RE: Issue-5: RE: Dropping the Local Overbooking Multiplier (LOM) meth od from D S-TE specs?
"Francois Le Faucheur (flefauch)" <[email protected]>
| Newsgroups | gmane.ietf.tewg |
|---|---|
| Message-ID | <[email protected]> |
Sanjaya, Dimitry, I could be wrong here (apologies if that is the case), but I have the impression the point of disconnect between you two is that: Dimitry's pretty much arguing that: - if you need per-CT-per-Link ratios and can tolerate some inaccuracy in how a CT affects the global constraints, then you do NOT need LOM. While Sanjaya is pretty much arguing that: - if you need per-CT-per-Link ratios and want accuracy in how a CT affects the global constraints, then you do need LOM. Both are true. The trick is in accuracy of how a CT impacts other CTs who share a global constraint. Hope that helps Francois >> -----Original Message----- >> From: Choudhury, Sanjaya [mailto:[email protected]] >> Sent: 05 June 2003 15:06 >> To: [email protected] >> Subject: RE: Issue-5: RE: Dropping the Local Overbooking >> Multiplier (LOM) meth od from D S-TE specs? >> >> >> >> Hi Dimitry! >> > -----Original Message----- >> > From: Dimitry Haskin [mailto:[email protected]] >> > Sent: Wednesday, June 04, 2003 11:13 AM >> > To: 'Choudhury, Sanjaya'; [email protected] >> > Subject: RE: Issue-5: RE: Dropping the Local Overbooking Multiplier >> > (LOM) meth od from D S-TE specs? >> > >> <snip..> >> > > >> > > 2. For a moment let is assume that preemption *is* >> > > deployed in the domain. Can you explain me how I >> > > will achieve the following [please be specific in >> > > what the user will need to do and the BC model in >> > > use]: >> > > >> > > case-1: want to overbook data-traffic without >> > > overbooking voice traffic. >> > > >> > > case-2: Want to overbook the voice traffic by >> > > 10%, but want to overbook data traffic by 40% >> > > >> > > case-3: want to overbook gold-traffic by 10%, >> > > overbook silver-traffic by 15%, overbook >> > > normal-traffic by 20% >> > >> > What is the catch? Why would not you do just that? I.e., >> > overbook each CT >> > independently to their respective designated level. >> >> I am still trying to understand _how_ (you propose to >> overbook each CT independently)? DSTE-PROTO proposes >> 3 approaches: >> (i) LSP Size overbooking i.e user specifies a >> higher/lower bw requirement for the LSP during >> creation) >> (ii) Link Size overbooking (user modifies the >> max_reservable_bw) >> (iii) LOM >> Which of these above three approaches, should the >> administrator use (obviously you are not indicating >> (iii) ? >> >> Or are you proposing to manipulate the BCs (for >> individual CTs)??? >> >> > Or, I supposed, you have missed the aggregate constraint from >> > you examples >> > which would make them slightly more illustrative. Not that I >> > want to help >> > you to make your case ;) >> >> Your help will be appreciated. >> >> Thanks, >> sanjay >> >> > >> <snip..> >> > >> >>