RE: reminder: things todo in august and shortly thereafter

["Ash, Gerald R (Jerry), ALCTA" <gash@att.com>]

Hi Tricci,

Thanks for your comments.  Please see some additional comments below.

Regards,
Jerry

> -----Original Message-----
> From: Tricci So [mailto:tso@caspiannetworks.com]
> Sent: Monday, September 03, 2001 1:26 AM
> Cc: 'te-wg@ops.ietf.org'
> Subject: Re: reminder: things todo in august and shortly thereafter

> Many thanks for your info regarding Class-Type priority definitions and
LSP
> bandwidth modification selection example.

> I can see that the amount of path computation would be very extensive with

> the solution that you proposed.  May I ask, have you done any performance 
> analysis to determine if such solution is scalable for a large core
network, 
> let's say 10's thousands of LSPs?

Yes, a comparative scalability analysis was done, and is presented in
Section 3.3 of
http://search.ietf.org/internet-drafts/draft-ash-mpls-diffserv-te-alternativ
e-02.txt.  Actually, the main point of our I-D was to point out ways to
increase the scalability of DiffServ TE.  

On a more general note, one can question scalability even without
introducing multiple class-types with DiffServ TE.  For example, see our
discussion of IGP scalability in
http://search.ietf.org/internet-drafts/draft-ash-ospf-isis-congestion-contro
l-00.txt.  In that draft we propose mechanisms to improve current IGP
scalability.