Internet Core Routing - Ethernet

Stephen Sprunk ssprunk at cisco.com
Mon Sep 30 15:30:03 UTC 2002


Thus spake "Bob Martinez" <bobmartinezzz at hotmail.com>
> >Hmm... so if somebody posts to the list with the problem, and somebody else
> >saw that same issue and got a fix from the vendor, they need to send it to
the
> >vendor for comment, or they can say "Oh, you're being bit by bug (can't say
> >because it would identify the vendor) in a (vendor model you can't say)
> >several hops upstream from you".
>
> Is this a problem?  Not on my team.

When, say, every ISP in the world is bouncing half of its BGP links every time
they come up because vendor A's bug aggravates vendor B's bug, it's helpful to
know what the problem is.  You might have encountered this in your own lab, or
magically figured it out before the thousands of others on NANOG seeing the same
problem, but odds are you won't.

If vendor information is relevant to the discussion, it has merit.  Vendor
information purely for marketing or bashing is not; this is long established on
NANOG and seems to work well.

> >And how did we learn the names?  Let's see.. Cisco, Juniper, Proteon, Bay,
>
> Dude, this is exactly what I'm talking about.  You didn't mention my vendor.

Wah.  If you choose to use unpopular hardware, that's your choice.  Cisco and
Juniper account for around 98% of the Internet, so don't pretend you can ignore
the implications of that.

> >Now, did you actually *buy* and *use* all of that gear yourself?
>
> Friends in the community are my most trusted resource.

NANOG is a lot of people's link to the "community".

> No mistakes to report, sir.  NANOG keeps me informed.  It is invaluable when
> used for operations.  The Melissa virus, the WTC disaster, business
> disasters.  I didn't mention the names of service providers, but I consider
> them to be vendors as well.

That's funny, I remember a *lot* of vendor names being mentioned during the WTC
attack and subsequent complications.  Maybe not equipment vendors, but certainly
service vendors.  There was a lot learned there, and if you ignore who was
helping who and which vendors did the learning, you're going to be behind the
next time it happens.

> A vendor that releases an official press release is more than welcome on
> NANOG.  Up to you gues.  I'm just aginst opinionated vendor information
> coming from NANOG.  Moderate yourself.

Press releases are marketing, and thus are not appropriate for NANOG.  If
someone wants to provide on-topic technical input, that should be welcome
regardless of which ISP/vendor/etc they work for, or whether they hide their
affiliation behind yahoo/hotmail.

> I can build a network with any vendor.  Just like you.  Vendor does NOT
> MATTER TO ME!  If you're posting on NANOG and you don't feel the same,
> perhaps you'd better unsubscribe.

Do you really think the equipment/circuit/service vendors you choose do not
affect how you design your network or the resultant services and availability
you deliver to your customers?

As Randy says, I encourage all of my competitors to think that way.

> The other option may be to study Ethernet.

Ethernet isn't a terribly difficult concept.

> Everything else you said (below) makes you look like the mad scientist from
> my POV.  My R&D budget is down 72% from last year.  RPR?  RUCRAZY?  Look
> forward to seeing you on the battlefield while I chop you to shreds.  Not a
> single ATM project on the map nowadays huh?  MPLS is out there, but who
> needs it 'cept for certain situations sometimes driven by business needs.
> When I need it, I hope it is over Ethernet.

Ethernet is just another tool in the box.  It has its uses, just like T1's,
SONET, ATM, and even Token Ring.  Perhaps it's more useful today than it was
yesterday; we'll undoubtedly have a shiny new toy tomorrow though.

> Wouldn't you rather have a reliable, redundant L2 core with isolated
> failures and fast recovery?

No.  You haven't dealt much with STP if you think this is a good idea.

> There are many examples of Ethernet in the internet core today.

You'd be hard pressed to find any ISP world-wide that didn't use Ethernet.

> I know of many service providers using another vendor besides the
> two favorites on NANOG.

Just about every ISP uses other vendors.  Your point?

> Ethernet: 2  (Mike Lieber did not challenge Ethernet as a technology)
> Valdis:   1

Ethernet is _a_ technology.  It's not the only one, and it never will be.

S




More information about the NANOG mailing list