SR Policy Group SPRING Working Group F. Yang Internet Draft China Mobile Intended status: Standards Track C. Lin Expires: April 24, 2023 Y. Qiu New H3C Technologies October 21, 2022 Intelligent Routing Method of SR Policy draft-yang-sr-policy-intelligent-routing-00 Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html This Internet-Draft will expire on April 24 2023. Copyright Notice Copyright (c) 2021 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Yang, et al. Expire April, 2023 [Page 1] Internet-Draft SR Policy Intelligent Routing Method October 2022 Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Abstract Segment Routing is a source routing paradigm that explicitly indicates the forwarding path for packets at the ingress node. An SR Policy is associated with one or more candidate paths, and each candidate path is either dynamic, explicit or composite. This document describes an intelligent routing method for SR Policy based on network quality in MPLS and IPv6 environments. Table of Contents 1. Introduction ................................................ 2 2. Terminology ................................................. 3 3. Requirement Scenarios........................................ 3 3.1. Requirement 1 .......................................... 4 3.2. Requirement 2 .......................................... 5 4. Intelligent Routing Method for SR Policy .................... 5 4.1. Processing model ....................................... 5 4.2. Flow Classification .................................... 6 4.3. Flow Steering .......................................... 7 4.4. Intelligent Routing .................................... 7 4.5. Network Quality Measurement ............................ 8 4.6. Flow Forwarding......................................... 9 5. Examples of intelligent routing ............................. 9 6. IANA Considerations ........................................ 11 7. Security Considerations .................................... 11 8. References ................................................. 12 8.1. Normative References .................................. 12 8.2. Informative References ................................ 12 9. Acknowledgments ............................................ 12 Authors' Addresses ............................................ 13 1. Introduction Segment routing (SR) [RFC8402] is a source routing paradigm that explicitly indicates the forwarding path for packets at the ingress node. The ingress node steers packets into a specific path according to the Segment Routing Policy (SR Policy) as defined in [I-D.ietf- spring-segment-routing-policy]. In order to distribute SR policies to the headend, [I-D.ietf-idr-segment-routing-te-policy] specifies a mechanism by using BGP. An SR Policy is associated with one or more candidate paths. A composite candidate path acts as a container for grouping SR Yang, et al. Expires April, 2023 [Page 2] Internet-Draft SR Policy Intelligent Routing Method October 2022 Policies. As described in section 2.2 in [I-D.ietf-spring-segment- routing-policy], the composite candidate path construct enables combination of SR Policies, each with explicit candidate paths and/or dynamic candidate paths with potentially different optimization objectives and constraints, for load-balanced steering of packet flows over its constituent SR Policies. For convenience, the composite candidate path formed by the combination of SR policies is called parent SR policy in [I-D.jiang-spring-parent-sr- policy-use-cases]. This document describes an intelligent routing method for SR Policy based on network quality in MPLS and IPv6 environments. 2. Terminology The definitions of the basic terms are identical to those found in Segment Routing Policy Architecture [I-D.ietf-spring-segment- routing-policy] and [I-D.jiang-spring-parent-sr-policy-use-cases]. 3. Requirement Scenarios Take the networking shown in Figure 1 below as an example to illustrate the current problems. PE1 and PE2 are the two access endpoints of the IP telecom network. There are many service flows between PE1 and PE2 that have different requirements for forwarding quality, such as voice, video and Internet services. Generally, from PE1 to PE2, voice services with low delay requirements are forwarded along the shortest path PE1- >P4->PE2. The Video traffic is forwarded along path PE1->P1->P2->PE2. The forwarding path of other traffic is PE1->P1->P3->P2->PE2. Yang, et al. Expires April, 2023 [Page 3] Internet-Draft SR Policy Intelligent Routing Method October 2022 +-------------+ | | | Controller | | | /+-------------+\ / \ +---- ----+ / \ / +-----+ \ / +----| P3 |----+ \ / | +-----+ | \ | | +-----+ +--+--+ +--+--+ +-----+ | PE1 |-----| P1 |---------| P2 |-----| PE2 | +--+--+ +-----+ +-----+ +--+--+ | | | +-----+ | +----------------+ P4 +----------------+ +-----+ Figure 1 3.1. Requirement 1 In Section 4 of [I-D.jiang-spring-parent-sr-policy use cases], the flow steering method of parent SR policy through flow characteristics is introduced. Specify