Simple Two-Way Active Measurement Protocol Extensions for Performance Measurement on LAG

Link Aggregation Group (LAG), as defined in , provides mechanisms to combine multiple physical links into a single logical link. This logical link offers higher bandwidth and better resiliency, because if one of the physical member links fails, the aggregate logical link can continue to forward traffic over the remaining operational physical member links. Usually, when forwarding traffic over a LAG, a hash-based or similar mechanism is used to load-balance the traffic across the LAG member links. In some cases, the link delays of the member links are different because they are over different transport paths. To provide low delay service to time-sensitive traffic, we have to know the link delay of each member link of a LAG and then steer traffic accordingly. That requires a solution that could measure the performance metrics of each member link of a LAG. However, when using Simple Two-Way Active Measurement Protocol (STAMP) to measure a LAG's performance, the LAG is treated as a single logical link/path. The measured metrics reflect the performance of one member link or an average of some/all member links of the LAG. In addition, for LAG, using passive or hybrid methods (like alternative marking or iOAM ) can only monitor the link crossed by traffic. It means that the measured metrics reflect the performance of some member links or an average of some/all member links of the LAG. Therefore, in order to measure every link of a LAG, using active methods would be more appropriate. This document defines extensions to STAMP to implement performance measurement on every member link of a LAG.

This document intends to address the scenario (e.g., ) where a LAG (e.g., the LAG includes three member links) directly connects two nodes (A and B) . The goal is to measure the performance of each link of the LAG.

To measure performance metrics of every member link of a LAG, multiple sessions (one session for each member link) need to be established between the two hosts that are connected by the LAG. These sessions are called micro-sessions for the remainder of this document. All micro-sessions of a LAG share the same Sender Address, Receiver Address. As for the Sender Port and Receiver Port, the micro-sessions may share the same Sender Port and Receiver Port pair, or each micro session is configured with a different Sender Port and Receiver Port pair. But from simplifying operation point of view, the former is recommended. In addition, with micro-sessions, there needs a way to correlate a session with a member link. For example, when the Reflector receives a Test packet, it needs to know from which member link the packet is received, and correlate it with a micro-session. That is a new functionality for STAMP as defined in and . Upon receiving a Test packet for a micro-session, the receiver uses the receiving link's identifier to correlate the packet to a particular session. In addition, Test packets may need to carry the member link information for validation checking. For example, when a Session-Sender/Session-Reflector receives a Test packet, it may need to check whether the Test packet is from the expected member link.

The micro-STAMP-Test protocol is based on the STAMP-Test protocol and with the following extensions.

The LAG Member Link ID TLV is defined to carry the LAG member link identifiers associated with a micro-STAMP session. The member link identifiers are used for the Session-Sender and Session-Reflector to check whether a Test packet is received from the expected member link. The detailed procedures are defined in Section 3.1.2. The format is as below:

Sender Member Link ID (2-octets in length): it is defined to carry the LAG member link identifier of the Sender side. The value of the Sender Member Link ID MUST be unique at the Session-Sender. Reflector Member Link ID (2-octets in length): it is defined to carry the LAG member link identifier of the Reflector side. The value of the Reflector Member ID MUST be unique at the Session-Reflector.

The micro-STAMP-Test reuses the procedures as defined in Section 4 of STAMP with the following additions: The micro-STAMP Session-Sender MUST send the micro-STAMP-Test packets over the member link associated with the session. The micro STAMP Session-Reflector MUST send the reflected Test packets over the receiving member link. The configuration and management of the association between a micro-STAMP session and the Sender/Reflector member link identifiers are outside the scope of this document. When the Session-Sender sends a Test packet, the LAG Member Link ID TLV MUST be carried, and the Sender member link identifier associated with the micro-STAMP session MUST be put in the Sender Member Link ID field. If the Session-Sender knows the Reflector member link identifier, it MUST set it as the Reflector Member Link ID field's value. Otherwise, the Reflector Member Link ID field MUST be set to zero. The Session-Sender uses the Sender Member Link ID field's value to check whether the reflected Test packet is received from the member link associated with the correct micro-STAMP session. Therefore, the Session-Reflector MUST copy the Sender Member Link ID value to the reflected Test packet. The Session-Reflector uses the Reflector Member Link ID value to check whether a Test packet is received from the member link associated with the correct micro-STAMP session. The Reflector member link identifier can be obtained from pre-configuration or learned through the data plane (e.g., learned from a reflected Test packet). How to obtain/learn the Reflector member link identifier is outside of this document's scope. When the micro-STAMP Session-Reflector receives a Test packet, it MUST use the receiving member link to correlate the Test packet to a micro-STAMP session. If there is no such a micro-STAMP session, the Test packet MUST be discarded. Suppose the Reflector Member Link ID is not zero. In that case, the micro-STAMP Session-Reflector MUST use the Reflector member link identifier to check whether it is associated with the micro-STAMP session. If it is not, the Test packet MUST be discarded and no reflected Test packet will be sent back to the Session-Sender. If all validation passed, the Session-Reflector sends a reflected Test packet to the Session-Sender over the receiving member link. The micro-STAMP Session-Reflector MUST put the Sender and Reflector member link identifiers associated with the micro-STAMP session in the Sender Member Link ID and Reflector Member Link ID fields, respectively. The Sender member link identifier is copied from the received Test packet. When the micro-STAMP Session-Sender receives a reflected Test packet, it MUST use the receiving member link to correlate the reflected Test packet to a micro-STAMP session. If there is no such a session, the reflected Test packet MUST be discarded. If a matched micro-STAMP session exists, the Session-Sender MUST use the identifier carried in the Sender Member Link ID field to check whether it associates with the session. If the checking failed, the Test packet MUST be discarded.

This document requires the IANA to allocate the following the TLV type from the "STAMP TLV Types" sub-registry.

This document does not introduce additional security requirements and mechanisms other than the ones described in apply to this document.

The authors would like to thank Min Xiao, Fang Xin for the valuable comments to this work.