| MPLS Working Group | R. Gandhi, Ed. |
| Internet-Draft | C. Filsfils |
| Intended status: Standards Track | Cisco Systems, Inc. |
| Expires: September 7, 2020 | D. Voyer |
| Bell Canada | |
| S. Salsano | |
| Universita di Roma "Tor Vergata" | |
| M. Chen | |
| Huawei | |
| March 6, 2020 |
Performance Measurement for Segment Routing Networks with MPLS Data Plane
draft-gandhi-mpls-rfc6374-sr-02
Segment Routing (SR) leverages the source routing paradigm. RFC 6374 specifies protocol mechanisms to enable the efficient and accurate measurement of packet loss, one-way and two-way delay, as well as related metrics such as delay variation in MPLS networks using probe messages. This document utilizes these mechanisms for Performance Delay and Loss Measurements in Segment Routing networks with MPLS data plane (SR-MPLS), for both SR Links and end-to-end SR Policies. In addition, this document defines Return Path TLV for two-way performance measurement and Block Number TLV for loss measurement.
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). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.
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."
This Internet-Draft will expire on September 7, 2020.
Copyright (c) 2020 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 (https://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 Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Service provider's ability to satisfy Service Level Agreements (SLAs) depend on the ability to measure and monitor performance metrics for packet loss and one-way and two-way delay, as well as related metrics such as delay variation. The ability to monitor these performance metrics also provides operators with greater visibility into the performance characteristics of their networks, thereby facilitating planning, troubleshooting, and network performance evaluation.
Segment Routing (SR) leverages the source routing paradigm and greatly simplifies network operations for Software Defined Networks (SDNs). SR is applicable to both Multiprotocol Label Switching (SR-MPLS) and IPv6 (SRv6) data planes. SR takes advantage of the Equal-Cost Multipaths (ECMPs) between source and transit nodes, between transit nodes and between transit and destination nodes. SR Policies as defined in [I-D.ietf-spring-segment-routing-policy] are used to steer traffic through a specific, user-defined paths using a stack of Segments. Built-in SR Performance Measurement (PM) is one of the essential requirements to provide Service Level Agreements (SLAs).
[RFC6374] specifies protocol mechanisms to enable the efficient and accurate measurement of performance metrics in MPLS networks using probe messages. The One-Way Active Measurement Protocol (OWAMP) defined in [RFC4656] and Two-Way Active Measurement Protocol (TWAMP) defined in [RFC5357] provide capabilities for the measurement of various performance metrics in IP networks. However, mechanisms defined in [RFC6374] are more suitable for Segment Routing when using MPLS data plane (SR-MPLS). [RFC6374] also supports "direct mode" Loss Measurement (LM), which is required in SR networks.
[RFC7876] specifies the procedures to be used when sending and processing out-of-band performance measurement probe replies over an UDP return path when receiving RFC 6374 based probe queries. These procedures can be used to send out-of-band PM replies for both SR-MPLS Links and Policies [I-D.ietf-spring-segment-routing-policy] for one-way measurement.
This document utilizes the probe-based mechanisms defined in [RFC6374] for Performance Delay and Loss Measurements in SR networks with MPLS data plane, for both SR Links and end-to-end SR Policies. In addition, this document defines Return Path TLV for two-way performance measurement and Block Number TLV for loss measurement. The Performance Measurements (PM) for SR Links are used to compute extended Traffic Engineering (TE) metrics for delay and loss and can be advertised in the network using the routing protocol extensions.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.
ACH: Associated Channel Header.
DM: Delay Measurement.
ECMP: Equal Cost Multi-Path.
G-ACh: Generic Associated Channel (G-ACh).
GAL: Generic Associated Channel (G-ACh) Label.
LM: Loss Measurement.
MPLS: Multiprotocol Label Switching.
NTP: Network Time Protocol.
PM: Performance Measurement.
PSID: Path Segment Identifier.
PTP: Precision Time Protocol.
SID: Segment ID.
SL: Segment List.
SR: Segment Routing.
SR-MPLS: Segment Routing with MPLS data plane.
TC: Traffic Class.
TE: Traffic Engineering.
URO: UDP Return Object.
In the reference topology shown in Figure 1, the sender node R1 initiates a performance measurement probe query and the responder node R5 sends a probe response for the query message received. The probe response is typically sent back to the sender node R1. The nodes R1 and R5 may be directly connected via a Link enabled with Segment Routing or there exists a Point-to-Point (P2P) SR Policy [I-D.ietf-spring-segment-routing-policy] on node R1 with destination to node R5. In case of Point-to-Multipoint (P2MP), SR Policy originating from source node R1 may terminate on multiple destination leaf nodes [I-D.voyer-spring-sr-replication-segment].
+-------+ t1 Query t2 +-------+
| | - - - - - - - - - ->| |
| R1 |---------------------| R5 |
| |<- - - - - - - - - - | |
+-------+ t4 Response t3 +-------+
Sender Responder
Figure 1: Reference Topology
One-way delay and two-way delay measurement procedure defined in Section 2.4 of [RFC6374] are used. Transmit and Receive packet loss measurement procedures defined in Section 2.2 and Section 2.6 of [RFC6374] are used. One-way loss measurement provides receive packet loss whereas two-way loss measurement provides both transmit and receive packet loss. For both SR Links and end-to-end SR Policies, no PM session for delay or loss measurement is created on the responder node R5 [RFC6374].
For Performance Measurement, probe query and response messages are sent as following:
The In-Situ Operations, Administration, and Maintenance (IOAM) mechanisms for SR-MPLS defined in [I-D.gandhi-mpls-ioam-sr] are used to carry PM information in-band as part of the data traffic packets, and are outside the scope of this document.
As described in Section 2.9.1 of [RFC6374], MPLS PM probe query and response messages flow over the MPLS Generic Associated Channel (G-ACh). A probe message for SR-MPLS Links contains G-ACh Label (GAL) (with S=1). The GAL is followed by an Associated Channel Header (ACH), which identifies the message type, and the message payload following the ACH as shown in Figure 2. The probe messages are routed over the SR Links for both delay and loss measurement. For SR-MPLS Links, the TTL value is set to 1 in the SR-MPLS header for one-way and two-way measurement modes.
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | GAL (value 13) | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 1|Version| Reserved | GAL Channel Type | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 2: Probe Message Header for an SR-MPLS Link
As described in Section 2.9.1 of [RFC6374], MPLS PM probe query and response messages flow over the MPLS Generic Associated Channel (G-ACh). A probe message for an end-to-end measurement for SR Policy contains SR-MPLS label stack [I-D.ietf-spring-segment-routing-policy], with the G-ACh Label (GAL) at the bottom of the stack (with S=1). The GAL is followed by an Associated Channel Header (ACH), which identifies the message type, and the message payload following the ACH as shown in Figure 3. For SR-MPLS Policies, the TTL value is set to 255 in the SR-MPLS header.
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Label(1) | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ . . . . . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Label(n) | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | GAL (value 13) | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 1|Version| Reserved | GAL Channel Type | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 3: Probe Message Header for an End-to-end SR-MPLS Policy
The SR-MPLS label stack can be empty (as shown in Figure 2) to indicate Implicit NULL label case.
For SR Policy performance measurement, in order to ensure that the probe query message is processed by the intended responder node, Destination Address TLV (Type 129) [RFC6374] can be sent in the probe query message. The responder node only replies with Success in Control Code if it is the intended destination for the probe query. Otherwise, it MUST return 0x15: Error - Invalid Destination Node Identifier.
In one-way performance measurement mode [RFC7679], the PM sender node can receive "out-of-band" probe replies by properly setting the UDP Return Object (URO) TLV in the probe query message. The URO TLV (Type=131) is defined in [RFC7876] and includes the UDP-Destination-Port and IP Address. In particular, if the sender sets its own IP address in the URO TLV, the probe response is sent back by the responder node to the sender node. In addition, the "control code" in the probe query message is set to "out-of-band response requested". In this delay measurement mode, as per Reference Topology, timestamps t1 and t2 are collected by the probes. Only timestamps t1 and t2 are used to measure one-way delay. The one-way mode is applicable to both SR-MPLS Links and SR-MPLS Policies.
In two-way performance measurement mode [RFC6374], when using a bidirectional path, the probe response message is sent back to the sender node on the congruent path of the data traffic on the reverse direction SR Link or associated SR Policy [I-D.ietf-pce-sr-bidir-path] using a message with format similar to their probe query message. In this case, the "control code" in the probe query message is set to "in-band response requested". In this delay measurement mode, as per Reference Topology, all timestamps t1, t2, t3, and t4 are collected by the probes. All four timestamps are used to measure two-way delay. The two-way mode is applicable to both SR-MPLS Links and SR-MPLS Policies.
