x
MPLS

MPLS-TP vs. PBB-TE

5:20 PM -- It's only an opinion poll, but based on the results, it looks as if MPLS Transport Profile (MPLS-TP) could end up with a bigger role in the network than Provider Backbone Bridging - Traffic Engineering (PBB-TE).

The audience poll was taken during a connection-oriented Ethernet (COE) session at the Light Reading virtual tradeshow, Packet-Optical Transport Evolution, which happened today. Here's the tally:

Table 1: What's COE Look Like in Five Years?
PBB-TE in metro/aggregation, MPLS-TP in regional core
34.6%
MPLS-TP in metro/aggregation, PBB-TE in regional core
12.0%
MPLS-TP everywhere
30.6%
PBB-TE everywhere
9.3%
Other
13.3%
Source: Audience poll, LR Packet Optical Transport Evolution virtual tradeshow




Panelists Nick Cadwgan, product line management director of Gridpoint Systems Inc. , and Lars Pederson, CTO of Tpack A/S , noted that they expect many operators to use both technologies, fitting each to the appropriate networks that are already in place. Other interpretations are welcome, of course.

— Craig Matsumoto, West Coast Editor, Light Reading

Frank 12/5/2012 | 4:00:38 PM
re: MPLS-TP vs. PBB-TE

Craig, is there any possibility that you might describe the makeup of the audience responding to this poll, by industry segment (BBP, ILEC, ISP, FirstMileProviders, etc.), geographic locations, roles playd, and such? TIA.

Sterling Perrin 12/5/2012 | 4:00:37 PM
re: MPLS-TP vs. PBB-TE

Frank,


I was moderator on this panel at the virtual event. I don't have the breakout of responses, but it is always a mix of service providers and vendors from around the world.


I thought these results were very interesting. My read is that MPLS-TP will own the core without question while PBB-TE will have to fight it out with MPLS-TP for metro/aggregation networks. In metro/aggregation, the debate is still very much alive as both technology choices scored very highly. Given the knocks that PBB-TE has received over the past year or so (BT, NT, NSN, etc.), I thought it's showing in this poll was a very positive sign for the newly standardized technology.


Sterling

Frank 12/5/2012 | 4:00:35 PM
re: MPLS-TP vs. PBB-TE

Thanks, Sterling. Let me try this another way. If you'll agree to consider for a moment that the world is made up of two opposing views, those of netheads and bellheads, how would you say the crowd was divided, and to which flavor of transport did each group align? Or were there no netheads in attendance, to speak of, at all? Curious ... Frank

Sterling Perrin 12/5/2012 | 4:00:32 PM
re: MPLS-TP vs. PBB-TE

Well, we didn't ask attendees "Are you a bellhead or a nethead?" when they registered.


Are you thinking that bellheads will favor PBB-TE and netheads will favor MPLS-TP? Or are you thinking that PBB-TE and MPLS-TP are both bellhead choices and that netheads will go for IP/MPLS?


Sterling

Frank 12/5/2012 | 4:00:31 PM
re: MPLS-TP vs. PBB-TE

Yes, generally along the lines as you indicated in your last sentence. Perhaps with a little basic IPoverDWDM thrown in, too.


That's some pretty fun imagery, btw, i.e., asking registrants in advance if they are netheads or bellheads. Maybe next time you should do exactly that, and follow up by handing out complimentary headgear to match their allegiances as each attendee walks through the gate.

Kdavison 12/5/2012 | 4:00:30 PM
re: MPLS-TP vs. PBB-TE

Frank, I would argue that we are finally (after two years of debate) seeing acknowledgement across the industry –- be that Net heads, Bell heads, Optical, Ethernet or Router Vendors, as well as Service Providers -- that we must separate the service layer from the transport layer, while ensuring we operationalize capability across the two layers to support new IP based services/business models. In doing so, the focus on the transport layer has driven the convergence of packet and optical and the standardization of two “transport-based protocols” -- 802.1Qay (PBB-TE) and MPLS-TP -- to deliver a Data Transport Resource layer that, when engineered (using a combination of Traffic Engineering, Resource Management and Traffic Management), can deliver the key attributes of predictability and determinism while maintaining the flexibility of Packet/Ethernet. What the poll provides is yet another “datapoint” and confirmation that Connection Oriented Ethernet (COE)/Connection Oriented - Packet Switching (CO-PS) has moved from a debate on “why bother”, to “how do we deploy”. Given that, PBB-TE is now standardized and therefore can leverage Commercial Silicon being rolled out, driving down cost and improving interoperability. MPLS-TP, on the other-hand, is still in the requirements phase and although significant work is going on focused on the “framework & requirement specs”, we have yet to see the implementation details that will follow. Timing is the key issue, here in my opinion

 


 

If COE (CO-PS) is required in NG Transport (which appears to be the case) and needs to be  rolled out during 2010, I would argue the industry should leverage the standard COE (PBB-TE) in the Access/Aggregation and metro Core Network (due to Ethernet dominance, large number of nodes and access points) and focus on emerging/pre-standard MPLS-TP to address the Core Router Virtualization Interconnect issues (e.g. reduce non revenue generating ports for transit traffic), where MPLS Operational tools kits can be leveraged/extended into  the Transport Layer to operationalize MPLS-TP.

 

Both protocols provide similar (and potentially the same attributes) and may even leverage the same OAM (Y.1731) protocol; this is one example of the implementation details yet to be refined. However the key issue remains… both need to be engineered for transport.




HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE