unit is configured to: equate the received message with the intermediate node to an ordered aggregate OA, and add the received message to the same ordered scheduling queue according to the mapped message color information. ;

The first-level scheduling control packet discarding unit is configured to: in the case of performing congestion control on the traffic, preferentially discarding the packet whose color information is yellow in the ordered scheduling queue.
The traffic congestion control apparatus according to claim 14, wherein the first-level scheduling control packet discarding unit is configured to preferentially discard the packet whose color information is yellow in the ordered scheduling queue according to the following manner:

According to the queue congestion discarding policy of the weighted random pre-detection, the packet with the yellow color of the packet in the ordered scheduling queue is discarded.
The traffic congestion control apparatus according to claim 13, wherein the congestion control module further comprises a second level scheduling control EXP encoding unit and a second level scheduling control message forwarding unit, wherein:

The second level scheduling control EXP encoding unit is configured to: encode an EXP field of an outer label of the packet according to the service type and the packet color information of the PHB of the packet mapping, where the EXP field includes: a first field of the service type of the message, and a second field for identifying the message color information of the message;

The second-level scheduling control packet forwarding unit is configured to: perform forwarding priority control on the packet according to the service type of the packet mapping.
A single board includes an EXP encoding chip and a buffer, wherein:

The EXP encoding chip is configured to: encode an EXP field of an outer label of the packet according to a service type of the PHB and a packet color information after the traffic congestion control is performed, wherein the The EXP field includes: a first field for identifying a service type of the packet, and a virtual private network for identifying whether the packet is a committed bandwidth in a virtual link entering a head node of the operator edge router PE The second field of the service message;

The buffer is coupled to the EXP encoding chip and configured to: buffer a message encoded by the EXP encoding chip.
The board of claim 17, the board further comprising a scheduler, wherein:

The scheduler, coupled to the buffer, is configured to perform a first-level scheduling control process according to the ingress node PHB, and perform at least port scheduling, tunnel scheduling, virtual link scheduling, and the first-level scheduling control process. Flow-level scheduling, where the port scheduling is used to ensure that traffic of multiple ports does not interfere with each other. The tunnel scheduling is used to ensure traffic control of multiple tunnels in a port, and the virtual link scheduling is used to ensure virtual The flow control of multiple virtual links in the link is used to ensure the scheduling and forwarding requirements of services of different service types in the tunnel carried by the virtual link.
The board according to claim 18, wherein the scheduler is further configured to: during the packet header encapsulation process of the packet, preferentially forward the service type in the second-level scheduling control process according to the ingress node PHB Forwards packets with higher priority.
A board, including a processor and a scheduler, where:

The processor is configured to: according to the EXP field mapping of the packet received by the intermediate node, the per-jump is issued as a PHB, where the EXP field includes: a first field for identifying a service type of the packet, And a second field for identifying a virtual private network service message in the committed bandwidth in the virtual link of the first node of the operator edge router PE;

The processor maps the first field to a service type of the PHB, and maps the second field to packet color information. When mapping the second field, the packet in the committed bandwidth is mapped. Green, packets outside the committed bandwidth are mapped to yellow;

The scheduler, coupled to the processor, is configured to perform two-level scheduling control of traffic according to the mapped PHB, where the first level scheduling control is used to control bandwidth, and the second level scheduling control is used to control message Priority forwarding.
A computer program comprising program instructions, when the program instructions are executed by a PE head node, an intermediate node, and a tail node, such that the PE head node, the intermediate node, and the tail node can perform any one of claims 1-9 The outer label encoding method.
A carrier carrying the computer program of claim 21.