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

RE: CSPF related....



Title: Message
IMHO, simple constraints like availability (Simple boolean constraints) are easy to manage across a network in a link state routing protocol. More complex accumulative constraints like (Max allowable delay) are harder to manage across a network for use in a link state routing protocol. Complex accumulative type of constraints seem to lend themselves better to explicit source routing scenarios. Explicit source routing (as mentioned in many drafts) does not need to be standardised.
 
Curious to hear other opinions.
 
Vik
-----Original Message-----
From: Vishal Sharma [mailto:v.sharma@ieee.org]
Sent: Monday, October 14, 2002 9:32 AM
To: Satish Jamadagni; ccamp@ops.ietf.org
Subject: RE: CSPF related....

Satish,
 
The CSPF is one of the few things that distinguishes different system vendor or network planning tool
vendor offerings from one another. As such, it is unlikely that there would be a move to standardize it.
The CSPF (especially an off-line version) could, for example, use information beyond what is provided
by either RSVP-TE or OSPF-TE.
 
The basic idea of what a CSPF module/process should expect from OSPF-TE and RSVP-TE is
clear (the bandwidth parameters and, of course, topology, from OSPF-TE, and from RSVP-TE those
reserved resources that alter the bandwidth parameters of links), but I am not aware of any _draft or standards work_
that delves into details of these.
 
If there is some such work, I'd be interested in knowing about it too.
 
-Vishal
-----Original Message-----
From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On Behalf Of Satish Jamadagni
Sent: Saturday, October 12, 2002 6:11 AM
To: ccamp@ops.ietf.org
Subject: CSPF related....

Hi All,
I was curious to find out if any work has been done towards standardizing CSPF (Constrained Shortest Path First) related issues. At least a framework on what a typical CSPF module should expect from say OSPF-TE, REVP-TE or may be even RSVP-TE or OSPF modules taking CSPF algorithms or algorithm classes (in the literature) into account. What will CSPF do (or may not do) to LMP and vice versa etc??
 
I am aware of a draft on the nature of constraints by Kompella et al, Any thing else apart from that???
 
Thanks a lot
 
Regards,
Satish Jamadagni.