x
Optical/IP

MPLS Bigwigs Get Edgy

PARIS -- Many of the big-name boffins of the IP world, along with about 600 delegates, have turned up here at the MPLS World Congress ’05 this week to review the status quo of developments underpinning carrier convergence hopes.

The overall message is that the focus of Multiprotocol Label Switching (MPLS) discussions has moved on to the services and applications it enables, rather than the technology itself.

“Customers are literally asking for MPLS now. It has a brand name,” said Kevin Macaluso, VP and general manager of Alcatel's (NYSE: ALA; Paris: CGEP:PA) IP division, in the conference’s introductory remarks.

Progress has also been made in addressing a couple of the issues holding back deployment of Layer 2 VPNs -- namely the scaleability and manageability of Virtual Private LAN Service, the multipoint Ethernet service that runs over MPLS infrastructure. This (and continuing vendor squabbling) is evident from the MPLS interoperability demo at the show (see Juniper: The VPLS Odd-Ball?).

The focus of attention is now shifting towards three particular topics:
  • Extending MPLS into access networks
  • Standardizing the way carriers can link together their VPN services to meet geographic coverage requirements
  • Finding more scaleable multicast technologies to support triple-play applications incorporating video, voice, and data


MPLS in Access Networks
“We’re starting to see MPLS getting pushed into access networks,” said Andrew Malis, chairman and president of the MPLS/Frame Relay Alliance and chief technologist of advanced data products at Tellabs Inc. (Nasdaq: TLAB; Frankfurt: BTLA), in a debate on the first day of the conference.

In terms of business services, a number of vendors are now offering equipment that extends MPLS pseudowires to customer premises equipment.

One such vendor, RAD Data Communications Ltd., participated in the show’s interoperability demonstration with a number of unannounced products. These include a VPLS network termination unit, called the ETX-510; a device for carrying legacy ATM connections over pseudowires, the ACE-3200; and a device for carrying TDM over pseudowires, the IPmux-14.

“We’re really showing some stuff out of the R&D lab,” said Yuri Gittik, RAD’s director of business development. RAD has discontinued development of the carrier edge device that was in its original plans in this area (see RAD Plans Ethernet Access Push).

Other vendors with MPLS CPE include World Wide Packets Inc. and Overture Networks Inc., according to Malis. Tellabs resells both vendors’ equipment.

In his introductory remarks, Alcatel’s Macaluso said the next “potential explosion” of MPLS use could occur in the residential service market, supporting bundling and triple-play services.

This implies MPLS being extended into homes as well as business premises in the future -- an idea that was greeted with skepticism by other speakers. “All you need out there is a multiplexing layer,” said Bruce Davie, a Cisco Systems Inc. (Nasdaq: CSCO) Fellow. “It seems a bit of a stretch to say there’s a lot of value in pushing MPLS out there.”

“You’ll see MPLS being used in the aggregation layer,” said Malis. In other words, it might extend as far as DSLAMs. At least one carrier, Telefónica SA, is already aggregating traffic from 200,000 residential users in this way, according to Gary Holland, director of marketing for EMEA at Riverstone Networks Inc. It’s linking IP DSLAMs into the infrastructure used to support VPLS services for business users.

“I’m not going to say MPLS should go this far and no further,” said Kireeti Kompella, a Juniper Networks Inc. (Nasdaq: JNPR) Fellow. “People don’t converge because MPLS is sexy. They do it because it makes business sense. MPLS comes with a cost."

Interconnect
The next big challenge for MPLS-based services is to standardize the way Layer 2 VPNs can be linked together to address geographic coverage requirements, according to Malis.

Although carriers can reach distant locations by leasing circuits from other carriers, it’s far more cost effective to interconnect with other carriers offering VPN services, according to Charlie Muirhead, CEO of Nexagent, who cited a study showing savings of between 30 percent and 70 percent. Nexagent provides a way for multiple carriers to jointly offer services (see Nexagent Promises VPN Nirvana and Carriers Join Nexagent Program).

Other carriers are working directly with counterparts in other countries. Cable & Wireless plc (NYSE: CWP) gave an example of this at the conference, in a session describing the engineering and commercial aspects of its MPLS interconnect agreement with Germany’s Arcor AG.

“QOS is the area where the biggest challenge is,” said Cisco’s Davie. “It’s possible to do inter-domain QOS now, but it’s taking way too much effort to do it.”

By standardizing interconnect for Layer 2 VPNs, carriers could avoid a lot of this effort, according to Malis. Peering points, equivalent to the ones that exist on the Internet, could be established.

This is relatively easy for Layer 3 VPNs, which are based on BGP (Border Gateway Protocol). The challenge is to “stitch together pseudowires for Layer 2 peering,” said Malis. The Internet Engineering Task Force (IETF) has drafted an RFC for this, and Malis expects to see results, and possibly vendor implementations, by the end of the year.

Multicast
Yakov Rekhter, a Juniper Fellow, gave a keynote address in which he tried to light a fire under the development of technology to support widespread use of multicast in MPLS networks -- possibly because Rahul Aggarwal, a Juniper Distinguished Engineer, is behind the latest effort to establish another IETF RFC in this field, called Point to Multipoint Traffic Engineering Label Switching Paths.

At the show, Juniper is demonstrating its use for a video distribution service developed by British Telecommunications plc (BT) (NYSE: BTY; London: BTA).

