Can Virtualization Enable the Next Big Bang?
The move to virtualization will enable large equipment manufacturers such as Alcatel-Lucent to move innovation into the commercial market faster by creating more "insertion points" for new technology and allowing good ideas to develop independently of existing products, the head of Bell Labs said today. (See Bell Labs Toasts Big Bang Discovery With Prize, Israeli Office.)
Marcus Weldon, president of Bell Labs and CTO of Alcatel-Lucent (NYSE: ALU), spoke to Light Reading in advance of today's 50th anniversary celebration of one of Bell Labs' biggest discoveries: the strongest evidence of the Big Bang Theory, identified by Arno Penzias and Robert Wilson, when they were trying to enhance radio communications for AT&T. (See AlcaLu Breathes New Life Into Bell Labs.)
As part of that celebration, Alcatel-Lucent is announcing a new Bell Labs Prize, an award of up to $100,000 for innovations addressing the big challenges facing the IT and communications industries. The award winners then work with Bell Labs to bring their ideas to market.
"This is an open call for ideas that are 10X game changers," says Weldon. Those who produce the best ideas in the annual competition will be teamed with an internal researcher to help refine their ideas. All of those will be presented to a panel and the winner announced in January.
Weldon, who took over Bell Labs in December, says the institution had become a bit too academically focused and needed to get back to focusing on the "grand challenges in the industry -- the things that confronted us in telecom which we now call ICT."
The advent of SDN and NFV plays into that shift in thinking in two ways: Not only will innovations be able to reach the marketplace faster through virtualization, but the myriad choices that virtualization creates require the kind of basic math and science solutions for which Bell Labs is justifiably known, Weldon says.
In today's telecom world, getting new things that come out of the research lab requires working closely with the existing product teams, because any innovations have to be embedded in the existing systems -- many of which are proprietary and monolithic, Weldon admits.
"There wasn't an obvious way that you could take a piece of the code or a functional module and adapt that functionality module and give it back to the product," he explains. " You had to deeply innovate with the product teams -- and that is a very time-consuming business."
NFV and SDN will separate functional elements as well as the control and data planes to allow for "more insertion points for innovation," Weldon adds.
"From a research standpoint, it means we could innovate and run a functional module -- which could run in the cloud next to the piece of software that is the ALU software and not have to deeply integrate it other than to service chain it together."
At the same time, virtual systems will have complex data plane processing requirements, he notes, and those are things Bell Labs has historically been able to address.
"SDN, when it is done properly, particularly between the IP and optical layer, is a complex optimization problem between sending a packet over an IP network or sending a packet over an optical network or both," Weldon says.
As SDN "escapes" the data center into the wide area network to become transport SDN, network operators will have many options for optimizing the performance of their network and compute capabilities. Multiple data centers connected by multiple networks, all with resource varying by the second in terms of availability and utilization create a massive math and algorithmic problem.
"Increasingly our IP and optical business lines are coming to Bell Labs and saying 'Hey this multi-layer optimization between IP and optical is really hard -- you guys are really good at that, help us out,'" Weldon says.
So Bell Labs researchers are deeply embedded in those teams just writing algorithms for how to compute the right path, when there are so many degrees of freedom.
The Bell Labs Prize hopes to identify people with expertise to share who either don't want to or can't develop their ideas on their own and team them with Bell Labs staff who are expert at bringing new ideas to market.
Penzias and Wilson were experts imported into Bell Labs to try to help AT&T improve satellite communications, working with satellites initially deployed on balloons and later in space. They had the skill and intellect to recognize that what they were noticing "off the side," as it were, were key clues to the origin of the universe.
"That's the kind of thing we want to be doing -- classic Bell Labs solving a real-world problem but having the diversity of disciplines, intellect and experience to see stuff off to the side that is remarkable and game-changing and insightful," Weldon says.
— Carol Wilson, Editor-at-Large, Light Reading