the mapping relationship between color and flow characteristics (such as DSCP) for parent SR policy. Upon receiving a packet with the specified destination address, the device searches for the SR policy containing the color value mapped to the flow characteristics of the packet in the parent SR policy. This method can well meet the scenarios where different services have different forwarding quality requirements. If there are multiple equivalent policy paths or backup paths for services with the same characteristics, only one valid constituent policy will always be matched within the parent SR policy according to the flow characteristics. Take Path1 (PE1->P4->PE2) and Path2 (PE1->P1->P2->PE2) in Figure 1 for example. The transmission delay of both paths can meet the requirements of voice service. When the remaining bandwidth on Path1 is relatively small, even if Path2 is relatively idle, the voice traffic of new access users will not be shared on Path2. The voice traffic cannot share the bandwidth on the equivalent policy path. When quality of the constituent SR policy is degraded, Yang, et al. Expires April, 2023 [Page 4] Internet-Draft SR Policy Intelligent Routing Method October 2022 it cannot switch to the backup constituent policy paths with better quality. To solve the problem in this scenario, this document proposes an intelligent routing method to lower down the impact of network quality degradation for SR policy. 3.2. Requirement 2 In the SRv6 network, the controller is usually centrally deployed to manage all network nodes. The controller collects performance measurement data from the network node, calculates the measurement results, and then programs the forwarding path to the head node according to the service requirements. For the customers who have high requirements for SLA such as finance industry, this method cannot meet the customer's requirements. Firstly, the node on the path will report measuring data to the controller, and then the centralized controller calculates the path per quality restriction criteria, and then the controller notifies the source node to schedule the path of traffic. The whole control loop is too long and it is difficult to guarantee the SLA requirements of customers in this way. In some large networks, such as SD-WAN network, the number of network nodes is huge, and the path calculation complication on the controller is too high. In order to better meet these requirements, this document proposes an intelligent routing method for SR policy based on network quality requirement. The head end node selects the optimal path according to the current network quality to improve the path switching speed and forwarding performance. 4. Intelligent Routing Method for SR Policy 4.1. Processing model The path priority is assigned to the SR policy forwarding path manually by the controller. Each path with quality requirement will be assigned with a priority value. The lower the value, the higher the priority. That is, when there is a group of qualified paths, best path will be selected with higher priority. Configure multiple SR policy paths for the service flows with specified characteristics. Assign the corresponding path priority to each path according to the priority order of the path. If the traffic needs to be shared by multiple SR policies, assign the same priority and sharing weight values to these SR policies. If there is Yang, et al. Expires April, 2023 [Page 5] Internet-Draft SR Policy Intelligent Routing Method October 2022 a backup path for the SR policies, lower priority value should be used according to the quality requirements. After receiving the service packet with the specified characteristics, when the network quality is good, the traffic is forwarded through the path with high priority. When the network quality degradation is happened on the high priority path, such as the packet loss rate exceeds the acceptable range, switch to the next high priority path of the service. If the quality of the high priority forwarding path is restored and the specified quality requirements are met, the traffic is switched from the low priority forwarding path to the high priority forwarding path. When there are multiple paths with the same priority, the traffic will share the bandwidth on these paths with the same priority according to the weight value. According to the processing logic, the SR policy intelligent routing model can be divided into five units, including Flow Classification, Flow Steering, Intelligent Routing, Flow Forwarding, and Network Quality Measurement, as shown in Figure 2 below. +----------------+ +----------+ +-------------+ +------------+ | | | | | | | | | Flow |->| Flow |->| Intelligent |->| Flow | | Classification | | Steering | | Routing | | Forwarding | | | | | | | | | +----------------+ +----------+ +------+------+ +------------+ ^ | +-----------------+ | | | Network Quality | | Measurement | | | +-----------------+ Figure 2 The functions of each uint are described below. 