[c-nsp] New Catalyst 6k chassis

[Saku Ytti]

On (2013-06-26 23:30 +0000), Dobbins, Roland wrote:

> > But can cisco afford to have three quite similar product lines,
> > that are expensive to maintain? 
> 
> Cisco isn't really a unitary company, it's a loose confederation of semi-feudal fifedoms, each with its own P&L.   Effectively, they're separate companies utilizing a common branding/marketing framework and shared administrative resources.

This has overhead and I can understand why some people don't like having
competing products in the market.

But I still think it's overall positive thing and not something you should
change. I don't want to touch the technical merits of current portfolio (I
fully believe each nexus7k, cat6k and ASR9k have market, but I understand
how it can lead to a debate).

What this situations gives to Cisco is ability to change, as dropping the
ball does not mean your entire portfolio failed. When you have option to
fail, you can be more aggressive with your outlook.

Take Juniper, ok, it's not single JunOS by far, but it's mostly single RPD,
which is extremely naive by modern standards, relying on programmers
diligence on adding enough yields to the code so there can't be use-case
where certain part of code runs long-enough to cause your ISIS or BGP to
flap.
  Of course Juniper has technical competence to fix this, and redesign the
architecture to be much more modern and robust, relying more on proven OS
scheduling, use more than one core, use more than 4GB of memory.
  But as it obviously mostly works, any major redesign is huge risk, which
would need very ambitious manager to root for, failure would be
career-breaking.

Starting new product family form scratch and take the ideas there, prove
them or fail them and backport what works is much more secure way to do it.

Another, similar beneficial point in fragmentation is, that every project
has more competence/ideas than the product can implement, so some people
are left frustrated as they don't have chance to contribute at the level
they want to. With fragmentation you have more surface where talent can
prove their ideas. I believe this is major reason why Linux is such a
success, clearly GPL is inferior license to BSD from business POV, but
because Linux is so fragmented it receives lot more time from top-tier
talent.


-- 
  ++ytti