Re: Mail regarding draft-litkowski-rtgwg-spf-uloop-pb-statement

[Mike Shand <[e-mail]>]

I have been assigned as Routing Directorate QA reviewer for this document.

The following web page contains a briefing on the QA process.

​https://trac.tools.ietf.org/area/rtg/trac/wiki/RtgDirDocQa


The document provides a useful summary of the issues leading to 
micro-loop formation, especially in mixed vendor networks and provides 
examples of the possibilities to adversely affect the the micro-loop 
behaviour by inconsistent choices of parameters and algorithms among the 
routers in the network.

I confess that I find the timing diagrams quite hard to follow. One of 
the penalties of being constrained by the need to use ASCII art. But I 
wonder if something could be done to make them a little easier to follow 
without losing the essential information?

There are some very simple mitigation techniques, such as delaying the 
locally triggered  SPF/FIB installation more than a remotely triggered 
one, which it may be helpful to mention.

The document goes on to propose future work to standardise some behaviours.

Clearly this work is at an early stage and the trade-offs between 
standardisation and allowing vendors freedom to innovate for the benefit 
of their customers must be carefully considered.

The document seems like a good starting point for this work.


In reading the document I spotted a few items which it would be as well 
to address.

General:

1. micro-loop or microloop. The terminology is used inconsistently. RFC 
5715 uses micro-loop
2. There are numerous instances of awkward usage of english. It would be 
helpful to address these at some stage.

Nits:

3. "   We will call SPF delay, the delay timer that exists in most
    implementations that makes codes to wait before running SPF
    computation after a SPF trigger is received."

The phrase "makes codes to wait" is somewhat contrived. How about "that 
specifies the required delay"?

4. "   Routers have more and more powerful controlplane and dataplane that
    reduce the Control plane to Forwarding plane overhead during the
    convergence process.  Even if FIB update is still reasonably the
    highest contributor in the convergence time for large network, its
    duration is reducing more and more and may become comparable to
    protocol timers.  This is particular true in small and medium
    networks."

I don't understand what is meant by "may become comparable to protocol 
timers"? Are you suggesting that the FIB update latency WAS greater than 
the protocol timers, but has now been reduced to a comparable value?

The reference to small and medium networks is also confusing, since in 
my experience it is actually the small and medium networks which are 
subject to the LARGEST FIB update times as a result of the deployment of 
under powered hardware.


5. "   In multi vendor networks, using different implementations of a link
    state protocol may favor micro-loops creation during convergence time
    due to deprecancies of timers."

deprecancies? Do you mean discrepancies?

6. "4.2 Exponential Backoff"

"   o  First delay : amount of time to wait before running SPF. This
       delay is used on when SPF is in fast mode."

I assume "is used only when SPF" is what you mean.

and similarly in the next bullet



Mike