[j-nsp] OSPF reference-bandwidth 1T

[Saku Ytti]

On Thu, 24 Jan 2019 at 10:57, Mark Tinka <mark.tinka at seacom.mu> wrote:

> > And you have shorter paths with inferior bandwidth which you do not
> > want to use, you'll rather take 9x10GE links than 1xGE to reach the
> > destination? It boggles my mind which network has _common case_ where
> > bandwidth is most indicative of best SPT.
>
> In some economies, shorter paths can be more expensive than longer ones.

I don't disagree, I just disagree that there are common case where
bandwidth is most indicative of good SPT.

Consider I have

10GE-1:
PE1 - P1 - P2 - P3 - P4 - P5 - P6 - P7 - P8 - PE2

10GE-2:
PE1 - P1 - P2 - P3 - P4 - P5 - P6 - P7 - P8 - P9 - PE2

10GE-3:
PE1 - P1 - P2 - P3 - P4 - P5 - P6 - P7 - P8 - P9 - P10 - PE2

1GE:
PE1 - PE2

In which realistic topology

a) in 10GE-1 + 1GE, I want to prefer the 10GE between PE?
b) in 10GE-2 + 1GE, I want to balance between the paths
c) in 10GE-3 + 1GE, I want to prefer the 1GE

All these seem nonsensical, what actually is meant '1GE has role Z,
10GE has role X, have higher metric for role Z', regardless what the
actual bandwidth is. I just happens that bandwidth approximates role
in that topology, but desired topology is likely achieved with
distance vector or simple role topology and bandwidth is not relevant
information.

Even when if the P boxes are in same pop, each device adds some 5-10km
of latency. So we'd prefer 40-80km latency over direct connection. Why
did that direct 1GE exist when would you realistically fall back to
using the lower bandwidth link? It seems ridiculously arbitrary and
not indicative of any design.

I'd like to see mock-up topology, where bandwidth metric makes any
sense at all, and is more frequently right than role or latency.

-- 
  ++ytti