The MPLS WG Archive

[Curtis Villamizar <curtis@fictitious.org>]

In message <05707214338CD5119BFF0040A5B170D3028897C7@mail3.tellium.com>, Bala R
ajagopalan writes:
> 
> Curtis:
> 
> First, thanks for your patient response.
> Needless to say, a bit of exaggeration is
> necessary for humor (I hope there was some
> humor in this).

Yes.  I can parse humor.  [seems like an oxymoron as stated. :) ]

> I wasn't really following the thread on G.709,
> so this was not the issue. There are a number
> of G.xxxx documents related to ASON that use
> IETF protocols and this is what I was referring
> to. 
> 
> Now, I don't think that all solutions developed
> outside of IETF (using IETF protocols) are clean.
> However, I do think that so far, there has not
> been an adequate effort in IETF to understand
> the problem formulations coming from outside.
> This is the crux of my message.

There are also individuals, some of whom played a key role in ASON,
who could not make a convincing argument at times in the IETF, and
have gone outside the IETF where questionable requirements will
receive less carefull scrutiny, then come back to the IETF and expect
a rubber stamp.  Some of the same people that dominated IPLDN, IPATM,
ROLC and ION within the IETF went outside IETF and gave us ASON.

To be perfectly honest with you, I personally put no more credibility
in ASON that I did in the ITU IP-QoS documents of the mid-90s.  The
ITU has stepped out of bounds (again).  They deserve to be ignored.

> In this regard, yes, I do believe that you have
> been prominent in rejecting the entire notion
> of the UNI. I personally don't have an addiction
> to UNI (or any of the G.xxxx stuff), except that 
> some number of service providers
> (the potential customers of the technology) wanted
> this and an effort was made to deliver the features
> they wanted. Is there a better way of providing
> the same UNI functionality using GMPLS protocols? This
> is what one would like to hear from IETF, rather than
> a dismissal of the whole concept. Correct me if
> I'm wrong, but the present Andersson draft, IMO, 
> doesn't address this problem. 

The ISPs certainly did not drive this.  The SPs, wannabe ISPs, maybe.
I do think the vendors pushed this and a few people at SPs, mostly in
legacy transport, were willing to stand behind it.

> Clearly, I won't place any bets on
> the success of UNI in the market place (what
> market?). Neither will I place any bets on GMPLS 
> at this point in time.

OK.  I tend to agree, however GMPLS may become successful.  The future
of GMPLS is not certain, however if GMPLS dies it will be very likely
to due to core optical switching turning out to be even less useful
than currently anticipated.

> Thank you.
> 
> Bala

Curtis

ps - for a related dose of humor see
<http://www.avici.com/technology/whitepapers/bigbangtheory.pdf>
keeping in mind that at the time of publication (April 1, 2000) it was
herassy to suggest that optical switching was going to be anything
less than enormously successful or suggest that all optical switching
would not be the dominant technology of IP backbones or suggest that
exponential growth of the Internet would not be sustained
indefinitely.