Just my 1 cent :-)
PBT alone doesnt have much steam to replace MPLS in the core and the edges as it can only deliver E-Line service, but PBT with PBB has all it takes to do so.
MPLS is currently on the 2.5 layer on the OSI.
Hence it is dependent on Layer 2 and Layer 3 to be implemented (exclude GMPLS).
This had been a boon to provide such a wide array of services on the prevalent last mile and core infrastrcuture (TDM/SONET/SDH/ATM/FR)
till now and till distant future.
Assume, some years down the line carrier ethernet becomes a reality and you have all the fiber deployed in the core and well populated edges and for the far edges you are using mid-band ethernet. Then this PBB with PBT is a definately a replacement for MPLS based services like L2 and L3VPN's.
The only reason being it can provide all the MPLS based services can provide, with lesser cost, as there is no overhead of purchasing and deploying L3/MPLS devices in the SP cloud hence reducing the cost for the SP. As it run on Layer2.
As for PBB/PBT it can deploy E-line and E-lan service hence providing all that could be provided by VPLS/EoMPLS/L3VPN and that too without the issues of resiliency, mac learning/flooding, with deterministic QOS.
Realistically now in current time we can think of using PBT in the core only (cross connecting your core locations like you do in you SDH/SONET environment) and connect your Edges running MPLS clouds.
Now when the complete transition/replacement of MPLS with PBB/PBT or even with something else will happen (there are other contenders as well to MPLS apart from PBB/PBT but still in evolutionary stages) , only time can tell.
But whenver this happens, we really dont have to worry about having a forum like this for PBB/PBT as there is no complex control plane to understand and deal with as it is with MPLS. Which also means we wont have much work to do :-).
HTH-Cheers,
Swaroop
