Seven vendors are touting the world's first GMPLS multivendor interoperability demo. Should you care?

October 14, 2002

4 Min Read
GMPLS Demo: Missing the Point?

The MPLS Forum is today staging the world’s first multivendor interoperability demonstration of GMPLS (generalized multiprotocol label switching) at the Next Generation Networks conference in Boston (see MPLS Forum Demos GMPLS).

The demo targets technology that might one day help carriers automate control of their networks -- affording them massive savings in opex even while they respond much more rapidly to customer demands.

All the same, plenty of people think the demo’s focus on interoperability is misplaced. Right now, carriers have much more basic questions to ask about GMPLS, according to Tom Nolle, president of CIMI Corp. consultancy. "Carriers need to know, Why the hell do I need this? How can I make money on it?" Multivendor interoperability won't even be an issue initially, he says.

Nonetheless, participants say the demo, titled "GMPLS: Proof Of Concept," is a landmark effort to prove the viability of GMPLS, which extends the idea of Multiprotocol Label Switching (MPLS) -- automatically setting up and tearing down virtual connections over packet-based networks -- to embrace other layers, including the optical layer, in carrier backbones (see Optical Signaling Systems).

Agilent Technologies Inc. (NYSE: A), Cisco Systems Inc. (Nasdaq: CSCO), Data Connection Ltd. (DCL), Juniper Networks Inc. (Nasdaq: JNPR), NetPlane Systems Inc., NetTest, and Sycamore Networks Inc. (Nasdaq: SCMR) are demonstrating a forum-sponsored test initially conducted at the University of New Hampshire InterOperability Laboratory (IOL) from September 30 to October 4, 2002. The demo is set to start this afternoon at the Next Generation Networks conference at the Boston Marriott Copley Place Hotel.

The test shows the transmission of control messages over an OC48 (2.5 Gbit/s) link using a core of SN 16000 switches from Sycamore, with edge routers from Cisco and Juniper, aided by software from DCL and Netplane with emulated devices from Agilent and NetTest. (For a whitepaper and diagram, see www.mplsforum.org/NGNevent.html.)

Proponents are careful to stipulate the limits of the demo. "This demo points out that GMPLS is now in a position to work nicely," says Ananda Sen Gupta, VP of marketing at the MPLS Forum and product manager at Agilent. He's careful to say this does not signal that the participants think GMPLS is ready for deployment in carrier networks today. "The point is to show something that really works. We're not aiming to show something that's perfect," he says. Indeed, the test and demo also should help ongoing standardization efforts at the Internet Engineering Task Force (IETF), he says, by providing the standards-setters with consensus and proof-of-concept.

But the demo doesn't seem to address some key issues about GMPLS. One source, who asked not to be named, says a key area of interest among vendors and carriers alike is GMPLS's ability to support networks that combine the bandwidth efficiency of mesh networks with the topological efficiency of rings. This demo isn't concerned with architectural issues like this.

On the plus side, the demo proves that peer model GMPLS works. This gives GMPLS backbones some of the same characteristics as the Internet, enabling operators to jointly offer services over their combined backbones. But it remains to be seen how this method is implemented in real life. For instance: Will carriers be willing to show competing operators what's going on inside their backbones? Will device vendors be willing to have their gear overseen by an integrated management system they don't provide? (For more on these issues, see Light Reading's archived Webinar: Automating the Optical Network, hosted by Scott Clavenna, president of PointEast Research LLC and director of research at Light Reading.)

"In Asia, there are a number of carriers looking hard at GMPLS because they have the benefit of starting from scratch," writes Clavenna in an email. He says GMPLS is likely to ramp up slowly, with no significant deployments for at least three years among the big U.S. carriers and European PTTs. But in Asia, CLECs and smaller incumbents already are looking at peer-model GMPLS, he says, in order to tie together IP and transport networks that don't have strict organizational boundaries between them.

— Mary Jander, Senior Editor, Light Reading
www.lightreading.comMovers and shakers from more than 100 companies – including Agilent Technologies and Cisco Systems– will be speaking at Lightspeed Europe. Check it outat Lightspeed Europe 02.

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like