SDN architectures

The Good, the Bad & the Distinctly Unwelcome

Light Reading readers have seen much change in the global communications market in recent years, mulled over and discussed in varying depths -- and in some cases it's been acted upon.

As a result, we've got too used to the rip-roaring debates about PBB/TE vs MPLS/TP and the minutiae of Metro Ethernet Forum (MEF) Ethernet definitions, slicing up what basically is bandwidth.

This is all great sport, if that's your thing, but it's hardly headline news.

However, the steady technological evolution of networks from TDM through to packet and now the mercurial software-defined networking (SDN) is now fueling debate like never before.

The gleeful revenge of the operators, as they look to take back the agenda from the traditional network equipment providers, is real and will fundamentally change how many think about architectures and services.

But what really sets this debate apart from previous ones is the way the cozy bit of banter between two old foes has been gate-crashed by a bunch of folk who not only have no respect for the fine art of telecoms, but are declaring that networks are too complicated and need to be completely rethought. (And that's if they thought about them in the first place.)

Now everyone is at it, with valuations that many of the "real" equipment makers can only dream about.

So where did it all go wrong, and what's next?

Some people are adopting a "better the devil you know" approach, while the telco traditionalists and vendors try to create new myths around the lack of performance of generic silicon and the sheer naivety of these new entrants.

But hang about, folks: During the past five years we've seen a bookseller with a Christmas rush problem become the largest cloud computing vendor in the world, a bunch of open source software cobbled together as the future of networking get sold for insane multiples, and the world's largest search engine starting to buy up fiber-to-the-home. They are akin to a bunch of unruly teenagers after their first drink. But ... they are exposing inefficiencies and busting myths through sheer bravado, ignorance and arrogance in equal measure.

The response to this is what some, but not all, have been doing for a long time -- be an engineer and be paranoid. If you see something and think it can be done better, faster and for less, then you're not alone. THINK hard and long about what it is you do and if change is working, embrace it.

But if you're not changing, you should ask yourself, how safe are you really?

— Matthew Finnie, CTO, Interoute Communications Ltd.

sam masud 8/16/2013 | 11:14:43 AM
Re: New blood - at long last We were told technology would make life easier and give us more leisure time, but I still don't see long vacation times (in the US) and/or people working fewer hours. Likewise, SDN, to succeed, will have to fulfill its promise of reducing capex/opex, but I have a strong feeling that it may actually raise the complexity of operating networks. The difference will be that without SDN, it might be far more costly to operate networks to provide the services that will be needed going forward.
brookseven 8/15/2013 | 12:13:57 PM
Re: Gleeful revenge.... So all that time Telcordia/Bell Labs/Western Electric were pawns of the vendors....

Good to know.



@mbushong 8/15/2013 | 10:56:27 AM
New blood - at long last I like that there is new blood, a little bit bold and maybe too brash at times. The networking industry has been stagnant because it has been driven by the same people with the same ideas (or tired derivatives) for more than a decade. When the bubble burst, the money left and went elsewhere. Without a real lure, it left the industry chock full of beards that were graying and growing. 

SDN, if nothing else, is bringing the money back in off the sidelines, and that is luring new, younger talent. And that is bringing in new ideas. The new ideas might not all be good - heck, they might not even be practical. But they force us all to be on our toes a bit. And that is good for everyone IMO.

-Mike Bushong (@mbushong)

Carol Wilson 8/15/2013 | 8:43:43 AM
Re: Gleeful revenge.... I'd agree that what separates SDN/NFV from previous efforts to move telecom networks out of the proprietary realm (I'm thinking about IMS, and, for us old farts, AIN and IN variants) is that there are other forces at play such as Amazon and Google that are forcing the hand of telecom network operators. 

I still want to see how the massive legacy architectures of telecom can be transformed in a meaningful way. As even Google is learning in its FTTH experiments, networks aren't easy. I hope it comes down to smart people doing the right things. 
Interoute CTO 8/15/2013 | 8:42:24 AM
Re: Gleeful revenge.... It's a wonderful wake up call for all concerned everyone including us and our kind needs to be a bit sharper and smarter. Looking forward to it.
Ray Le Maistre 8/14/2013 | 9:24:49 PM
Gleeful revenge.... Woah... don't hold back Matt! That sounds like a few years of feeling trapped by vendor constraints....

Will dealing with these new gate-crashers be any different though?
Sign In