x
Optical/IP

Cisco Tracks PBT Standards Process

PBT (Provider Backbone Transport), the new Ethernet flavor championed by Nortel Networks Ltd. and BT Group plc (NYSE: BT; London: BTA), is on the first rung of the standards ladder at the Institute of Electrical and Electronics Engineers Inc. (IEEE) and is under examination by a number of other industry bodies.

The news is important for PBT's supporters, as becoming a standard will help legitimize the technology in the eyes of carriers and vendors alike. It will also rile some companies, particularly Alcatel-Lucent (NYSE: ALU), that don't want PBT to gain any market credibility, as they have rival technologies to sell. The main direct rival to PBT is Transport MPLS, or T-MPLS.

So what's all the fuss about? PBT is a new brand of Ethernet that strips out some of the technology's inherent complexity and enables service providers to create a managed, traceable, point-to-point Ethernet tunnel across a network that offers Sonet/SDH-like performance. As a result, the OA&M (operations, administration, and maintenance) attributes are very familiar to carrier operations staff, so it's a technology they should feel comfortable with and confident about deploying and managing. (See BT Likes Nortel's New Ethernet Flavor and PBT Means What?)

Those attributes have attracted the attention of a number of carriers. BT has been the most vocal in its support, working closely with Nortel, but BCE Inc. (Bell Canada) (NYSE/Toronto: BCE) and Verizon Communications Inc. (NYSE: VZ) have also expressed an interest in PBT. (See BT Rethinks 21CN Core Strategy.)

And that recognition has led more vendors, including Siemens Communications Group to promote the technology. And now Cisco Systems Inc. (Nasdaq: CSCO) is showing interest and believes the technology has some potential. A spokeswoman says Cisco believes "there are plenty of options to do what PBT does," but Cisco will monitor the Ethernet technology's progress through the standards bodies, "and we'll be adhering to the resulting standards. We're tracking it."

Like Cisco, the carriers want a technology that's standards-based. So with BT so keen on Nortel's story, the operator has been helping the vendor get PBT recognized by the IEEE.

Nortel's director of carrier Ethernet in Europe, Mervyn Kelly, says the vendor's strategy has always been to get PBT standardized as quickly as possible.

"We want it to be widely adopted, so we are pushing it through the IEEE as fast as we can," says Kelly. "Sure, it'll allow other vendors to catch up with us, but if the technology isn't adopted by other companies then it won't get taken up [by carriers]. It's not in our interest to be the only guys out there deploying it."

So what has been the process at the IEEE? Kelly says Nortel and BT "socialized PBT" at the IEEE about 12 months ago, again in July, and then in November submitted it for a "formal agreement to proceed" vote, which was passed with a high majority, with no votes against and just a few abstentions.

Now PBT is in the PAR (project approval request) stage, and is referred to in IEEE documents as PBB-TE (Provider Backbone Bridge Traffic Engineering). There will be a further submission in February 2007 for the "formal agreement to proceed" part of the process.

Once that agreement is in place, "it's hard to tell how long before the technology will be standardized," but, Kelly says "the rule of thumb is about a year" at the IEEE.

The Nortel man says PBT is not rocket science. "There's no actual new technology, which makes it safe and risk free for carriers," he says. "Essentially, two things are being examined in the IEEE standards process. The first is a process -- how to use existing Ethernet technology, including VLAN identifier plus destination address" to create a deterministic path for the Ethernet tunnel.

Second, says Kelly, "it's standardizing a method of switching off the 'broadcast unknown' feature in Ethernet. There's currently no standard way of doing that."

But the IEEE is not the only standards body looking at PBT. According to Kelly: The TM Forum has a work item looking at the OSS requirements for PBT; the Broadband Forum has been presented with information for consideration by vendors that might want to incorporate PBT into their DSLAMs; there's an Internet draft paper with the Internet Engineering Task Force (IETF) to look at how GMPLS would manage PBT (see PBT: Stray Thoughts); and BT has submitted information to the International Telecommunication Union (ITU) 's Study Group 15, which looks at optical and other transport network infrastructures. "The ITU is doing a lot of stuff around Ethernet OA&M that is directly relevant to PBT," says Kelly.

Meanwhile, PBT continues to evoke passionate debate among vendors. At the "Future of Optical Networking: Europe," a Light Reading conference held in London earlier this week, a panel of vendor executives was asked whether they thought PBT was more suitable for carriers than T-MPLS.

Massimo Leo, director of product marketing in the optical multiservice networks division of Alcatel-Lucent, was most forthright. "T-MPLS is better. That's it," he said. Though he did elaborate after the laughter subsided, saying that T-MPLS is already a standard at the ITU, "which is the proper place to do transport standards," and backed by the IETF, "which is the proper place to do packet standards."

