Hi,
I have a general question about relationship between SRLG support defined in
the TI-LFA
draft<https://datatracker.ietf.org/doc/html/draft-ietf-rtgwg-segment-routing-ti-lfa-09>
and various flavors of micro-loop avoidance defined, e.g., in RFC
8333<https://datatracker.ietf.org/doc/html/rfc8333> and SR Micro-Loop Avoidance
draft<https://datatracker.ietf.org/doc/html/draft-bashandy-rtgwg-segment-routing-uloop-14>.
(All authors of RFC 8333 are also authors of the TI-LFA draft).
1. The TI-LFA draft:
* Explicitly claims support of protection against a "single SRLG failure"
i. Handling
of SRLG failures defined in the TI-LFA draft means that the repair path is
computed assuming simultaneous failure of all the links in a given single SRLG
ii. "Single
SRLG failure" may result in detection multiple simultaneous (or almost
simultaneous) failure of multiple links by the IGP
* Uses SR to extend the local and remote LFA FRR mechanisms defined in
RFC 5714 and RFC 7490 respectively
i. The
remote LFA mechanisms are well know for potential of micro-loops if the IGP in
the PLR converges faster than IGP in its neighbors
ii. If not
taken care of, these micro-loops can effectively negate any gain provided by
these mechanisms
1. RFC 8333 defines a simple mechanism for preventing this by forcing
delayed IGP convergence at the PLR. However, it explicitly states that it
* Assumes that there has been a single link failure as seen by the IGP
area/level.
* States that "if this assumption is violated (e.g., multiple links or
nodes failed), then regular IP convergence must be applied"
2. The SR Micro-Loop Avoidance draft:
* Neither explicitly states that it is only applicable to a single
network event
* Nor is explicitly restricted to handling just a single network event
* However, at least one production level implementation explicitly
states that:
i. Events
related to two endpoints of the same link are treated as a single network event
ii.
Micro-loop avoidance procedures are discarded if IGP detects multiple network
events
1. Micro-loop avoidance is not explicitly mentioned in the TI-LFA draft.
*The reference to the SR M
I would highly appreciate any clarification of expectations with regard to
handling of micro-loops in the SRLG protection scenarios in TI-LFA draft.
Regards, lots of thanks in advance, and my vbest New Year wishes,
Sasha
Notice: This e-mail together with any attachments may contain information of
Ribbon Communications Inc. and its Affiliates that is confidential and/or
proprietary for the sole use of the intended recipient. Any review, disclosure,
reliance or distribution by others or forwarding without express permission is
strictly prohibited. If you are not the intended recipient, please notify the
sender immediately and then delete all copies, including any attachments._______________________________________________
rtgwg mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/rtgwg
