x
<<   <   Page 2 / 8   >   >>
davallan 12/5/2012 | 3:39:34 PM
re: Verizon Also Shunning PBT If PBT solves a problem that other mature technologies solve, what then is T-MPLS and MPLS-TP?, other than a duplication of PBT attributes.

T-MPLS is dead in standards as a result of the IETF reasserting design authority. What the IETF will eventually produce, MPLS-TP, is several years away from producing a standard. PBT is the only interoperable technology available on a standards track in this space...

And yes ASICs already exist from several vendors. PBT has been "must have" on silicon for about 2 years. And I've not heard any complaints from the already deployed tier 2s....

D
majordomo 12/5/2012 | 3:39:34 PM
re: Verizon Also Shunning PBT 1. PBT is solving the Ethernet (Q-Q)scaling problem. This problem is not solved by MPLS, due to its inherent cost. Hence most aggregation networks are based on Etheret (Q-Q)
2. Most implementations today are based on NPs. Asics are just for price point (if nay) and not prohibiting the use of the technology.

Most carriers that today use Q-q with its limited managenmet will have much better experience with the management suits offered for PBT (e.g. Soapstone, Gridpoint)
spelurker 12/5/2012 | 3:39:33 PM
re: Verizon Also Shunning PBT FWIW, the automobile is a horrible analogy. PBT is a new technology trying to supplant other existing technologies. Compare it to replacing the internal combustion engine with something else (steam/rocket/fuel-cell/squirrel-on-treadmill).

Now the "solved problem" issue becomes relevant. A modelT and a Nissan Pathfinder are merely different models of the same thing solving the same problem, reusing the same solution. [same fuel, essentially the same spark plugs, dashboard gauges, etc.] Compare either of them to a hydrogen fuel-cell car, and now you need to spend time on standardization & infrastructure to support the new technology. The question of "is it worth it" is pretty natural, since the science/engineering mindset is to solve problems, [why solve the same one more than once when there are lots of other more pressing issues]

Mark Seery 12/5/2012 | 3:39:33 PM
re: Verizon Also Shunning PBT Two of the most falacious arguments in any market are "solved-problems" and "one trick ponies".

For example, consider this solved problem:

http://jeremydenk.net/blog/wp-...

There are many cars on the street and there would never be any new ones if people raised the "solved problem" flag every time someone tried to work on a new one - at some abstract level every car is a repeat of the same pattern: four wheels, internal combustion engine, steering wheel,....it is the differences that define cars, not the abstract commmonalities. I remember when the MEF started there were people saying that it had "no right to exist" because it did not add anything new. I remember people who said MPLS-TE was a solved problem because of the existence of SONET/SDH. Beware of the "solved problem" argument it often cloaks a prejudice and bias that is the road to the loss of liberty. What ever happened to the idea of letting a thousand flowers boom? If something adds no value the market will sort that out. MPLS-TE is a traffic engineering mechanism for MPLS, it is not a traffic engineering mechanism for Ethernet, and it is certainly not an address-based traffic engineering mechanism. If people want an MPLS solution instead of an Ethernet based solution, fine, make the argument. But only at the abstract level of four wheels, an internal combustion engine, and a steering wheel is connection-oriented Ethernet already a solved problem (it is clearly not a solved problem from the perspective of the IEEE - you know, the people who define[d] what Ethernet is).

A related argument is the one trick pony argument. Examples of "one trick ponies" are: WDM, LDP, Fiber, DSL, a manufacturing line set up to produce one product at high scale, a startup with one product, a software programmer that only knows "C" or Linux etc. It is an absolute irrelevancy whether something is a "one trick pony". There are many things which add value by specialization - in fact most of the people reading this would not be earning the money they do if not for some form of specialization. The relevant question is does the product have any buyers - once again a question the market is quite capable of working out for itself.

