[j-nsp] Spine & leaf

[Mark Tinka]

On 25/Jun/18 19:37, Scott Whyte wrote:

> In balance then, we have better filtering versus less config, which
> has already been noted can (must) be completely automated.  Where
> one's shop is on the NetDevOps curve probably has a lot of impact on
> the decision, which is unfortunate.

If you have proper IP addressing design, you don't need to worry about
filtering infrastructure routes that you know should never leave your AS.

I think the last time I ever played with filtering on an IGP was when I
was under the impression L1 and L2 levels in IS-IS were a good thing a
la Area 0, Area 1, e.t.c. in OSPF. That was the IS-IS Route Leaking
feature, which turned out to be a totally useless exercise for 3 years,
since MPLS still required /32 IPv4 addresses for stitch LSP's together
anyway.

In the end, I went to a flat L2 design for IS-IS and have never looked back.

Agree that BGP offers plenty of filtering opportunities, but if you have
full control of your internal infrastructure routing, and unless you're
limiting which of your devices sees which addresses (either due to
service design or hardware constraints), why would you want to filter
the global infrastructure view each device needs from the other?
Wouldn't you end up running the risk of partitioning your internal
infrastructure network (again, unless you have specific service design
or hardware limitation concerns to deal with)?

Mark.