4.2. Flow Classification After receiving the traffic, the head node first needs to label the traffic with forwarding class according to classification configuration. Yang, et al. Expires April, 2023 [Page 6] Internet-Draft SR Policy Intelligent Routing Method October 2022 The head node can match flow characteristics in its ingress interfaces (upon any field such as Ethernet destination/source/VLAN/TOS or IP destination/source/DSCP or transport ports or application attribute etc.) and color them with an internal per-packet forwarding-class variable. 4.3. Flow Steering According to the forwarding class variables determined by the Flow classification, the header node selects the matching forwarding path, that is, selects the SR policy or the parent SR policy representing a group of policies. If multiple SR policy forwarding paths are configured for the traffic flow with the specified characteristics, all valid SR policies will be retrieved and handed over to the Intelligent Routing unit to select the optimal forwarding path. 4.4. Intelligent Routing According to the SR policy(policies) provided by Flow Steering, the Intelligent Routing unit obtains the current quality of each SR policy path from the Network Quality Measurement unit. Based on the mapping between the quality and the priority of intelligent routing, it selects the forwarding path with the highest priority and the best quality requirements for the traffic flow. When the network quality is good, the traffic is forwarded by the policy with high priority. When the network quality of the high priority SR policy degrades, such as the packet loss rate exceeds the acceptable range, switch to the secondary high priority SR policy path forwarding. Similarly, when the next higher priority SR policy forwarding path cannot meet the forwarding quality requirements, switch to the lower priority SR policy path. If the quality of the high priority forwarding path gets better and meets the specified quality requirements, the traffic can be recovered from the low priority forwarding path to the high priority SR policy forwarding path. If multiple SR policies in the parent SR policy have the same priority, when the traffic flow selects the path of this priority, the traffic flow will share the load on these SR policy paths. In order to better control the switchover process of SR policies between different priorities, the quality threshold and switchover waiting time can be specified. If the value of network quality degradation consistently exceeds the switchover threshold during the Yang, et al. Expires April, 2023 [Page 7] Internet-Draft SR Policy Intelligent Routing Method October 2022 waiting time, select the SR policy with the next highest priority from the highest priority to the lowest. In contrast, if the network quality continues to improve, after exceeding the failback waiting time, the forwarding path can be switched to a high priority SR policy that meets the current network quality requirements. In order to avoid frequent path switching when the network quality is unstable, if the current path can meet the forwarding quality requirements, the head node can choose not to automatically switch back to the higher priority path in case of the quality of the higher priority path is restored. The device can provide a configuration for automatic failback, and add a restore waiting timer. Only after automatic restore is allowed and the restore waiting timer is timeout, the forwarding path switch from the current path that meets the quality requirements to the path with higher priority. 4.5. Network Quality Measurement The Network Quality Measurement unit regularly monitors the quality of all effective forwarding paths according to the measurement cycle, records the current performance measurement data of the path, and reports it to the Intelligent Routing unit, which decides whether to switch paths. The following network quality parameters of forwarding path can be used for path scheduling: Jitter Latency Packet loss Available bandwidth Bandwidth utilization Current traffic statistics Other forwarding performance parameters The quality parameters of network forwarding path can be obtained through active or passive performance measurement methods, such as iOAM, STAMP, TWAMP, etc. The network quality parameters can be calculated by the controller and distributed to the head end node, or calculated by the head end node according to the network Yang, et al. Expires April, 2023 [Page 8] Internet-Draft SR Policy Intelligent Routing Method October 2022 measurement data. The measurement method and quality parameter acquisition method are beyond the scope of this document. 4.6. Flow Forwarding The service flow is forwarded according to the path determined by the Intelligent Routing unit. When there are multiple paths with the same priority, the traffic will share the load among these SR policy paths with the same priority according to the weight value. 