Improvements, yes. 50 to 100 times? At what price reliability and consistency? We ought take certain things with a very large grain of salt. SCMR has a vision of "liquid bandwidth," which I happen to find appealing and agree with for the most part.
But even they admit that the traffic engineering and algorithms needed, along with network awareness through real time surveillance and capacity forecasting to perfect mesh with optimum efficiency, is still a ways off.
The liquidity they seek will come with some degree of unpredictability and risk, when anticipating and meeting dynamic traffic demands, when all of the pipe's constructs begin to approach max fill (which means that they've reached their limit of bandwidth on demand, under any given set of circumstances).
So, what do they do? They reserve certain amounts of bandwidth for priority customers' use, the ones who demand all of the nines. And at some point this is the same as having a sonet standby ring facility (the counter-rotating part that is idle) in the waiting, in case there is a failure on the primary route (which is one of the criticisms of SONET, and one of the ways that SCMR will improve efficiencies through the use of mesh).
Even mesh doesn't allow you to get something for nothing. But it's a step in the right direction. If mesh network operators (such as today's ISPs, for example) can garner enough customers with lowered expectations in terms of QoS and class of service, CoS (which will allow them to oversubscribe their network to more profitable levels), who will wait in line, so to speak, when things get congested, then they will reduce their risks and make out okay. Offsetting this optimism is the hope that many have that their internet connections will be real time, or near real time, in the future... to support voice and video. The IETF is working on this, with the hope of being able to satisfy most users' preferences in the future. But they ain't there yet. Not be a long shot.
Having said all of this, it's useful to remember that despite their claims centering on mesh, they are still able to build networks the old way for those who prefer SONET rings, and still provide mesh to others.
More on "liquid bandwidth" below. While the article below highlights ODSI [which SCMR has had a hard time selling to the rest of the industry], the concepts which define liquid bandwidth apply in a similar way to other optical frameworks that are under consideration, too.
zdnet.com |
