The MPLS WG Archive

["Brijesh Kumar" <bkumar@ennovatenetworks.com>]

In his previous mail, Yanhguang writes,

> -----Original Message-----
>
> See my comments below:
> I guess my point is misunderstood. SONET protection switching
> uses dedicated
> overhead bytes for signaling and end-to-end coordination.
> This is why SONET
> protection is fast and deterministic. MPLS uses message-based
> signaling to
> coordinate protection switching. With only best effort IP,
> there is no way you
> can guarantee to deliver these signaling messages on time. So
> there has to be
> infrastructure to treat these signaling messages specially.
> This is why I
> mentioned QoS.

It appears that we are perfectly in sync. You raised a very valid
point regarding the need to accommodate QoS in MPLS protection
switching. The point I was trying make was that even without QoS, it
is very hard to obtain SONET like reliability in the system with the
mpls local protect mechanisms. There are many more issues such as
provisioning for fall back capacity, and ability to survive
multi-point failures.


Cheers,

--brijesh
Ennovate Networks Inc.