[c-nsp] non-BGP ISP redundancy

[David Barak]

--- RH Lists <lists at 101.net> wrote:

> I have several clients (all via one consultant) who
> swear by their
> Stonegates. These are local branches of "big 4"
> accounting companies,
> insurance companies, large hotels.  So I would
> imagine they work well.

I have encountered Stonegates in a couple of
customers' environments, and each time, they
introduced oodles of problems.  One of the major
issues was that they don't actually speak a routing
language, so unless you're willing to let them make
ALL of your routing decisions, you can end up with
some pretty kludgy environments (like GRE all over the
place to get around them...)
 
> That being said, in an SP environment, I don't think
> there's any reason not
> to use BGP.

Hear, hear!  I would even extend that to the
enterprise environment: the questions which should be
asked (and rarely is) of the non-BGP device vendors is
why their decision algorithms would be better than the
routing decision algorithms - do they have the number
of knobs of BGP?  The determinism and speed of
Dijkstra?  Will they interoperate with other vendors'
equipment?  Is their interdevice communication
protocol well-documented and are the attacks against
it known and patched, or are they relying on security
through obscurity?  What is their roadmap for new
feature support?  (As an example, when I last looked
at Stonegate, they didn't support QoS, and had zero
ability to do VoIP over a WAN, 'cause their added
jitter was terrible.  They may have changed this.)

BGP works, and is well understood.  use it.

David Barak
Need Geek Rock?  Try The Franchise: 
http://www.listentothefranchise.com


		
____________________________________________________
Start your day with Yahoo! - make it your home page 
http://www.yahoo.com/r/hs