He added: "PBT was initiated by Nortel, and now there are a number of vendors staring through the window to see what will happen."

One of those technology peeping toms is Ciena Corp. (NYSE: CIEN). Marketing director Vinay Rathore said Ciena "hasn't taken sides" because it's waiting to see what carriers want. "I see an advantage to PBT because of the simplicity of the technology, but both [PBT and T-MPLS] appear to do exactly the same thing."

Nick Cadwgan, director of product marketing at Meriton Networks Inc. , said his company can "see the need for a connection-oriented methodology. The IEEE is the home of Ethernet, and this [PBT] is a way of making Ethernet more carrier-grade. It's starting down the standards path now, but the customers will decide."

But Meriton's not throwing its weight behind just one camp. "PBT will actually use some of the exact same attributes as T-MPLS. We intend to support both," said Cadwgan.

— Ray Le Maistre, International News Editor, Light Reading

Page 1 / 6   >   >>
xbar 12/5/2012 | 3:32:25 AM
re: Cisco Tracks PBT Standards Process PBT is undependable as a technology in large SP environments. Yet another story from the NT transmission marketing group that does not have a clue in IP complexities.

Seems NT learned nothing from decades of industry IP routing and Ethernet switching experience.

xbar

digits 12/5/2012 | 3:32:24 AM
re: Cisco Tracks PBT Standards Process Re:
"If PBT has merit - customers will buy it. If it does not, they won't."

====
At the Wireless Backhaul event Light Reading held in London this week, one of the carrier speakers, from French mobile operator SFR (nearly 20 million customers) was asked whether he was waiting for standards to be pinned down for a technology he had talked about deploying etc.

His basic response was -- I tested it in my lab and it worked, I tested it in the field, and it worked. End of story.

Sure, becoming standardized will help any technology, but at the end of the day, if a technology works for a carrier, fits the business plan etc etc they'll use it.

PBT can be argued about until the cows come home, but let's see if anyone gets any purchase orders.

The feeling I am getting from industry debates is that the MPLS camp is REALLY worried that carriers will divert capex that would normally go on MPLS-based tech and towards PBT.

If there wasn't a real chance this tech could do the job for carriers, and if there was little chance that any Tier 1s would deploy it, PBT would not be one of the hottest new acronyms on the telecom block.

I expect to see PBT slammed in presentations and positioning papers. The fact that this is already happening means it's a technology worth tracking.

Ray
Say_Yes_2_MPLS 12/5/2012 | 3:32:24 AM
re: Cisco Tracks PBT Standards Process Ray,

Gǣone of the carrier speakers, from French mobile operator SFR (nearly 20 million customers) was asked whether he was waiting for standards to be pinned down for a technology he had talked about deploying etc.

His basic response was -- I tested it in my lab and it worked, I tested it in the field, and it worked. End of story.Gǥ

Did he perform the test using kit from different vendors? My guess is no if the standards havenGt been finalised. Getting equipment from different vendors that do adhere to the relevant standards to interoperate can be difficult enough due to different interpretations or things that are not fully defined.

Standardisation itself may not always be important, but interoperability is if you donGt want to get tied down to a single vendor. Standardisation doesnGt guarantee interoperability but it provides a greater probability that vendors equipment that meets the standards will interoperate.

It may make sense to go with a proprietary solution from a single vendor day 1 to gain a competitive edge. However, I would expect that most providers would be looking to standardise any technologies deployed as soon as possible in order to be able to negotiate better deals from alternate suppliers and to eliminate the risks involved in a single vendor solution. Unless of course the vendor is financially backed by the provider and maybe by the government, in which case the provider may not be interested in dealing with other suppliers and may not consider using them as their sole supplier a risk.

SY2M
Say_Yes_2_MPLS 12/5/2012 | 3:32:24 AM
re: Cisco Tracks PBT Standards Process xbar,

undependable = not worthy of reliance or trust. True trust in something can only be built up over time based on outcomes/results, although something may be considered worthy of trust before been given a chance to prove itself through association with (or guarantees from) a trustworthy source.

As pointed out by High-Tide, there are no large PBT deployments and therefore you are not in a position to be able to comment on whether or not PBT is dependable or otherwise. You may be of the opinion that Nortel is not a trustworthy source, and therefore may GspeculateG that Nortel PBT products are not worthy of trust. However, you do not have any evidence to support this as the Nortel products have not been deployed and tested in real world large scale deployments. You may also be of the opinion that PBT as a technology is flawed in some way, i.e. that it wonGt scale. Although there are a number of people that think we donGt need PBT because MPLS can provide the same functionality, I have not heard anyone say that they think PBT as a technology is flawed in some way.

In order to provide some useful input rather than negative statements that cannot be backed up, perhaps you could explain why it is that you think PBT as a technology will not scale in large SP environments (either technically or commercially)?

