If there was one sentiment that summed up the seventh annual ATM Migration Challenge conference in Washington, D.C. last month it was this: future public networks will have MPLS at the core, not ATM, and it is just a question of when and how each individual carrier will get there. For some, it will be quite some time.
As the chairman of the event, which is hosted by show organizer Marcus Evans, I went in more bullish on MPLS than ATM, a reverse of my attitude when I chaired the conference last time.
What changed? For one, MPLS has gained stability and wider support. In May, 21 vendors demonstrated MPLS interoperability at SuperComm.
Two, more companies are publicly talking about MPLS futures. AT&T Corp. CTO Hossein Eslambolchi recently told me he envisions AT&T having an MPLS core with an IP control plane. And three, as it becomes more evident that future applications will be based on IP, ATM looks increasingly long in the tooth.
But people at last week’s event convinced me that ATM might be longer lived than I presumed. As one speaker from a carrier said, “ATM is the cow that is generating revenue: We need to keep milking it.”
With carrier capital equipment expenditures down 40 per cent compared with two years ago (by some calculations), it may be a while before service providers are in a position to build out MPLS cores.
Some carriers, however, are using MPLS with success. Ron Bonica, senior manager of network design with WorldCom Inc.’s Very high performance Backbone Network Service (VBNS+), reported to the group that contrary to some criticisms you hear about the standard, MPLS is stable and working fine.
Where work is needed is on the ATM/MPLS interworking so carriers can deploy MPLS cores and use ATM at the edge. The ATM Forum and the IETF have created solutions for carrying ATM over MPLS, and they have taken their respective positions (read, fight) to the International Telecommunication Union.
Thankfully, a compromise was worked out at a recent ITU meeting in Japan under which four ATM-MPLS-ATM encapsulation modes will be recognized, and the process can start moving forward.
The upshot: better technology or not, MPLS won’t supplant ATM overnight because there is too much sunken investment. It will take some time for MPLS to make inroads, especially in the current economic climate.
Dix is editor in chief of Network World (U.S.). He can be reached at email@example.com.