[email protected]
[Top] [All Lists]

Re: composite link - candidate for respin, maybe

Subject: Re: composite link - candidate for respin, maybe
From: Curtis Villamizar
Date: Thu, 01 Apr 2010 01:26:49 -0400
In message <C7D6F5FF.924C%[email protected]>
Tony Li writes:
>  
> >>>   5.  LSP signaling is needed to indicate a preference for placement
> >>>       on a single component link and to specifically forbid spreading
> >>>       that LSP over multiple component links based on flow
> >>>       identification beyond the outermost label entry.
> >>  
> >> Do you meant to imply that there would also be a way to indicate a
> >> preference for diverse placement of a hierarchical LSP?
> > 
> > 
> > That is definitely one of the choices.  Just as is done today.
>  
>  
> Could we please explicitly call out all of the alternatives here, please?
> Just to be crystal clear...
>  
> Thanks,
> Tony


OK.

   5.  LSP signaling is needed to indicate a preference among the
       following two behaviours.

         a.  Placement of the LSP on a single component link.  This
             option specifically forbids spreading that LSP over
             multiple component links based on flow identification
             beyond the outermost label entry.

         b.  Load balancing using flow identification information
             beyond the outermost label entry.  The preferred flow
             identification method and flow distribution method must
             be a method whose support is indicated in the link
             advertisement.

I was under the impression that the exact set of mechanisms is out of
scope for the requirements document.  If we were to put in any load
balancing mechanism at this time we would document simple hash based
load balancing by citing RFC2991 and RFC2992.  We would also leave
codepoints to allow other methods to be specified.  We might even set
aside a set of experimental codepoint available for "unspecified
vendor magic" to be assigned real code points later.

For the requirements document the only requirement is "extensible".

This is supposed to fit in 5 pages after all.  The point was *not* to
constrain the requirments by specifying a particular implementation.

Curtis
_______________________________________________
rtgwg mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/rtgwg

<Prev in Thread] Current Thread [Next in Thread>