Rice Researchers Developing Optical Network To Master Big Data Management

• By Dian Schaffhauser
• 09/23/13

A grant-funded network initiative at Rice University will be researching development of high-performance, low-power, optical networking devices and programmable packet switches at the same time it provides a new way for researchers on campus to manage the piles of data they're generating in their projects. BOLD — short for "Big data and Optical Lightpaths-Driven Networked Systems Research Infrastructure" — received a $900,000 three-year grant from the National Science Foundation to enable this inter-disciplinary research.

The project is the brainchild of T.S. Eugene Ng, an associate professor of computer science and of electrical and computer engineering at the Houston university. The project will bring together the research disciplines of nano-photonics, networked systems, and big data applications.

"Experiments produce mountains of data, and there is often no efficient way to process that data to make discoveries and solve problems," Ng said. And the challenge goes beyond just moving data, he noted. "We also need to develop transformative ideas in the network control software, operating systems, and applications so that they can keep up with a faster network. Above all, for this network design to be appealing to industry, it has to be energy-efficient, scalable, and nonintrusive to the end user."

Ng and his fellow researchers are pursing the development of an energy-efficient optical network that can feed "rivers of data" to Rice's supercomputers.

BOLD will be a hybrid network that combines both electronic and optical switches. Optical switches have been around for years, but they operate differently from the electronic ones normally used in data centers, which means the two types aren't interchangeable.

Optical networking devices have several benefits, Ng said. They "consume very little power and can support enormous data rates, but they must first be configured, for example, by moving microelectromechanical mirrors into position, to establish a circuit. Electronic switches don't have moving parts, so they don't have that pesky delay."

Ng's idea is to create an optical switch without moving parts. These new "silicon-photonic switches" will be built by co-researcher Qianfan Xu, an assistant professor of electrical and computer engineering, whose specialty is creating ultracompact optical devices on chips — the nanophotonics part of the work.

On top of the devices will be a layer of operations that can "analyze data flow and demand, all the way up to the application layer, and dynamically allocate network resources in the most efficient way," Ng said.

Other researchers will tackle the work of optimizing network design and performance and developing the algorithms and testing for BOLD.

One of those testers, Rice computational and applied mathematician Bill Symes, does 3D seismic analyses, which routinely generates tens to hundreds of terabytes of intermediate data that must be loaded, cached, recalled, modified, and saved many times over. His work is especially attractive to oil and gas companies, which are continually seeking better models for translating seismic data to find the optimal drill spots.

"Results from the inter-disciplinary research enabled by BOLD will lead to future big data processing system architectures that dramatically speed up a wide range of computational scientific discoveries," the NSF said of Ng's research project. "Research enabled by BOLD could inform the design of future nation-wide networking infrastructures by showing how optical networks can be harnessed as shared 'cloud' resources."