The MPLS WG Archive

[David Charlap <David.Charlap@marconi.com>]

"Paola Iovanna (ERI)" wrote:
> 
> I'd like to know more information about how RSVP-TE  performs
> bandwith decrease modify in a MPLS scenario.  It is clear the
> procedure followed in order to perform bandwidth increase while it
> is not clear how it is released the bandwidth in case of bandwidth
> decrease.

Same way as an increase.  One of two mechanisms:

1: Have the ingress change the TSPEC it sends.  This should propagate
through the network and result in an updated FLOWSPEC propagating back
in Resv messages.

When the bandwidth is decreasing, you'll probably see a quicker
response, since RSVP implementations usually clip the FLOWSPEC to the
level of the TSPEC.  It should work fine, assuming the router allows LSP
modification in the first place.

2: Make-before-break.  Signal a new LSP with a different LSP ID.  Use SE
style to share resources with the old one.  Once you get the Resv back
for the new LSP, cut your traffic over to the new one and tear the old
one.

-- David