>> Sorry, but it is always true that by removing information you always
>> lose optimality (you increase stretch). Whether that removal is done at
>> the edge or in the core, the result is always the same. There are two
>> ironclad rules of routing:
>>
>> - Removing information decreases optimal routing.
> 
> No.

Yes. :-)

In fact, one of the points people are arguing here is, "in these
circumstances, the stretch increases."

1. Under normal operation, this proposal doesn't increase stretch (that
I know of).
2. Under broken conditions, any proposal may (probably will, in fact)
increase stretch.
3. In fact, normal aggregation increases stretch, so...

> Also the critical piece is that the determination when and what should
> be removed must be 100% automated. There is no way where manual
> intervention would be required for any scheme to work right. And your
> proposal does require manual NOC intervention as you have already
> confirmed today.

Robert --methinks you're putting words in my mouth. Beyond the basic
configuration, there's no "manual intervention." If you consider
configuring the option to be "manual intervention," then we have a fully
manual routing system already in place. :-)

> The proposal from Pedro however is automated, adjust itself to
> topology changes as well as removed only redundant information where
> such removal does not impact a single bit routing optimality.

This is a very simplified version of the proposal Pedro made --in fact,
this proposal actually predates Pedro's. I like Pedro's proposal, but I
consider the two different solutions parallel, rather than opposed.
There's no reason one AS couldn't implement one, and another the other
--it really isn't going to matter in the long run.

:-)

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

Reply via email to