RE: Progressing BC Models was RE: Progressing MAR

[Jean Philippe Vasseur <jvasseur@cisco.com>]

Hi Jerry,

At 17:49 14/03/2003 -0600, Ash, Gerald R (Jerry), ALABS wrote:
> >> Also, if one does too much soft preemption, and forces a lot
> >> of flows off of short routes onto much longer, inefficient
> >> routes, this would (perhaps severely) degrade network
> >> performance.  There are studies which show that effect.
> >> One of the applications of bandwidth reservation mechanisms
> >> (as in MAR), used in large-scale networks today, is to
> >> protect against such an effect.
>
> > I am not sure how soft preemption, referring to the recent draft, forces
> > flows off short routes onto longer ones.  Are you referring to preemption
> > itself, regardless of whether its hard or soft?
>
> The draft says:
> "a new preemption pending flag helps more gracefully mitigate the re-route 
> process of displaced LSPs."
>
> It's the re-routing of preempted LSPs (onto longer paths) that I'm 
> referring to.
So you're referring to preemption regardless of whether it is a hard or 
soft preemption. Indeed preempted TE LSPs may be rerouted along longer path 
but how could it be different ? If a link cannot handle the total amount of 
requested bandwidth, some TE LSP (with higher priority) can get the 
bandwidth, other TE LSPs (having a lower priority) will follow a longer 
path. Soft preemption just ensures that soft preempted TE LSPs get a chance 
to be gracefully rerouted (make before break) without traffic disruption.

JP.

> Jerry