[c-nsp] Latest iteration of core upgrade - questions

[Mark Tinka]

On Sunday 01 November 2009 03:02:21 am Rick Ernst wrote:

> > ---  It was an "in", but now it's "at".  I can still
> > argue it being
>
> appropriate as a border/"upstream" device and also as
> aggregation/"customer".

You probably want to try separating both functions where 
possible, otherwise your routing policies on a multi-
function box may get too complex (I've been in bad 
situations where border routers had to double as route 
reflectors - not very pretty).

> > --- One 720x per upstream, split into dual cores.

Sounds good.

> > We
> > had also considered
>
> landing upstreams directly on the 7600s, but this allow
> for a core device failure without losing upstream
> capacity.

Again, wherever possible, try separating those functions.

> --- I've looked at other vendors, but a big reason for
> sticking with Cisco is we have the in-house knowledge. 
> Changing vendors while re-architecting a production
> network seems to be a bad idea.

Fair enough - it's always best to go with what you're 
comfortable handling.

> --- What is the benefit in having 4 devices instead of 2?
>  It seems like you'd just be passing the same traffic
> through double the number of devices.

Like I'd said, you'd only grow to 4x (2x for edge + core 
aggregation, and 2x for border + core aggregation) if it 
became necessary. You'd normally find this in PoP's where 
you've got a lot of upstream service concentration, 
typically your flagship PoP when you started operations.

Depending on how many border routers you have (as well as 
what other devices may be sitting at this layer), there may 
be a need for a number of Ethernet ports.

Furthermore, assuming border + edge switch aggregation were 
collapsed into a single device, failure of either would 
affect Internet traffic for customers connecting to the same 
PoP. However, assume traffic to the Internet is coming in 
from another PoP, which connects to your core routers - 
here, a failure of a combined border + edge core switch 
affects both the local and remote PoP's. If you had 2x core 
switches dedicated for your border + core aggregation, 
remote PoP's would still have Internet access assuming the 
main PoP was their exit to the rest of the world.

Again, these are all dynamics respective to an individual 
business. As I'd mentioned, it's typically considered only 
where necessary.

> -- I had actually considered another pair of 7600s at the
> aggregation layer, but we currently have ~300 ports in
> use and the cable management is a nightmare.   The 4948s
> let us to a "top-of-rack" design and run back to the
> core.  We could have done the same thing with a pair of
> 7600s and dumb layer-2 switches, but using the 4948s
> allows incremental upgrades/migration.

Understand - this where I think the Nexus 7000 series may 
excel.

Cheers,

Mark.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: This is a digitally signed message part.
URL: <https://puck.nether.net/pipermail/cisco-nsp/attachments/20091102/dce26633/attachment.bin>