The MPLS WG Archive

[Rohit Dube <rohit@xebeo.com>]

On Fri, 22 Nov 2002 10:02:28 -0500 David Charlap writes:
[snip]
=>Any one of these would be fine for me.  My main concern here is that 
=>MPLS vendors will start implementing those features, and some may ignore 
=>the GMPLS approach, thinking that those drafts are not applicable to 
=>non-generalized MPLS.  This may lead to interop issues.
=>
=>Of course, vendors often ignore drafts anyway, but at least this way 
=>they'll know that there is something out there that they can choose to 
=>implement or ignore.  And those that need to pull pieces of GMPLS into 
=>their MPLS implementations will know how much must be pulled in to 
=>properly support the features under development.

Excellent points David. And in fact this is what happened with the 
"bidirectional lsps for classical mpls" draft that I presented at the
IETF. 

A quick solution is needed for this class of problems!

Regards,
--rohit.