Re: [Idr] draft-uttaro-idr-bgp-persistence-00

[<[e-mail]>]

>> Sure. But given the small number of RR boxes, and the high
requirements
>> on these boxes, it's harder to justify a two or three vendor
sourcing.
>
>I would think just the very opposite. Small number of RRs and their
high
>requirements IMHO fully justify more then one vendor sourcing.

Somehow we disagree on this. 
Regarding the small number of RR, if you need 10 RR, it's hard to go to
the buying, testing, integration and ops teams asking for a new vendor
for 5 only boxes.
Regarding dual sourcing, to keep redundancy, you would need to restrict
yourself to the smaller intersection of features and performance. This
can be limiting.


>Of course fun starts where it is hard to find such diverse RRs with all
>required feature sets.

My point. Both features and scale.

> Those concerned about such issues are looking
>actively in solving the problem fundamentally at it's roots and not
>cover it with yet another curtain.

Dual sourcing helps but does not guaranty that both RR will never fail.
You said it yourself. So this does not solve the fundamental problem.

However, again, given the low probability of double failure, it's harder
to justify work on this.

>Cheers,
>R.