5. Examples of intelligent routing The application of intelligent routing is described in detail in L3VPN over TE scenario. The networking is shown in Figure 3 below. CE1 and CE2 belong to the same L3VPN and access the public network through PE1 and PE2 respectively. There are three services between CE1 and CE2: voice, video and Internet access. The traffic from CE1 to CE2 can be forwarded through three paths: Path1 (PE1 ->P4 ->PE2), Path2 (PE1->P1->P2-> PE2), and Path3 (PE1 ->P1 ->P3 ->P2 ->PE2). Among them, Path1 has the shortest path and the highest priority. The second priority is Path2. Path3 has the lowest priority. Voice service use service ID 1 to identify and forward on Path1 first. When the delay requirement of Path1 cannot be met and Path2 can be met, switch the voice service to Path2. If the delay of Path2 fails to meet the voice service requirements and Path3 meets the requirements, switch the service to Path3. Video service use service ID 2 to identify and forward it on Path2 first. When the packet loss rate of Path2 fails to meet the requirements and Path3 can meet the requirements, switch the video service to Path3. If the packet loss rate of Path3 fails to meet the video service requirements and Path1 meets the requirements, switch the video service to Path1, the path reserved for voice service. Other services are identified by service ID 3, and priority is given to forwarding on Path3. When the remaining bandwidth of Path3 is insufficient, find the path with enough remaining bandwidth in the order of Path2 and Path1, that is, share the traffic to the path previously reserved for video and voice. Yang, et al. Expires April, 2023 [Page 9] Internet-Draft SR Policy Intelligent Routing Method October 2022 +----+ +--| P3 |--+ | +----+ | | | +-----+ +-----+ +----+ +----+ +-----+ +-----+ | CE1 |---| PE1 |---| P1 |----| P2 |----| PE2 |---| CE2 | +-----+ +--+--+ +----+ +----+ +--+--+ +-----+ | | | +----+ | +-----------+ P4 +------------+ +----+ Figure 3 The configuration on the head node PE1 includes the following three parts. These configurations can be directly configured on the node or distributed through the controller. 1. Define three intelligent routing policies, and specify the threshold of network quality, path priority and the corresponding path color value for routing. intelligent-routing-policy irp1 traffic-delay threshold 1000 priority 1 mapping-to color 100 priority 2 mapping-to color 200 priority default mapping-to color 300 intelligent-routing-policy irp2 packet-loss-rate threshold 5 priority 1 mapping-to color 200 priority 2 mapping-to color 300 priority default mapping-to color 100 intelligent-routing-policy irp3 remaining-bandwidth threshold 50M priority 1 mapping-to color 300 Yang, et al. Expires April, 2023 [Page 10] Internet-Draft SR Policy Intelligent Routing Method October 2022 priority 2 mapping-to color 200 priority default mapping-to color 100 2. Configure forwarding paths. sr-policy policy-A (color 100, PE4_SID) segment-list SID_PE1, SID_P4, SID_PE4 sr-policy policy-B (color 200, PE4_SID) segment-list pe1, p1, p2, pe2 sr-policy policy-C (color 300, PE4_SID) segment-list pe1, p1, p3, p2, pe2 3. Configure corresponding intelligent routing policies for services with specified characteristics. parent-sr-policy sr-policy-1(color 10, PE4_SID) service-ID 1 use intelligent-routing-policy irp1 service-ID 2 use intelligent-routing-policy irp2 service-ID default use intelligent-routing-policy irp3 6. IANA Considerations This document has no IANA actions. 7. Security Considerations This document does not introduce any security considerations. Yang, et al. Expires April, 2023 [Page 11] Internet-Draft SR Policy Intelligent Routing Method October 2022 8. References 8.1. Normative References [I-D.ietf-spring-segment-routing-policy] Filsfils, C., Talaulikar, K., Voyer, D., Bogdanov, A., and P. Mattes, "Segment Routing Policy Architecture", draft-ietf-spring-segment- routing-policy-22 (work in progress), March 2022. [I-D.ietf-idr-segment-routing-te-policy] Previdi, S., Filsfils, C., Talaulikar, K., Mattes, P., Rosen, E., Jain, D., and S. Lin, "Advertising Segment Routing Policies in BGP", draft- ietf-idr-segment-routing-te-policy-18 (work in progress), June 2022. [I-D.jiang-spring-parent-sr-policy-use-cases] Jiang, W., Cheng, W., Lin, C. and Qiu, Y., "Use Cases for Parent SR Policy", draft-jiang-spring-parent-sr-policy-use-cases-00 (work in progress), July 2022. [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., Decraene, B., Litkowski, S., and R. Shakir, "Segment Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, July 2018, . 8.2. Informative References TBD 9. Acknowledgments The authors would like to thank the following for their valuable contributions of this document: TBD Yang, et al. Expires April, 2023 [Page 12] Internet-Draft SR Policy Intelligent Routing Method October 2022 Authors' Addresses Feng Yang China Mobile Beijing China Email: yangfeng@chinamobile.com Changwang Lin New H3C Technologies Beijing China Email: linchangwang.04414@h3c.com Yuanxiang Qiu New H3C Technologies Email: qiuyuanxiang@h3c.com Yang, et al. Expires April, 2023 [Page 13]