Follow Datanami:
September 24, 2013

Rice University Deploying BOLD Optical Network Strategy for Big Data Processing

Isaac Lopez

Researchers at Rice University have a BOLD new networking strategy for handling the massive data sets that are generated in its distributed research labs. The new plan involves the creation of a new customized hybrid optical network – which includes a new type of optical switch engineered at Rice.

Funded by a grant from the NSA, the project, dubbed BOLD, may have future application for enterprises. BOLD, which is short form for “Big data Optical Lightpaths-Driven Networked Systems Research Infrastructure, will culminate in an optical big data highway, with paths ultimately leading to Rice’s supercomputers.

“Advances in computing and sensing technologies have led to a similar problem across many disciplines in science and engineering today,” said T.S. Eugene Ng, associate professor of computer science, and principal investigator for BOLD in a recent article. “Experiments produce mountains of data, and there is often no efficient way to process that data to make discoveries and solve problems.”

In order to address these challenges, Ng says they’re working on what his group believes to be transformative ideas that attack the problem holistically, from the network control software down to the operating systems and applications so that they can keep up with the fire hose of data coming into their environments. “Above all,” says Ng, “for this network design to be appealing to industry, it has to be energy-efficient, scalable, and nonintrusive to the end user.”

The fabric of BOLD will come from an array of technologies, including optical data-networking switches, as well as standard mechanical switches. Both have their drawbacks and advantages, says Ng. Where electronic switches are generally plug and play, optical switches have intricacies built into them that add a level of complexity for their transfer pay-off.

“There’s a trade-off” says Ng. “Optical networking devices consume very little power and can support enormous data rates, but they must first be configured, for example, by moving microlectromechanical mirrors into position, to establish a circuit. Electronic switches don’t have moving parts, so they don’t have that pesky delay.”

Interestingly, there is a third of switch that will be added to the infrastructure – a new silicon-photonic switch engineered at Rice built in the laboratory of co-principle investigator, Qianfan Xu. Professor Xu, who is an assistant professor for the Department of Electrical and Computer Engineering at Rice, will be responsible for building the switches, which they say don’t have the moving parts that traditional optical switches have.

To take advantage of the speed boost that they are hoping to affect, Ng says that they researchers at Rice are working on a new intelligent layer that can analyze data flow and demand (all the way up to the application layer), and dynamically allocate network resources as efficiently as possible.

The project, which is kicking off presently, is expected to run over the next three years thanks to the NSF funding grant, but Ng says they hope the project will have lasting implications at the university, not to mention the industry at large.

Related items:

CERN Turns to Google for Datacenter Direction 

Interplay between Research, Big Data & HPC 

Big Science and the Gift that Keeps on Giving 

Datanami