Juliusz,

I understand that this is a complex feature that would require quite a bit of 
work to get right.  There would also be work involved in standardizing the 
IS-IS autoconf and IS-IS source-specific routing extensions.  In addition, we 
might want/need to standardize a mechanism for running IS-IS over IP.  That 
would be considerable amount of work.

I don't think it is likely to be a lot of work, though, compared to the effort 
to standardize Babel which would include:

- Starting an IETF WG to do the work, and understanding that the WG would have 
change control over Babel.
- Substantial work to bring the spec up to the level of detail/clarity required 
for and IETF protocol.
- Consideration of many IETF requirements, perhaps most notably security 
requirements.
- Finding a second set of people who are willing to do a complete, independent, 
fully-interoperable implementation.

Personally, I think that Babel is an interesting Routing Protocol with 
substantially different applicability than IS-IS, OSPF or BGP, and it might be 
good to standardize it, whether Homenet adopts it or not.  However, I think 
that the time to do that would be a serious problem for Homenet because of the 
time pressure to finalize a standard for Homenet.

Margaret



On Mar 24, 2015, at 2:32 PM, Juliusz Chroboczek 
<[email protected]> wrote:

> Dear all,
> 
> At todays meeting, the claim was made (at least twice) that adding
> dynamically computed metrics to IS-IS is "just a feature".  I strongly
> disagree with this assessment -- it's an open research problem, and
> a difficult one at that.
> 
> Any interesting metric (packet loss, delay, etc.) will cause a negative
> feedback loop, which will lead to oscillations.  In Babel, there are three
> mechanisms that cope with the oscillations caused by feedback loops:
> 
> 1. the protocol is loop-avoiding, which means that even when oscillations
>   happen they don't normally cause packets to be lost;
> 2. the protocol uses delayed updates, which means that even when a metric
>   is varying the amount of network traffic remains controlled;
> 3. the protocol uses a hysteresis mechanism which limits the frequency of
>   oscillations.
> 
> IS-IS is fundamentally based on the notion that a topology change is
> propagated throughout the network in a timely manner and SPF is recomputed
> by all nodes -- it has no loop-avoidance mechanism other then timely
> reconvergence.  If implemented naively, a dynamically computed metric will
> cause repeated flooding, repeated SPF computation, and repeated transient
> loops.
> 
> I'm sure these issues can be solved, and I'm pretty confident that Henning
> can tell you how; at any rate, it would be a very interesting research
> project.  However, it is a difficult one, and the three techniques used in
> Babel do not apply directly to a link-state protocol.
> 
> -- Juliusz
> 
> _______________________________________________
> homenet mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/homenet

_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to