[c-nsp] Upgrading to 40G

[Mark Tinka]

On Saturday, March 01, 2014 02:39:07 PM Phil Mayers wrote:

> My view is that the real problem is treating ops and
> development as two different things done by two separate
> groups of people - a pervasive attitude in the IT
> industry as a whole, that leads to un-operable crap
> being thrown over the wall from development, while at
> the same time development has no idea what ops needs.

And this creates a huge gap.

Folk that translate RFC's into code may not necessarily 
understand how their choice of implementation affects 
network operation.

Some vendors have BU's that monitor these lists, and when 
all the planets align, some actually go out and ask whether 
implementing a feature (a certain way) has operational value 
or not. This happens less often than it should, but it shows 
that there is a gap that needs filling..

Operators have, for years, tried to get their word into 
vendors. And the usual answer - "Show me the money". One 
poster on a competing vendor's -nsp list suggested that 
requests for new features and capabilities should be posted 
via a web site, rather than trying to channel these through 
your AM, because more than likely, those go into /dev/null 
for the majority of operators.

Are operators talking more to their vendors than they are 
the IETF? Yes. But some may argue that the IETF are mostly 
vendors, so...

This is not an easy problem to solve - and if vendors 
continue to use the "AM's-go-and-find-out-how-many-of-your-
customers-have-requested-this-feature" or "how-large-is-
your-customer's-deal" approach, we'll never fix this 
problem. I appreciate that vendors have finite resources as 
to do other areas in life, but we also can't ignore the 
problem entirely.

Mark.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
URL: <https://puck.nether.net/pipermail/cisco-nsp/attachments/20140301/beca20ec/attachment.sig>