The MPLS WG Archive

[Dimitrios Pendarakis <DPendarakis@tellium.com>]

One benefit of the Bundle message, which might be more
relevant in the optical domain, is that it potentially 
could be used to signal set-up of multiple LSPs (lightpaths) 
at the same time.

There was some discussion in the context of the OIF UNI
about signaling multiple lightpaths simultaneously; but
it was not included in the final UNI 1.0 spec.
The idea was that it would be beneficial to path computation, 
for example.

Dimitris


> -----Original Message-----
> From: Yousuf Hasan [mailto:yhasan@cisco.com]
> Sent: Thursday, October 25, 2001 3:27 PM
> To: Feng, Mark
> Cc: 'Arvind, K'; 'mpls@uu.net'
> Subject: RE: RFC2961 BUNDLE message
> 
> 
> At 12:07 PM 10/25/2001 -0700, Feng, Mark wrote:
> >One thing it saves is the IP header, since multiple RSVP 
> message will be
> >packed into one.
> 
> The Summary Refresh helps solve the scaling issue a little 
> more, but it can only be used for messages in steady state.  
> Thanks
> 
> -Yousuf 
> 
> 
> 
> >If the question is, by the Bundle message itself, helps the 
> scaling problem.
> >IMO, not a whole lot.
> >
> >- Mark
> >
> >> -----Original Message-----
> >> From: Arvind, K [mailto:arvind@tenornetworks.com]
> >> Sent: Thursday, October 25, 2001 11:47 AM
> >> To: 'mpls@uu.net'
> >> Cc: Arvind, K
> >> Subject: RFC2961 BUNDLE message
> >> 
> >> 
> >> Does the BUNDLE message offer any real benefit in terms of 
> >> reducing control
> >> overhead? There seems to be no significant change in the 
> >> overall amount of
> >> control traffic that is generated or processed as a result of 
> >> bundling (even
> >> though the traffic is bunched together resulting in fewer 
> (but larger)
> >> messages). Does this mechanism really help in scaling RSVP?  
> >> Could someone
> >> throw light on this? Thanks.
> >> 
> >> Regards,
> >> Arvind.
> >> 
> 
>