The MPLS-OPS Archive

[Robert Raszuk <raszuk@cisco.com>]

<you forgot to cc the group :)>

Hi Rick,

> 1) Is RSVP-TE fast reroute the most widely used and recommended failure
> recovery process?   What others are used out there and to they interoperate
> between vendors?

I would say No ! Not at all. It is just one of many ways for fast
recovery after the link/node failure. What are the two most used others:

SONET/SDH protection (50 ms range - link protection)
IGP convergence (could less then 1 s with optimised timers in ISIS)

It is worth to note as well that the two above are universal in a sense
that they woork for plain IP packets as well as LDP flows while FRR for
not only addresses the protection of MPLS TE LSPs. And yes we are
investigating the protection scope of FRR as well :).
 
> 2) Can I assign a 1 to N back-up tunnel for FRR?

Sure - at least in cisco :)
 
> 3) What is the average recovery time from the detection of the failure to the
> tunnel switch over?

After detection the switch itself is very fast. Less then 1 ms ususally.
The detection of link failure can take anywhere from few to tens of ms
and detection of node failure from tens to hundreds ms.

Rgs,
R.

> RickGall1@aol.com wrote:
> 
> Thank you Robert for your reply,
> 
> This is an area of concern to me also.  Here are my questions to the group
> and Robert,
> 
> 1) Is RSVP-TE fast reroute the most widely used and recommended failure
> recovery process?   What others are used out there and to they interoperate
> between vendors?
> 
> 2) Can I assign a 1 to N back-up tunnel for FRR?
> 
> 3) What is the average recovery time from the detection of the failure to the
> tunnel switch over?
> 
> Thank you all
> 
> Rick Gallaher

-------
The MPLS-OPS Mailing List
Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
Archive: http://www.mplsrc.com/mpls-ops_archive.shtml