Ethernet equipment

Verizon Preps God Box RFP

CHICAGO -- NXTcomm -- Verizon Communications Inc. (NYSE: VZ) is set to issue a major telecom infrastructure RFP (request for proposal) to the equipment community for a multifunctional, integrated optical/packet platform that could include support for PBT (Provider Backbone Transport), according to sources on the show floor here.

And those sources say that three companies have already been identified by the operator as the likely frontrunners: Alcatel-Lucent (NYSE: ALU), Fujitsu Ltd. (Tokyo: 6702; London: FUJ; OTC: FJTSY), and Nortel Networks Ltd. .

The carrier has supposedly been close to issuing the RFP for some time, but an internal battle about whether to request support for PBT, the controversial new approach to carrier Ethernet transport, has held up the document's issuance, sources say. (See PBT: New Kid on the Metro Block, PBT Stars at Ethernet Expo , PBT's Ethernet Appeal, Poll: Opex Gives PBT an Edge, and PBT Means What?)

The sources say Verizon is looking for technology that will incorporate a wide range of optical features as well as integrated packet switching. That gibes with recent developments in the vendor community, where more and more vendors are talking about such combined optical/packet platforms. (See Nortel's Noisy With Optical Opportunity, Cisco Shows Some Optical Love, Alcatel-Lucent Gets Tropical, Ciena Continues Ethernet Push, Fujitsu Updates the MSPP, and Meriton Tackles Ethernet Transport.)

The sources add that this is Verizon's follow-up to its initial next-generation optical transport rollout, an engagement that was bagged by Tellabs Inc. (Nasdaq: TLAB; Frankfurt: BTLA) last year. This, in effect, will be the carrier finding a second source for new optical technology. (See Tellabs Victorious at Verizon.)

The big question on the show floor here is not when the RFP will be issued -- vendors are expecting it soon -- but whether the RFP will include a request for PBT support. Sources say Verizon is still deciding and hearing arguments from backers of carrier Ethernet technologies and supporters of a dedicated Multiprotocol Label Switching (MPLS) approach. Verizon has been a major purchaser of MPLS technology to date. (See Verizon Taps Layer 2 and Verizon Expands MPLS Plans.)

The inclusion of PBT would play towards Nortel, which has been the lead PBT evangelist in the vendor community, having worked on the technology with BT Group plc (NYSE: BT; London: BTA) for nearly two years. (See Nortel Pushes PBT Pact, Nortel Says 40-Gbit PBT Coming Soon, BT Likes Nortel's New Ethernet Flavor, and BT Rethinks 21CN Core Strategy.)

BT has already committed to deploying PBT in its 21CN next-generation network, and is helping to push PBT through the standards process at the Institute of Electrical and Electronics Engineers Inc. (IEEE) , where it is being called PBB-TE (Provider Backbone Bridge Traffic Engineering). (See Nortel, Siemens Win PBT Deals at BT, Nortel on PBT: Today BT, Tomorrow the World!, and Cisco Tracks PBT Standards Process.)

Some of the sources believe Verizon, which has reportedly been talking extensively with BT about how and why it's deploying PBT, will include PBT to at least keep its options open. That would also explain why Nortel has been identified as one of the three frontrunners in picking up the resulting business.

PBT's inclusion, though, is not assured. The alternatives would be that Transport MPLS (T-MPLS), which also enables protected point-to-point transport tunnels, could be requested instead, or that support for both PBT and T-MPLS might be requested by Verizon.

Those scenarios would favor Alcatel-Lucent and Fujitsu more. Fujitsu says its newly released Packet Optical Networking Platform supports both PBT and T-MPLS. Alcatel-Lucent is very much in the T-MPLS camp, but despite describing PBT as a "peripheral" technology, it has left the door open for PBT support, saying its overall position is "neutral." (See Fujitsu Updates the MSPP and AlcaLu: PBT Is Peripheral.)

Verizon declined to comment for this story.

— Ray Le Maistre, International News Editor, Light Reading

Page 1 / 5   >   >>
^Eagle^ 12/5/2012 | 3:06:30 PM
re: Verizon Preps God Box RFP I should point out a small but important inaccuracy.

To state that Tellabs won Verizon's optical transport network business is misleading. LU has a large part of VZ optical transport. Perhaps more installed base in leading edge DWDM than at VZ than Tellabs due to the LH contract LU had.

What Tellabs won for optical transmission was the METRO portion of the network. (of course this in addition to Tellabs wins at the edge for DLC terminals and PON systems).

In fact, I believe one of the motivators for ALU to pick up Tropic was because they lost the Metro DWDM RFQ to Tellabs.

Lucent was & is the primary supplier to Verizons Long Haul backbone network that they began building out before the MCI acquisition. Despite some thoughts to the contrary, where many felt that the buyout of MCI and MCI's fiber backbone would lead to Verizon ending and curtailing it's own backbone build out, Verizon has continued with that build out. One can suppose that the Verizon LH backbone and MCI's network was needed to give them the coverage and reach they needed / wanted.

So the incumbent for a large part of the VZ optical network is ALU in the form of Lucent platforms.

The rest of VZ installed optical networks have various incumbents. The VZ metro DWDM network prior to the recent award to Tellabs for the upgrade and new build outs was powered by boxes from Fujitsu, Lucent, and others. Tellabs had the core Digital Cross Connect business with some DCC's from Alcatel in the form of DSC boxes and Alcatel's own boxes, and SONET transport was Fujitsu and Lucent with some other vendors mixed in.

MCI's DWDM backbone is largely Nortel 1600G platforms deployed during the boom with newer stuff from other vendors ( I think I heard Siemens.. but not sure)

This new RFP is interesting therefore from several perspectives.

Different incumbents with unique advantages based on installed base, technology niche they currently dominate at VZ and long time relationships as suppliers. Also these incumbents all have the advantage of having their network management systems already integrated into Verizon's systems management and control.

Despite ALU's noise about neutrality on PBT, I would be surprised if ALU did not bend over backwards to increase their footprint at VZ in this deal and if VZ wants it, ALU will try to give it to them. They have the LH part via Lucent as stated above. Also would be surprising if NT did not use this as a tool to get back into control of one of their long time major accounts MCI. And Fujitsu will be very aggressive.

I don't think any one else stands a chance outside ALU, NT, or Fuji.

jes 12/5/2012 | 3:06:29 PM
re: Verizon Preps God Box RFP Where did Cisco stand with its optical lineup ? 15454 can do optical/packet.. but not the PBT...
comments ???
ipLogic 12/5/2012 | 3:06:29 PM
re: Verizon Preps God Box RFP 1 or 2 or 3 ?

1. If PBT (as promissed) can be build out of existing (off the shelf) hardware, then why should any vendor have problem implementing it with anything more than small Sw patch on the L2 switches ?

2. If PBT does require any specific Hardware and Sw developments, will it then keep to its promise to be cheap solution based on off-shelf hardware ?

3. Or maybe PBT is not yet complete solution, but rather just small marginal piece waiting for the *Big part* of the overall solution to be yet delivered ("invented" by reinventing some other technology).

So, either it is *nothing new* but rather reduced (not self sustained) technology of something that everybody already have, (someone trying to get famous by reinventing the wheel)

or, it is *something new*, and will need *overall* industry acceptance to deliver its promise of "being cheap". (someones credibility promising is in question)
(or the "invention" has weakness being dependent on industry acceptance before it delivers its promises)

or, it is still in *experimental* phase and will need Time and *Risk* of the ones that invest in it by not being able to realize the Real Total Cost of Ownership of such solution.

i mostly beleive the question is funny,
like "can proven BGP core router do Static routes and IP lookup ?"
(where everybody knows this is "baby" part of any serious router),

but then the much more important question follows "can you do Internet routing just with Static routes ?" (or how *serious* is this solution afterall)

t.bogataj 12/5/2012 | 3:06:29 PM
re: Verizon Preps God Box RFP Cisco made it very clear several times that they would keep an eye on PBT -- but would not support it.
t.bogataj 12/5/2012 | 3:06:24 PM
re: Verizon Preps God Box RFP "Can you do Internet routing just with static routes?"