SY2M
High-Tide 12/5/2012 | 3:32:24 AM
re: Cisco Tracks PBT Standards Process Dude - how can you make a statement like that? If PBT has never been deployed in a large scale - which is hasn't - how can you make any supportable argument as to it's performance?

MPLS is overly complex - PBT offers an alternative method that would greatly simplify operational maintenance. I just read a stratistic that up to 30% of the cost of a carrier's network is OPEX - equipment costs were quoted as being typicaLLY 5%. (Network World - Dec 11th - page 31- Metro Ethernet: Where's the beef?). I did not necessarily agree with all the points in the article - but driving cost out of the network is clearly in everyone's best interest. If new services are easier to activate and manage, customers would likely receive faster and perhaps even more reliable service, as mistakes would be less likely. Lower OPEX would allow the carrier to offer more competative rates while maintaining acceptable profit margins.

If PBT has merit - customers will buy it. If it does not, they won't. I would trust their judgement - in the end, their's is the only one that truly matters.

NT has made significant contributions to the standards bodies over the past 20 plus years. Your remarks are apparently the result of not being informed, or choosing to ingnore those contributions.

HT
Mark Seery 12/5/2012 | 3:32:23 AM
re: Cisco Tracks PBT Standards Process >> Are you claiming that these are not issues in the IP/MPLS world? <<

Most IP/MPLS vendors would assert there have been numerous interoperability tests over the last few years, in addition to multi-vendor MPLS networks in actual networks. So some level of interoperability can be claimed, despite whatever issues may still exist (and here we have to consider both PWE3 and MPLS).

That said, I would assert MPLS was deployed long before it was standardized, and long before people outside of a small circle knew how to implement it. As many know, it is in fact a formal ethos of the IETF process that protocols are deployed (by multiple suppliers) before they are standardized. In addition, we can observe that even SONET/SDH was well deployed before interoperability (at all levels of functionality) was even attempted.

All new technologies come across a similar list of complaints and issues. The question is not does this list of issues have to be addressed, because it does, the question is whether the technology has any value.
digits 12/5/2012 | 3:32:23 AM
re: Cisco Tracks PBT Standards Process I should point out, for clarity, that the SFR speaker was referring to a technology that carrier is using in its wireless backhaul infrastructure, and NOT to PBT.

I was using his comments to show that technology can be chosen and deployed by a carrier because it works and suits the purpose.

Nortel clearly doesn't believe it can generate a lot of business with its PBT offering until its shown to be based on an industry standard. BUt I bet there are plnety of carriers prepared to deploy it before it's signed off...

Ray
Munster 12/5/2012 | 3:32:23 AM
re: Cisco Tracks PBT Standards Process Say_Yes_2_MPLS

Interesting argumentation. If I understand you correctly, you are saying that PBT should not be considered because interoperability is an issue and that one should not get locked into one vendor.
Are you claiming that these are not issues in the IP/MPLS world?
juddy 12/5/2012 | 3:32:22 AM
re: Cisco Tracks PBT Standards Process I am bit stunned as to how low Light Reading can stoop to and put the gun on Cisco's shoulder to fire. The article is all about Nortel's PBT with quotes from Nortel and talks about BT's interests...but the title is "CISCO Tracks PBT Standards Process" and there is one line snippet about Cisco.

Ray: I am wondering where did you learn the "art" of such cheap sensationalism...Poor Nortel folks..must be sitting there licking their wounds after supplying all the info to you while Cisco gets the +ve or -ve publicity.

Cheap...

Judd
desiEngineer 12/5/2012 | 3:32:22 AM
re: Cisco Tracks PBT Standards Process Ray: "I expect to see PBT slammed in presentations and positioning papers. The fact that this is already happening means it's a technology worth tracking."

I see. So damned if you do, damned if you don't. If we just ignore PBT, the industry thinks we have no credible response, so it must be good. If we respond, then the press thinks it must be important.

What a load of crap!

PBT in its basic form, the attention-grabbing, cheap and without the complications of routing, MPLS, OAM, traffic mgmt, el cheapo ethernet switch, is the germ of an idea. That germ is interesting only because it is cheap and simple, like static routes.

Flesh it out any, which you have to, for the size and scale of ethernet services that providers want to deploy, and it doesn't work. There is the pre-emptive strike to that already - a GMPLS control plane for PBT. Next, the new ASICs that do traffic management, the CAMs, the processors, the heat, the fans, the board design, and you have a switch router.

I don't say PBT can't work. I think it's too late. It should have come out 4 years ago. MPLS had to go a long way to get to where it is today. You can't fairly compare, say, Ipsilon's tag-switching with MPLS today. You know, cheap little Sun workstation Siamese twin conjoined to a fast switch router, blah blah.

-desi
Page 1 / 6   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE