There's a certain amount of pressure felt by every academic researcher.
"We've gotten a lot of attention from outside sources for our work," said Peter LoPresti. He's referring to collaboration with Hazem Refai on projects to improve wireless communications -- especially during natural disasters or other catastrophes.
It's a point of pride for research to be influential. But LoPresti isn't boasting, as he recognizes the double-edged sword when advanced work attracts the attention of others.
"We would like to stay near the front of it. That's how you stay relevant. That's how you get grants," LoPresti said. Others have one advantage: "They've got access to a lot more computing resources than we do," he said.
Hazem Refai and Peter LoPresti
That could change in a matter of weeks, however, with the opening of the Tandy Supercomputing Center in Tulsa's city hall. A ribbon-cutting ceremony is planned for Thursday, May 23, with skilled researchers like Refai and LoPresti getting to use the computer before it's opened up to private industry and other users this summer.
The project is overseen by a collaborative effort known as the Oklahoma Innovation Institute.
The cost, along with a few other related programs, is roughly $6.75 million, according to information from the institute, a not-for-profit endeavor that's received financial support from philanthropic groups and the federal government as well as local colleges and universities.
"We do have support from the economic development administration to help build the actual supercomputer, all the way through to the city of Tulsa to give us a good deal, so to speak, to have space in One Technology Center," said David Greer, the institute's executive director.
The institute is paying $90,000 yearly to rent space for 12 racks for the computer and six cooling units, according to the city. The deal is a 60-month lease. "We have a separate meter and we pay our actual electrical expenses," said Mike Partain, a board advisor to the institute.
The spirit of collaboration has impressed researchers like Refai and LoPresti, who know first-hand that universities often benefit from partnerships with industry.
"If you look around the nation, I don't think you're going to see a partnership developed such as this," Refai said.
About two years ago, when first approached about the project, he admitted that he didn't see a critical need for it. The University of Oklahoma-Tulsa professor has devoted his career to researching better ways to operate wireless technology. At that time, Rezai said he felt he could make do with a supercomputer operating at OU's main campus.
But that's changed, he said.
"The bandwidth on the Norman supercomputer is becoming less and less available to us," Rezai said. LoPresti, an electrical engineering professor at the University of Tulsa, said that more disciplines are "tackling bigger problems," describing how TU's biology researchers collaborate with engineers on an array of sensors to track how pollen travels.
The pair has their own big problem they're trying to solve, one which emerged in some of the darkest days of American history.
"We had 9/11, and we had Katrina," said Refai. Along with the loss of lives, those tragic events highlighted flaws in wireless communications, with frantic but ineffective radio calls made in times of crisis. "There was no common communications among first responders," Rezai added.
Rezai and LoPresti collaborated on a way to help fix the problem by bolstering an alternative to mainstream wireless communications. The now-ubiquitous Wi-Fi networks, for example, use radio waves.
But light beams can also transmit highly complex information.
They wrote a proposal -- eventually funded by the National Science Foundation -- to improve the accuracy of those light beams in terms of being able to interpret the light signals well enough to communicate and process useful information, like voice communication.
It's far from a simple problem, as the idea behind what's known as free-space optical networks involves having balloons or unmanned aerial systems carry equipment into the sky. These airborne stations then transmit light to each other and then to fiber-optic receivers. Computers capable of translating the information into data are the final stop for the signal. In this way, first responders may be able to talk to each other across great distances. Video and images can also be transmitted in an optical system, which may have more bandwidth than traditional networks used currently.
But having a system deployed in the air makes it difficult to be able to point the beam of light -- which may be a foot wide -- to where it needs to go. Rezai and LoPresti have been working on outfitting receiving technology with arrays of optical fiber.
The research goal, simply, is to "relax those conditions on how accurate you have to point," LoPresti said. Ultimately, a successful system will work even if a light beam isn't able to be aimed precisely -- or weather conditions vary -- so "you could still send some data through it," he said. "That way, you could drive a truck in and stick this thing -- tether it to a pole -- and within a few hours have a communication system running with the bandwidth of optical."
Many things affect the light beam, including the instability of small aircraft holding it as well as atmospheric conditions that affect the power of the beam.
The pair uses mathematics to calculate how these factors influence the light that gets to the fiber-optic receivers.
"It's not a complicated model, but it needs to be very accurate," LoPresti said. To even get even one data point on a personal computer, "this could be five hours, because of the level of detail, the sheer number of calculations it takes to do it," he said.
The calculation is well-suited to the type of programming made possible by a supercomputer, however. The power of light in one spot on the beam can be calculated independently from an adjacent spot in the beam. Such parallel methods of data analysis fit perfectly with the nature of Tulsa's supercomputer, which is essentially several powerful computers linked together to allow for many problems independent of each other to be worked on at the same time.
Refai said the importance of the supercomputer goes beyond individual research projects.
"Faster access to the data would give us an edge over others, but then, at the same time, it might open up the possibility for us to collaborate with industry, too," Refai said, describing how students can benefit as well from learning the programming techniques used with supercomputers.
Send all comments and feedback regarding City to email@example.com
Share this article: