RE: MPLS VPNs or not?

Mailman List Mirror (Read Only)
1

From: Vadim Antonov [mailto:avg@exigengroup.com]
Sent: Wednesday, August 08, 2001 3:45 AM
To: Christian Kuhtz
Cc: nanog@merit.edu
Subject: Re: MPLS VPNs or not?

<snip>

I'm known for making strong statements, you are right.
However, so far
I've heard no convincing argument against these positions;
and I tried to
solicit opinions by offering the opinions in question for the public
review. Frankly, i was disappointed by the "let's wait and
see which one
works" attitude. As Randy was kind to point out, that
attitude already
gave us a slew of DOA networking technologies, to the tune of billions
wasted.

I don't think UUNET considered it a waste. UUNET could not have grown
as quickly as it did during the mid to late 90s without L2 (Frame and
ATM) technologies. Fortunately for them, they did not have any pure IP
only zealots that prevented the pragmatic use of other technologies
in their networks. Otherwise they probably would not have been able
to outrun the other ISPs.

UUNET received two benefits from it:

  1. Speed, since at the time L2 switches were faster
      than routers, and
      2. Traffic engineering, which saved them money in transport
      costs.

Point 1 is no longer valid. Point 2 is still valid.

UUNET built bigger and better networks at the time because of this.
The market decided that UUNET was right. UUNET's shareholders were
well rewarded because of what you called this "waste".

I guess the real question should be how much market cap did other
companies lose because of certain people's zealotry? Any answers
Vadim?

Prabhu

2

I don't think you can make a strong argument on TE being that big of a
factor. It's more likely that the provider market's winner-take-all nature
combined with acquisitions and attrition contributed much more to their
growth then and benefits derived from traffic engineering.

Furthermore, it may be possible that complex traffic engineering is
becoming less of a factor as backbone networks become increasingly based on
physical transports that are less heterogenous (in terms of
log(backbone_speed/client access speed)).

3

I don't think UUNET considered it a waste. UUNET could not have grown
as quickly as it did during the mid to late 90s without L2 (Frame and
ATM) technologies. Fortunately for them, they did not have any pure IP
only zealots that prevented the pragmatic use of other technologies
in their networks. Otherwise they probably would not have been able
to outrun the other ISPs.

UUNET received two benefits from it:

        1. Speed, since at the time L2 switches were faster
            than routers, and
        2. Traffic engineering, which saved them money in transport
            costs.

Point 1 is no longer valid. Point 2 is still valid.

UUNET built bigger and better networks at the time because of this.
The market decided that UUNET was right. UUNET's shareholders were
well rewarded because of what you called this "waste".

I guess the real question should be how much market cap did other
companies lose because of certain people's zealotry? Any answers
Vadim?

Diddn't PSInet deploy L2 switching massively throughout their network?
What did the market decide about that? Could it be that UUNet's success
was due to other factors? BTW, I'm not sure that #1 above was ever true
in a large scale network.

KL

4

NETCOM did it too.

  I think he's got it all wrong.

  Selecting the wrong technology /can/ kill you, but it's unlikely
to make you a success.

  --msa

5

Did they every finally set the last switch they decom'ed on fire? Or was it
just the sledgehammer drill? I had to miss the party...

6

I don't think UUNET considered it a waste. UUNET could not have grown
as quickly as it did during the mid to late 90s without L2 (Frame and
ATM) technologies. Fortunately for them, they did not have any pure IP
only zealots that prevented the pragmatic use of other technologies
in their networks.

Did I ever argue against L2 switching? It is a fine way to do traffic
aggregation/deaggregation. Just don't do _routing_ with that.

Otherwise they probably would not have been able
to outrun the other ISPs.

UUNET received two benefits from it:

  1. Speed, since at the time L2 switches were faster
      than routers, and

Ghm. FR boxen were cheaper, not faster. ATM at some point was faster, but
was (still is?) quite flaky. Networks i engineered had plenty of L2
switches in them - for clustering in POPs.

      2. Traffic engineering, which saved them money in transport
      costs.

TE at the cost of 20% of available bandwidth wasted to cell tax? You must
be kidding. Do not forget that TE could be done at SONET level, too.

I guess the real question should be how much market cap did other
companies lose because of certain people's zealotry? Any answers
Vadim?

Absolutely. Stupidity of your regular analysts, who were rewarding
companies investing into the latest overhyped crap. I hope these times
are over by now, and companies will actually start looking at the bottom
line.

Now, if you find an example of a network going down _because_ of pure-IP
design (not because of stupid business tactics such as overexpansion,
idiotic acquisitions, or simple mismanagement), i'll agree with you. So
far, most of networks which went down (or were acquired at bargain
prices) were hybrid designs. Also note that UUNET _did not_ survive as an
independent ISP. I know few ex-UUNET folks who weren't very thrilled
because of that.

--vadim