Net-net: compare and contrast what are the pros/cons of an MPLS solution versus an Ethernet solution. That is all fair game. Other arguments seem kind of falacious to me, but could be molded in to relevant points - for example instead of saying something is a solved problem someone could assert that the marginal cost of providing Ethernet service over an existing MPLS infrastructure is less than building a new Ethernet infrastructure. If true that would seem to be a highly relevant point for network operators (and of course at an abstract level it is the argument any new technology/product has always faced when trying to challenge an incumbent technology).
mbdi 12/5/2012 | 3:39:33 PM
re: Verizon Also Shunning PBT PBT has always been a redundant idea, and with BT out of the picture, the technology will die a rightful death.

After all, PBT minus BT is just pee.
jepovic 12/5/2012 | 3:39:32 PM
re: Verizon Also Shunning PBT "PBT will Gǣwin outGǥ if it is a superior solution (better scalability, higher resilency, better predictability, lower cost, and other) than alternatives"

Of course not. Don't confuse technical success with business success. The history is filled with technically superior solutions that have lost the battle for market dominance.

The car analogy is horrible for telecom, because of the interoperability issue. If Toyota designs a five-wheel car, it will still be able to drive on the same streets as four-wheel cars. Telecom doesn't work like that. Any new technology adds complexity, competence need, OSS cost etc. That's why any new tech needs to be radically better, not just slightly better, than the incumbent protocols.

So far, the PBT support from major vendors and operators is limited. Worse, the support is decreasing (BT, NokiaSiemens). Frankly, I don't think the technical issues matters very much anymore.
litereading 12/5/2012 | 3:39:32 PM
re: Verizon Also Shunning PBT PBT will Gǣwin outGǥ if it is a superior solution (better scalability, higher resilency, better predictability, lower cost, and other) than alternatives, despite the bashing by the MPLS crowd. MP3Gs are winning out over CDGs (as CDGs did over LPGs) because it provided a better experience at a lower cost. If PBT does the same, customers will eventually "get it".
Mark Seery 12/5/2012 | 3:39:32 PM
re: Verizon Also Shunning PBT >> The question of "is it worth it" is pretty natural, since the science/engineering mindset is to solve problems, [why solve the same one more than once when there are lots of other more pressing issues] <<

Right. The problem of how to move people from point A to point B over a paved road has already been solved. Let's stop all new car designs because it is already a solved problem.

Is it worth it? I assume you mean, given the fact that I already have feature requests up the ying yang for IP/MPLS how do I prioritize PBB-TE and/or I am bored with traffic engineering and I want to work on something else? Perhaps all relevant exestential questions at the level of an individual product manager or engineer.

However that is only one of the relevant perspectives. The reason why markets provide multiple ways to "solve the same problem" is to provide choice. Choice because competition leads to improvements. Choice because consumer tastes and preferences are not the same. Choice because there are different niches and habitats in any ecosystem.
davallan 12/5/2012 | 3:39:32 PM
re: Verizon Also Shunning PBT Hi Spelurker

Calling PBT a "technology" is generous. It is a small feature added to Ethernet that takes advantage of other Ethernet standards (PBB, OAM etc.) to significantly broaden the addressable market for Ethernet...As I've said before, it is more a comment on how far Ethernet has come....

As for what PBT specifically does, as I noted in my other post, standardizaed competitors do not exist as of yet, they are still on the drawing board. PBT already exists...and interoperability has been demonstrated for some time...and standardization is well underway.

Given the IETF jointly with the ITU are just now starting to work on solving the same problem, your question "why solve the same one more than once..." perhaps should be redirected. ;-)

D
Mark Seery 12/5/2012 | 3:39:31 PM
re: Verizon Also Shunning PBT Jepovic,

What level of interoperability are you referring to? The road/fiber or the service/maintenance function? You think new cars don't generate changes to inventory systems, training requirements, new manuals, new supplier recommendations even when there are interoperable parts (and there are not always),....

I've already stated the important question might be the incremental cost for an operation/organizational unit within an operation. OSS is obviously part of that equation. That is a whole different discussion than whether something is a solved problem or not. It is all part of the question of whether it solves a problem in a better way or not.
<<   <   Page 2 / 8   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE