Optical/IP Networks

BT Rethinks 21CN Core Strategy

BT Group plc (NYSE: BT; London: BTA) is reconsidering its 21st Century Network (21CN) core network technology deployment plans, and may add a controversial new Ethernet approach to its IP/MPLS strategy, an executive from the British operator revealed today. (See BT Closes 21CN Deals, Touts IPTV.)

Speaking at the Carrier Ethernet World Congress in Madrid, Tim Hubbard, head of 21CN solutions strategy at BT, said the carrier is planning to use Provider Backbone Transport (PBT), a new flavor of Ethernet being heavily promoted by BT in league with Nortel Networks Ltd. , as a backhaul transport technology between its access nodes and metro nodes.

That will come as no surprise, as BT and Nortel have been jointly presenting PBT to the telecom standards bodies, and the carrier has been vocal about its support for the technology. (See BT Likes Nortel's New Ethernet Flavor.)

PBT, in a nutshell, strips out some of the complexity of Ethernet and enables service providers to create a managed, traceable point-to-point Ethernet tunnel across a network that provides SDH-like performance, so making Ethernet more manageable and scaleable.

This is achieved, using existing Ethernet switches, by turning off certain Ethernet functions, such as Spanning Tree Protocol and MAC learning, and deploying Ethernet OAM (operations, administration, maintenance) functions that are currently going through the Institute of Electrical and Electronics Engineers Inc. (IEEE) and International Telecommunication Union (ITU) standards bodies. (See Nortel Launches PBT.)

But Hubbard revealed that BT is considering using PBT in a broader context. He said the carrier had issued an Ethernet ITT (invitation to tender) -- a.k.a. an RFP (request for proposal) -- earlier this year because BT "is asking itself, do we need a complete IP/MPLS core? Since the arrival of technologies such as PBT, it is possible to use that in the core and reduce" the reliance on IP/MPLS. (See BT Issues 21CN Ethernet RFP.)

Such a move is still feasible for BT's 21CN buildout, stated Hubbard, as "our 21CN IP/MPLS core network is not fully deployed yet -- our options are still open."

One of those options is to "extend the Ethernet backhaul to give a single Ethernet infrastructure layer end-to-end," a move that would have significant implications for the 21CN core capital expenditure, said Hubbard.

And, presumably, be not such great news for BT's two 21CN core router vendor suppliers, Cisco Systems Inc. (Nasdaq: CSCO) and Lucent Technologies Inc. (NYSE: LU), which is providing Juniper Networks Inc. (NYSE: JNPR) gear. (See Lucent, Juniper to Supply BT 21CN and Cisco Picked for BT's 21CN.)

Such a move would also send shock waves around the global carrier community, which has, to date, reached a consensus that the world's next-generation networks will have IP/MPLS cores at their hearts.

But BT's motivation is not just the potential capex saving. Hubbard says there are some utility companies that require absolute iron-cast SLAs for alarms, such as the electricity and nuclear power generating companies, and the carrier is looking to pin down its guaranteed response times so that, for example, there is no chance of an alarm delay that could result in an explosion at an electricity generator plant. "There could be lives on the line over this," he stated.

Hubbard says vendor submissions for the Ethernet ITT are in and are being assessed. Vendors here in Madrid see Nortel as the obvious vendor to benefit from any PBT deployment.

The response here from delegates to BT's 21CN core rethink was one of surprise, and, from the MPLS supporters, incredulity. The PBT versus MPLS debate looks as if it will dominate proceedings during the Ethernet event. "I sense that the MPLS community is a bit troubled by PBT," stated Mark Lum, an independent consultant who was chairing the event's proceedings.

And the MPLS community could be in for more strife, as it seems BT is not the only Tier 1 carrier cozying up to PBT's charms. In presentations made at the Broadband World Forum Americas event in Vancouver last week, Verizon Communications Inc. (NYSE: VZ) and BCE Inc. (Bell Canada) (NYSE/Toronto: BCE) both expressed an interest in PBT in presentations made to delegates.

No surprise, then, that Nortel is not the only vendor here in Madrid espousing the benefits of PBT. Siemens Communications Group talked up the technology during a presentation today; packet switching vendor Tpack A/S has issued a white paper on the technology; and Extreme Networks Inc. (Nasdaq: EXTR) has expressed vocal support.

— Ray Le Maistre, International News Editor, Light Reading

Page 1 / 14   >   >>
dlock 12/5/2012 | 3:16:25 PM
re: BT Rethinks 21CN Core Strategy Lao,

>I am stuck between being too young to know, and too old to remember exactly why IP is CL-PS; ...is there anything can not be done without CL-PS in the core today?

How about the Internet? - for instance, jumping from one web-site to another, not even noticing how many connections you are making or where to. CL-PS buys you a stateless core network, which scales well. Your PC sees your connections but the network doesn't. As an alternative, imagine your PC dialling up (i.e. asking the network to create and maintain) a discrete connection to each of those web servers.
zmding 12/5/2012 | 3:39:39 AM
re: BT Rethinks 21CN Core Strategy I do not think there is "Ethernet over PBT". PBT is a modified version of native Ethernet e.g. by disabing STP or MAC learning, using MAC+VLAN ID as a label, applying Ethernet OAM, etc. Comparing Ethernet over MPLS, operators can save money by getting rid of the IP/MPLS layer.
I don't think PBT can achieve IP aware TE while it can achieve TE by using the label(MAC+VLANID) to set up various tunnels.
I am not sure about the third question. I am also very interesting in this PBT vs. MPLS debate.
JimKnopf 12/5/2012 | 3:39:39 AM
re: BT Rethinks 21CN Core Strategy From my point of view there is no way that PBT makes MPLS redundant - what it could make redundant is Ethernet over MPLS ...

Of course, PBT cannot support RFC2547bis, given that this RFC is used to implement IP VPNs and PBT does not look into the IP layer.

So, I believe PBT (if deployed) will coexist with MPLS. PBT can potentially be more cost-efficient as compared to Ethernet over MPLS, but future will have to show, if this really is true. However, it's definitely worth it to investigate connection-oriented Ethernet approaches (among them PBT) in more detail, because Ethernet transport services will become more and more important.

The interesting part with this is, if there is going to be a control plane on these approaches, and if yes, which mechanisms will be applied there ...
farfields 12/5/2012 | 3:39:39 AM
re: BT Rethinks 21CN Core Strategy I was under the impression that the main driver for development of Ethernet OAM was so that the telcos could have awareness of faults in the access network and at the demarcation to user equipment. The core network has sufficient OAM already, so why the noise about IP/MPLS? Are they trying to scare their current vendors to pay attention to tweaking the products to suit the current OAM overhead?
EthernetFan 12/5/2012 | 3:39:39 AM
re: BT Rethinks 21CN Core Strategy I am a little confused about PBT and its capabilities.

Can someone explain to me if PBT can accomplish following?

1. How " Ethernet over PBT approach" is able to accomplish LOWER CAPEX compared to "Ethernet over MPLS" approach ?
2. Can PBT achieve IP-based or IP-aware TE (IP Source Destination Addresses, Diff. Serve bits etc.) a la an MPLS LER device can do ?
3. Can PBT support similar services such as MPLS RFC 2547bis VPNs ?

If the answer to these questions are YES, then why DO WE NEED MPLS ???

If Ethernet is able to step up to being carrier grade with similar capabilities as MPLS, Ethernet wins, as it has consistently demonstrated since 1974. Isn't ???

laserboy 12/5/2012 | 3:39:38 AM
re: BT Rethinks 21CN Core Strategy SolitonWave, please get informed regarding PBT since it is real and it is a serious challenger to MPLS in certain parts of the network. Do you honestly believe that BT, after a serious investigation and comparison, is so uninformed regarding PBT? Many that have knocked Ethernet and its evolving incarnations have learned the hard way that Ethernet is an unstoppable freight train. PBT makes sense and it is real.
SolitonWave 12/5/2012 | 3:39:38 AM
re: BT Rethinks 21CN Core Strategy PBT is good Nortel Marketing... They took the concept of VLAN cross-connecting and added some nice technical buzwords. Luckily for them BT got hooked on it for some strange reason...
This is no challenger to MPLS. No fair comparison is possible...
sgan201 12/5/2012 | 3:39:37 AM
re: BT Rethinks 21CN Core Strategy If the goal is to deliver bits at the lowest possible cost, then, the cheapest mean is the best way. At the core, this seems to be true.

1) If the input and output to the core is Ethernet, why convert it to anything else?? Conversion costs money. Yes, you need OAM. But, other than that. You do not need much.

2) IP awarenes.. The edge box aka customer's router can choose to prioritize their IP traffic and send it out via different VLAN ID. This keep the service provoder's job easier and the customer can choose to prioritize the traffic anyway that they like.

What is MPLS supposed to be good for?? Where is the valued added for MPLS with all the complexity??

giles0 12/5/2012 | 3:39:36 AM
re: BT Rethinks 21CN Core Strategy so you build a big mesh of VLANs/PBT circuits between your PE routers?

hmmm... sounds rather like an architecture we tried about a decade back with ATM switches :(
davallan 12/5/2012 | 3:39:35 AM
re: BT Rethinks 21CN Core Strategy As PBT is a profile of Ethernet that can run side by side with other profiles on a bridge, the question shouldn't be PBT (in isolation) vs. MPLS, it should be "can Ethernet assume a lot of the role of the MPLS PSN?"

IMO Yes...
Page 1 / 14   >   >>
Sign In