Thanks, ipLogic, for finally posing the right question (comparing PBT TE to static routes). PBT's control plane, its manageability & scalability is THE key issue with PBT deployment. Using, say, G-MPLS for PBT TE makes PBT far from cheap and simple.

"1 or 2 or 3?"

As usual, the only correct answer is 42.

Putting it in your words, PBT is nothing new technically... just an excellent pact between N. and BT.
CardinalX 12/5/2012 | 3:06:23 PM
re: Verizon Preps God Box RFP I should point out there's a big difference between routing requirements for the Internet and routing requirements for transmission & transport. For a start the Internet is connectionless and transport networks are connection oriented.

Internet routing is highly dynamic, a fault on the other side of the world can change the routing on my AS. Transport routing is much more static. The duct & fibre network topology changes very slowly and we use protection triggered by OAM rather than reroute triggered by routing protocol timeouts to protect against faults in the transmission network.

The difference between Internet routing and Transport routing, in non-technical terms, is like the difference between planning a route when driving a car and planning a route when building a road.
open-minded 12/5/2012 | 3:06:23 PM
re: Verizon Preps God Box RFP t.bogataj,

Clearly the only word of advise I can offer to you is to educate yourself a little bit more on Ethernet OAM (.1ag/Y.1731) before making sweeping statements on the OAM capabilities.
t.bogataj 12/5/2012 | 3:06:23 PM
re: Verizon Preps God Box RFP CardinalX,

thanks for your thoughts.

However, the point that ipLogic made was not whether or not there were differences between "Internet" and "Transport" routing. The point was in an analogy between L3-routing and L2-learning (and restoration).

In non-PBT (i.e. connectionless) Ethernet, the network _learns_ how and where to forward frames, and does not need a complex TE (control plane) to administer the forwarding tables. PBT, on the other hand, takes all the "brain" out of Ethernet network and moves it into an (abstract) management layer.

In IP (and IP/MPLS, for that matter) networks, the network uses routing protocols to "learn" how and where to forward packets, and does not (necessarily) need manual management of the routing tables.

What PBT does with excluding learning capabilities from Ethernet, is analogous to using *none* of the routing protocols in IP networks.

Now, can you imagine the Internet to run without routing protocols, only with static routing entries? Can you really?

Can you then imagine a huge Ethernet network with hundreds of thousands of "connections" (i.e. tunnels) managed exclusively manually? I bet you can't. Which is exactly why PBT needs a complex TE (control/management plane)... say, a little help of G-MPLS.

Dinamics of topology changes does not really matter in this analogy (although I don't agree with your statement on higher dinamics of internet routing). What really matters is scalability and manageability.

By the way: OAM-triggered restoration is completely irrelevant for PBT. PBT creates point-to-point "connections", so it cannot make use of 802.1ag CFM since it uses multicast. Hence... no OAM capabilities in PBT.
CardinalX 12/5/2012 | 3:06:22 PM
re: Verizon Preps God Box RFP t.bogataj,

My point is that any analogy between connectionless IP routing and connection-oriented routing will be weak and we should not assume both have similar requirements that can be met with the same technology.

I don't have to image a transport network with hundreds of thousands of statically managed connections - I can see them in many a SDH or SONET network. Obviously these are not manually managed but they are statically routed and managed via an OSS.

I agree manageability & scalability is important and G-MPLS implementations are too constrained on both counts.

It takes little development to do 802.1ag on unicast addresses required for PBB-TE protection. ITU Y.1731 already includes unicast OAM.
materialgirl 12/5/2012 | 3:06:21 PM
re: Verizon Preps God Box RFP LVLT management talks about how they ditched those pricey L3 routers in favor of L2+ switches that "follow Ethernet price declines of 20% per year and cost 1/10 as much". I assume they are not using PBT, but are using some form of Ethernet. Since their network is large, and seems to work, I wonder what they are up to? If they are using Ethernet, how do they make it work, while others have to go to PBT?
Page 1 / 5   >   >>
Sign In