[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Proposal for Simplification of Overbooking



Wai Sum, Jerry -

I'm a little unclear as to your intention in the "integrated approach."  Is it
that the Local Overbooking Multiplier, instead of an independent representation
of a CT's overbooking policy (which is what the new term "OB(CTc)" would
signify), will now be rather a kind of correction factor, configured as needed
link by link, to other factors which are configured globally?

I think in general this wouldn't be a bad way to get compatibility with older TE
procedures. However, I'm still uncomfortable with how even network-wide per-CT
policies are supported using Link Size Overbooking.  And the integrated approach
appears to build on an understanding of how this is done: "For either per-domain
per-CT overbooking or per-link per-CT overbooking, we have Reserved(CTc) =
Tspec(LSPc) / OB(CTc)" and "OB(CTc) = LSOM * LSPOM(CTc) * LOM(CTc)."  

Specifically, how do you calculate Unreserved (CTc) using "LSPOM" when there are
per-CT policies and an aggregate constraint?  I don't see that BCc - Reserved
(CTc) / (LSOM * LSPOM) yields an effective value to advertise. After all, if CT7
has a policy of no overbooking, even though the link generally is allowed to be
overbooked, if BC7 = 50 and the CT's  reservation total is 50, then Unreserved
(CT7) should be 0.      

Thanks for clarifying, 
Sandy

Sanford Goldfless
192 Fuller St
Brookline MA 02446
617-738-1754
sandy9@rcn.com


-----Original Message-----
From: Lai, Wai S (Waisum), ALABS [mailto:wlai@att.com] 
Sent: Tuesday, June 10, 2003 10:53 AM
To: Francois Le Faucheur (flefauch); Choudhury, Sanjaya; Dimitry Haskin; sanford
goldfless
Cc: Ash, Gerald R (Jerry), ALABS; Jim Boyle; Ed Kern
Subject: Proposal for Simplification of Overbooking

Attached is an I-D that provides more details on our proposal to simplify
overbooking from an operational perspective.  Before posting it, we would like
to hear your comments.
Thanks, Wai Sum and Jerry

<<attachment: winmail.dat>>