Specifically, the probe response message is sent back on the incoming physical interface where the probe query message is received. This is useful for example, in case of two-way measurement mode for Link delay.
The Path Segment Identifier (PSID) [I-D.ietf-spring-mpls-path-segment] of the forward SR-MPLS Policy in the probe query can be used to find the associated reverse SR Policy [I-D.ietf-pce-sr-bidir-path] to send the probe response message for two-way measurement of SR-MPLS Policy unless when using the Return Path TLV.
The Loopback measurement mode defined in Section 2.8 of [RFC6374] can be used to measure round-trip delay for a bidirectional SR Path [I-D.ietf-pce-sr-bidir-path]. The probe query messages in this case carries the reverse SR Path label stack as part of the MPLS header. The GAL is still carried at the bottom of the label stack (with S=1). The responder node does not process the PM probe messages and generate response messages. In this delay measurement mode, as per Reference Topology, the timestamps t1 and t4 are collected by the probes. Both these timestamps are used to measure round-trip delay. The loopback mode for SR-MPLS Links is outside the scope of this document.
For two-way performance measurement, the responder node needs to send the probe response message on a specific reverse path. The sender node can request in the probe query message to the responder node to send a response message back on a given reverse path (e.g. co-routed path for two-way measurement). This way the destination node does not require any additional SR Policy state.
For one-way performance measurement, the sender node address may not be reachable via IP route from the responder node. The sender node in this case needs to send its reachability path information to the responder node.
[RFC6374] defines DM and LM probe query messages that can include one or more optional TLVs. New TLV Type (TBA1) is defined in this document for Return Path to carry reverse path for probe response messages (in the payload of the message). The format of the Return Path TLV is shown in Figure 4 and Figure 5:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type = TBA1 | Length | Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Return Path Sub-TLVs | . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 4: Return Path TLV
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Label(1) | . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ . . . . . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Label(n) | . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 5: Segment List Sub-TLV in Return Path TLV
The Segment List Sub-TLV in the Return Path TLV can be one of the following Types:
The Return Path TLV is Mandatory when used. If responder does not support this TLV, it MUST return Error 0x17: Unsupported Mandatory TLV Object. The PM sender node MUST only insert one Return Path TLV in the probe query message and the responder node MUST only process the first Return Path TLV in the probe query message and ignore other Return Path TLVs if present. The responder node MUST send probe response message back on the reverse path specified in the Return Path TLV and MUST NOT add Return Path TLV in the probe response message.
As defined in [RFC6374], MPLS DM probe query and response messages use Associated Channel Header (ACH) (value 0x000C for delay measurement) [RFC6374], which identifies the message type, and the message payload following the ACH. For both SR Links and end-to-end measurement for SR-MPLS Policies, the same MPLS DM ACH value is used.
The DM message payload as defined in Section 3.2 of [RFC6374] is used for SR-MPLS delay measurement, for both SR Links and end-to-end SR Policies.
The Section 3.4 of [RFC6374] defines timestamp format that can be used for delay measurement. The IEEE 1588 Precision Time Protocol (PTP) timestamp format [IEEE1588] is used by default as described in Appendix A of [RFC6374], with hardware support in Segment Routing networks.
The LM protocol can perform two distinct kinds of loss measurement as described in Section 2.9.8 of [RFC6374].
For both of these modes of LM, Path Segment Identifier (PSID) [I-D.ietf-spring-mpls-path-segment] is used for accounting received traffic on the egress node of the SR-MPLS Policy as shown in Figure 6. Different values of PSID can be used to measure packet loss per SR-MPLS Policy, per Candidate Path or per Segment List of the SR Policy.
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | PSID | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | GAL (value 13) | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 1|Version| Reserved | GAL Channel Type | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 6: With Path Segment Identifier for SR-MPLS Policy
As defined in [RFC6374], MPLS LM probe query and response messages use Associated Channel Header (ACH) (value 0x000A for direct loss measurement or value 0x000B for inferred loss measurement), which identifies the message type, and the message payload following the ACH. For both SR Links and end-to-end measurement for SR-MPLS Policies, the same MPLS LM ACH value is used.
The LM message payload as defined in Section 3.1 of [RFC6374] is used for SR-MPLS loss measurement, for both SR Links and end-to-end SR Policies.
The Loss Measurement using Alternate-Marking method defined in [RFC8321] requires to color the data traffic. To be able to compare the transmit and receive traffic counters of the matching color, the Block Number (or color) of the traffic counters is carried by the probe query and response messages for loss measurement. Probe query and response messages specified in [RFC6374] for Loss Measurement do not identify the Block Number of the counters.
[RFC6374] defines probe query and response messages that can include one or more optional TLVs. New TLV Type (value TBA2) is defined in this document to carry the Block Number (8-bit) of the traffic counters in the probe query and response messages for loss measurement. The format of the Block Number TLV is shown in Figure 7:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type TBA2 | Length | Reserved | Block Number | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 7: Block Number TLV
The Block Number TLV is Mandatory when used. If responder does not support this TLV, it MUST return Error 0x17: Unsupported Mandatory TLV Object. The PM sender node SHOULD only insert one Block Number TLV in the probe query message and the responder node in the probe response message SHOULD return the first Block Number TLV from the probe query messages and ignore other Block Number TLVs if present. In probe messages, the counters MUST belong to the same Block Number.
The procedures for delay and loss measurement described in this document for Point-to-Point (P2P) SR-MPLS Policies [I-D.ietf-spring-segment-routing-policy] are also equally applicable to the Point-to-Multipoint (P2MP) SR-MPLS Policies as following:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Replication SID | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | GAL (value 13) | TC |S| TTL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 1|Version| Reserved | GAL Channel Type | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 8: Query with Replication Segment for SR-MPLS Policy
An SR Policy can have ECMPs between the source and transit nodes, between transit nodes and between transit and destination nodes. Usage of Anycast SID [RFC8402] by an SR Policy can result in ECMP paths via transit nodes part of that Anycast group. The PM probe messages need to be sent to traverse different ECMP paths to measure performance delay of each of the ECMP path of an SR Policy.
Forwarding plane has various hashing functions available to forward packets on specific ECMP paths. For SR-MPLS Policy, sweeping of entropy label [RFC6790] values can be used in PM probe messages to take advantage of the hashing function in forwarding plane to influence the ECMP path taken by them.
The considerations for performance loss measurement for different ECMP paths of an SR Policy are outside the scope of this document.
The extended TE metrics for SR Link delay and loss computed using the performance measurement procedures described in this document can be advertised in the routing domain as follows:
This document describes the procedures for performance delay and loss measurement for SR-MPLS networks, for both SR Links and end-to-end SR Policies using the mechanisms defined in [RFC6374] and [RFC7876]. This document does not introduce any additional security considerations other than those covered in [RFC6374], [RFC7471], [RFC8570], [RFC8571], and [RFC7876].
IANA is requested to allocate a value for the following mandatory Return Path TLV Type for RFC 6374 to be carried in PM probe query messages:
IANA is requested to allocate the values for the following Sub-TLV Types for the Return Path TLV for RFC 6374.
IANA is also requested to allocate a value for the following mandatory Block Number TLV Type for RFC 6374 to be carried in the PM probe query and response messages for loss measurement:
| [RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997. |
| [RFC6374] | Frost, D. and S. Bryant, "Packet Loss and Delay Measurement for MPLS Networks", RFC 6374, DOI 10.17487/RFC6374, September 2011. |
| [RFC7876] | Bryant, S., Sivabalan, S. and S. Soni, "UDP Return Path for Packet Loss and Delay Measurement for MPLS Networks", RFC 7876, DOI 10.17487/RFC7876, July 2016. |
| [RFC8174] | Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, May 2017. |
The authors would like to thank Thierry Couture for the discussions on the use-cases for the performance measurement in segment routing networks. Authors would like to thank Patrick Khordoc for implementing the mechanisms defined in this document. The authors would like to thank Greg Mirsky for providing many useful comments and suggestions. The authors would also like to thank Stewart Bryant, Sam Aldrin, Tarek Saad, and Rajiv Asati for their review comments.
Sagar Soni Cisco Systems, Inc. Email: sagsoni@cisco.com Zafar Ali Cisco Systems, Inc. Email: zali@cisco.com Pier Luigi Ventre CNIT Italy Email: pierluigi.ventre@cnit.it