Core router bakeoff?

John A. Tamplin jat at
Thu May 7 22:03:28 UTC 1998

On Thu, 7 May 1998, Perry E. Metzger wrote:

> Arnaud Girsch writes:
> > > You don't want Bay and you certainly don't want 3Com. If your network
> > > is fairly slow (ethernets and T1s only) you can use PCs running a
> > > reasonable BSD and GateD. Otherwise, the only commercial choice is
> > > Cisco.
> > 
> > Just out of curiosity ...
> > Why not considering the 4th vendor, Cabletron, for this kind of equipment,
> > before using PCs.
> PCs are cheap and I know them well. I wasn't aware Cabletron even had
> a box with a BGP-4 implementation in it.

PCs are also designed with a mindset that saving $.10 on a component saves
millions, encouraging overly cheap designs.  Considering the typical PC
customer has no problem with rebooting their machine several times a day,
that gives them plenty of room to cut corners without pissing off their
primary client base.  This is not to say that you can't build solid hardware,
but the typical PC vendors simply do not have a level of quality sufficient
for 24x7 operation.

"Cheap" often winds up "expensive" when you count the cost of downtime.  We
run all Cisco routers and have had exactly one failure on any box in 4 years
(a power supply in a 4500).  While the software quality has gotten worse
lately and you have to be careful selecting which code to run, the software
has generally been as stable as the hardware.

> > They just released some announcement that major companies and campuses are
> > using their equipment for new backbones.
> That sounds like salesspeak. I've used lots of cabletron hubs and such 
> over the years, but they never seemed to have real routers.

As far as I know, Cabletron's router blades for their switches are just
Cisco 4500's with one of the NPM slots tied to the backplane.  I assume you
could run BGP on it, although performance might not be good enough.

John Tamplin					Traveller Information Services
jat at Traveller.COM				2104 West Ferry Way
205/883-4233x7007				Huntsville, AL 35801

More information about the NANOG mailing list