Rekhter said existing multicast technologies worked fine on today’s small-scale multicast deployments, but used his keynote to demonstrate that they wouldn’t scale to support conditions where a “provider edge” piece of equipment in an MPLS network had to support multicasts to 1,000 connected devices.

The need to support such a huge-scale multicast was questioned by Gurvinder (Bobby) Singh, IP multicast product manager for Cisco, who was in the audience. Singh asked whether Rekhter was using “scare tactics” by citing such a large number, noting that he was familiar with many of the world’s IP multicast projects and none of them came close to the size cited by Rekhter.

Rekhter responded by saying he’d been proved right in the past, when people thought he was crazy to suggest that MPLS equipment would be called on to support tens of thousands of IP VPNs.

Rekhter’s been sending the same message about multicast protocols every time he’s spoken recently, according to Malis. “He’s been a bit of a broken record,” he says.

— Peter Heywood, Founding Editor, Light Reading


For further education, visit the archives of related Light Reading Webinars:

gavlar 12/5/2012 | 3:26:16 AM
re: MPLS Bigwigs Get Edgy Seems they just had a big news release on this very subject.

Press Release Source: Broadwing Communications, LLC


Broadwing Communications to Launch Unique Converged Services Network
Monday February 14, 7:46 am ET
Opens the Door to a New Era in MPLS Virtual Private Networking


AUSTIN, Texas--(BUSINESS WIRE)--Feb. 14, 2005--Broadwing Communications, LLC, a consolidated subsidiary of Broadwing Corporation (NASDAQ:BWNG - News), today announced that it is completing the roll-out of a unique, Converged Services Network infrastructure based on Multiprotocol Label Switching (MPLS) that will enable both Layer 2 and Layer 3 Virtual Private Network (VPN) services for enterprises and carriers. On track for general availability in the second quarter of 2005, the Converged Services Network will dramatically exceed the capabilities of traditional MPLS networks.
Peter Heywood 12/5/2012 | 3:26:12 AM
re: MPLS Bigwigs Get Edgy Perhaps the bigger point to make here is that Layer 2 VPN deployments (VPLS etc) are now said to be growing faster than Layer 3 VPN deployments, and both are growing at a cracking pace, according to Infonetics statitics quoted by Malis.

On why I didn't mention the Broadwing deployment in the story...there's lots of deployments these days so why should I single out that particular one?

By the way, I did ask Malis why Tellabs didn't participate in the interoperability demo, and I didn't put his answer in the story either. It was that Tellabs was so busy doing customer trials that it didn't have any spare equipment or resources.

imlazar 12/5/2012 | 3:26:10 AM
re: MPLS Bigwigs Get Edgy Good, that is exactly what we've been focusing on at MPLScon for the last two years and what we'll be focusing on at this year's show in NYC this May.
gigeguy 12/5/2012 | 3:26:09 AM
re: MPLS Bigwigs Get Edgy Broadwing is just one example of what we'll be seeing more and more of in the future - MPLS-based convergence with pseudowires and VPLS replacing older separate L2 networks, and also adding new services. As you quoted Kiritti and others in the article, it makes economic sense for the carriers.
rriicc 12/5/2012 | 3:26:05 AM
re: MPLS Bigwigs Get Edgy Is there any knowledge of real numbers of how many paying users (companies) that are actually using e.g. MPLS/BGP-VPN etc. services? And the revenues behind it.. I'm not asking if a railroad company in Burma is offering the service, and I'm not asking about revenues of core routers sales that includes MPLS whether you want it or not.
I see a discussion here that looks the same as 5+ years ago. I see same conference clowns still talking about scalability issues, QoS, MPLS at the edge and blablablah.
Is it an economic reality behind this or is it just loads of people travelling to conferences on companies' expenses?
Just a simple question... ;-)
lilgatsby 12/5/2012 | 3:26:03 AM
re: MPLS Bigwigs Get Edgy The market is converging and it's not a PowerPoint dream like in 2000. There are plenty of examples to back this up. Is MPLS the standard today - no. Is it gaining credibility quickly through its ability to inexpensively carry guaranteed differentiated services - yes.

It's really pretty simple: enabling multiple tied services over a common infrastructure introduces new revenue streams. One could debate the QoS, resiliency, ROI, etc...but the bottom line is that this is a viable and comparitively inexpensive option for many cable companies and telcos to add new revenues to the bottom line.

Just my 2cents (no I'm not in marketing for one of these vendors).

lg

------------------------------------------------
Is there any knowledge of real numbers of how many paying users (companies) that are actually using e.g. MPLS/BGP-VPN etc. services? And the revenues behind it.. I'm not asking if a railroad company in Burma is offering the service, and I'm not asking about revenues of core routers sales that includes MPLS whether you want it or not.
I see a discussion here that looks the same as 5+ years ago. I see same conference clowns still talking about scalability issues, QoS, MPLS at the edge and blablablah.
Is it an economic reality behind this or is it just loads of people travelling to conferences on companies' expenses?
Just a simple question... ;-)
gigeguy 12/5/2012 | 3:25:45 AM
re: MPLS Bigwigs Get Edgy The service provider revenues for MPLS-based VPNs (according to analyst studies from yankee group, infonetics, etc.) was over US $3B in 2004 and growing at about 30%/year, so yeah, it's real. You would know this if you read network world instead of this rag. :-)
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE