When it comes to high-powered computers, increased energy consumption and lack of reliability continue to be problems researchers are trying to overcome. Unfortunately, some of the best current methods for achieving these goals have their flaws.
One example is seen in server consolidation, during which resources are shared across multiple applications and virtual CPUs. Without careful optimization, consolidation may result in sub-optimal energy efficiency or in high temperatures that increase cooling costs.
Coskun and her team’s initial work focused on designing server consolidation techniques that run on virtualized servers, targeting high-performance computing applications. They now have a working system that is able to improve energy efficiency on a real-life server.
“Working with VMware has proven to be a rewarding partnership because they’re always offering valuable feedback,” said Coskun. “They work with customers so they’re aware of problems we might not be able to observe in the lab setting.”
Since 2011, she and her team have published several research papers on their progress in improving energy efficiency through server consolidation.
“We have met the early milestones that we set with VMware,” said Coskun. “Now, we’re looking at more broad, realistic data center scenarios and exploring the reliability aspects in addition to improving the energy efficiency.”
Pleased with their initial results, VMware recently renewed Coskun’s funding, providing $75,000 for another year of research.
“Based upon promising early results, we are excited to extend our collaboration with Professor Coskun to investigate techniques for energy-efficient use of multi-core processors with virtualized workloads,” said Steve Muir, director of the VMware Academic Program (VMAP).
Part of Coskun and her team’s continued research will include testing their techniques on multiple server nodes that are connected over the network as a representative data center scenario.
“Professor Coskun plans to build on this base by exploring clustered environments and extending experiments to include general enterprise workloads is particularly interesting; we look forward to our continued relationship over the next year,” said Muir.
The renewed funding will allow Coskun to hire additional PhD students or postdoctoral researchers later this summer. She will also continue supporting student researchers, Can Hankendi (PhD ’15) and Samuel Howes (ECE ’14). Hankendi played a major role in developing a technique for automated resource management while Howes, newer to the project, focuses on setting up the virtualized experimental infrastructure for data center benchmarks.
“Can has single-handedly done most of the implementations and has put together great work so far,” said Coskun, “and Sam is rapidly learning about virtualization, which is a hot research topic right now.”
Truly a team effort, students not only help with the research; they attend company meetings and symposiums with VMware as well.
“VMware is very research focused so we have a lot of exposure to their own researchers and engineers,” said Coskun. “It’s been a beneficial collaboration on both sides, so we hope to keep this partnership going in the future.”
-Rachel Harrington (firstname.lastname@example.org)
May 19, 2011 -”VMware Funds Coskun’s Server Consolidation Research“
Improving how computer chips communicate with each other could potentially result in smaller, faster and more power efficient devices.
To reach that point, some researchers are studying on-chip optical communications, which rely on the ability to transmit light at high frequencies using light emitters and lasers. Unfortunately, the primary materials used in on-chip lasers tend to be expensive and difficult to integrate with silicon-based microelectronics circuitry.
Many electrical engineers believe that if light were more efficiently generated using a silicon-compatible material, costs of bandwidth could decline and optoelectronic devices could be massively produced at low cost. The problem, however, is that standard silicon materials do not efficiently generate light.
Boston University Associate Professor Luca Dal Negro (ECE), who has conducted much of his research on optical physics and semiconductor nanostructures, recently received a grant that could make silicon-based light sources a real possibility.
“What we’re hoping to do is create novel light sources that leverage distinctive optical resonances, known as gap plasmons,” said Dal Negro. “This would strongly confine electromagnetic fields at the nanoscale on a less expensive silicon platform.”
The Air Force Office of Scientific Research recently awarded Dal Negro $379,989 for his project, “Nanoscale Optical Emitters for High Density Information Processing Using Photonic-Plasmonic Coupling in Coaxial Nanopillars.”
According to his proposal, Dal Negro plans to design and engineer nanoscale-size optical cavities using silicon-compatible electronic materials to demonstrate “dramatically enhanced radiation rates, drastically reduced [size], and room temperature efficient light emitting arrays and laser structures.”
Dal Negro will receive the funding over a two-year period that began in early January.
-Rachel Harrington (email@example.com)
A growing amount of confidential or private information is shared using computers or phones that aren’t always secure, despite researchers’ best efforts.
Part of the reason for this, Professor Alexander Sergienko (ECE) said, is because current telecommunication security relies on complex coding schemes.
“It’s a very difficult and time-consuming task, but all of these codes could potentially be cracked using sufficient computer power,” said Sergienko. “Many previously secure messages have been accessed recently thanks to the development of more efficient and powerful computers.”
Sergienko and his research team at Boston University think quantum mechanics – and more specifically quantum cryptography – may be the solution.
When using quantum mechanics, the security of legitimate users is protected by not allowing a rogue party to make copies or clones of passwords or messages as they travel in the telecommunication fiber or free space.
“Quantum cryptography has the potential to code messages in a way that is unbreakable and secure forever,” said Sergienko.
To support this research, the Defense Advanced Research Projects Agency (DARPA) has awarded Boston University $1.3 million through the new program, “Quiness: Macroscopic Quantum Communication.” The funding is part of a larger $4 million grant that will be shared with researchers at the University of Maryland, Baltimore, and University of Rochester, who will work with BU to develop secure quantum cryptographic communication technology.
Sergienko will work closely with researchers such as Gregg Jaeger (CGS), an associate professor of natural sciences and mathematics, and University of Maryland’s James Franson, a professor of physics. Jaeger is the author of several related books including Quantum Information and Philosophy of Quantum Information and Entanglement. He also teaches quantum information and quantum theory at BU.
“Quantum cryptography has its roots in and is intertwined with the study of the foundations of quantum theory and quantum entanglement,” said Jaeger, who added that entanglement typically occurs when elementary particles interact then separate. “Entanglement is a property characteristic of the quantum world and is of tremendous interest to physicists and philosophers, as well as engineers working with quantum information.”
At this time, one of the drawbacks of quantum cryptography, which is sometimes known as quantum key distribution, is that it runs at a significantly slower speed than that of a regular telecommunication signal and cannot travel as far.
“The main task of the DARPA Quiness program is to develop new engineering solutions that would bridge this gap and allow quantum cryptography to run at faster speeds over longer distances,” said Sergienko.
Improving secure data exchanges has been a main focus of government and private agencies over the last decade, including the United States Department of Defense.
Sergienko, who has studied quantum optics for more than 20 years, has played an active role in searching for telecommunication security solutions. Working with BBN Technology, Inc. and Harvard University in 2004-06, he helped establish an operational Boston DARPA quantum network, a culmination of previous work.
He and his research team will continue to make advances in the field with this latest project.
-Rachel Harrington (firstname.lastname@example.org)
Holyoke center does environmentally friendly computing
So many machines, so few people.
Ultimately, 20,000 computers will fill an almost acre-sized floor of the Massachusetts Green High Performance Computing Center (MGHPCC). Yet just 10 or so workers have day jobs at the center, which opened earlier this month after BU, four other Bay State research universities, state taxpayers, and business partners poured almost $95 million and two years into its construction.
The hardware-to-human ratio highlights the nature of business at the 90,300-square-foot building approximately two hours west of Boston, in Holyoke, Mass., one of the largest supercomputing data centers east of the Mississippi River. Its computers will be accessed remotely for the most part by researchers at BU and its partners, working from their campuses. But unlike The Terminator’s vision of a dystopia where machines control mankind, the center exists to harness machines to human ends—specifically, to accommodate the five universities’ mushrooming computing needs.
“Some equipment will be general-purpose clusters available to all researchers at BU,” says Azer Bestavros, a College of Arts & Sciences computer science professor, director of the Rafik B. Hariri Institute for Computing and Computational Science & Engineering, and cochairman of the MGHPCC education and outreach committee. “Other equipment will be more special-purpose, and thus dedicated to specific research.” For example, he and seven BU colleagues have received a National Science Foundation grant to purchase, in tandem with several other universities, a computing cluster to be housed at Holyoke and linked to each member school. Research like this propelled BU’s admission this month to membership in the Association of American Universities, only the fourth school since 2000 invited into the elite research consortium.
In the next few months, the first computers will be installed, making the trip from the center’s loading dock up its six-ton-capacity freight elevator and into that sprawling computing room. Some will be relocated from the five campuses, freeing up space on each. (At BU, that will include space at 881 Commonwealth Ave., 111 Cummington Mall, the Physics Research Building at 3 Cummington Mall, and various schools’ data centers.) Others will be replaced altogether.
Center executive director John Goodhue guided a recent media tour around the austere corridors and cavernous rooms with their factory-like floors, explaining things as adroitly as a seasoned professor, although he’s actually a refugee from Cisco and various computer companies. The electric room has “the equivalent of the fuse box in the basement of your house,” he said, except that the fuses weigh 300 pounds each. The center will make 10 megawatts of energy a day available for computing, which is enough to run up to 10,000 homes.
The MGHPCC was designed as an environmentally benign computing colossus. Outside, most of the building is a huge rectangle, fronted by towering windows, in a neighborhood of old brick factories. A canal off the Connecticut River meanders by the property, hinting at one of its green aspects: more than 70 percent of the local electricity supply comes from renewable hydroelectric power. The building also meets green building standards, courtesy of its efficient power system and conservation measures.
For example, the electric room’s equipment can function under high temperatures, obviating the need for power-gulping air conditioning. The building’s electric system was installed directly under the computing floor, with shorter wires that minimize wasteful loss of electricity. The average data center requires nine megawatts of power for cooling, Goodhue said; MGHPCC will require less than three, partly by tapping the outside air in winter to cool the computers.
Half the computing floor is lined with black phone booth–sized cabinets, each built to hold about 30 computers. The other half is wide open for expansion and more computers, as is the entire nine-acre property, which has space for additional buildings as needed.
The center devotes an entire corridor to telecommunications equipment—“Computing doesn’t really make any difference unless you can tell people the answer,” noted Goodhue—including a dedicated link to every university in the consortium and thousands of sensors monitoring things from energy efficiency to smoke in the building. Dubbing this corridor “the building’s mouth and ears,” Goodhue said that the “goal is to make all the users of the machines feel like this building is just another building on their campus. The way we do that is by making sure we have very fast, very reliable communication to all of the users.”
BU’s academic partners in the center are Harvard, MIT, Northeastern, and the University of Massachusetts. The state contributed $25 million to the construction, reflecting “what is possible when government, academia, and business work together,” Governor Deval Patrick said. The center “will serve as an economic development model for the state and the nation for generations to come.” Additional funding came from EMC Corporation and Cisco Systems.
This article first appeared on November 27, 2012, in BU Today.
Synthetic biology is still a relatively new research field, but we already know that in the future, it might be used to complete tasks too risky for the average human such as explosives detection or oil cleanup.
To make these plans a reality, synthetic biologists typically introduce specific DNA into living organisms like bacteria in order to instruct them to carry out certain jobs. Often, doing so can be a trial-and-error process.
To help create a more systematic approach, Assistant Professor Douglas Densmore (ECE) and his research group designed Eugene, a human and machine readable specification language that can be used to formalize and improve the efficiency.
“Imagine knowing English but not knowing how to put the words together,” said Densmore. “We’re essentially creating rules to make the sentences only instead, we’re working with DNA.”
In 2011, Agilent Technologies awarded Densmore over $50,000 for this research. In addition to funding, Agilent also connected Densmore to a mentor in their company, Allan Kuchinsky, a principal project scientist in the Molecular Tools group. To continue their partnership, Agilent recently awarded Densmore another $40,000 to further his work.
Much of this funding has and will help support Ernst Oberortner, a postdoctoral associate in the Department of Electrical & Computer Engineering who has assisted Densmore in his research over the last year.
“When Ernst came, Eugene existed, but he took it back to the basics in order to get better results,” said Densmore.
As a Ph.D. student at the Vienna University of Technology, Ernst Oberortner had already completed a lot of work on domain-specific languages (DSL) and programming languages used in software development.
“During my first year on the project, I was learning synthetic biology and extending Eugene’s functionalities,” said Oberortner. “Now I’d like to augment the Eugene language to move it to the next level.”
Densmore chose to work with Oberortner because of his extensive skills in computer science.
“He’s a good example of how computer science and biology can come together,” said Densmore. “Ernst didn’t have a background in synthetic biology but he was able to apply his experience and move our research forward.”
New funding will be awarded at the start of 2013.
-Rachel Harrington (email@example.com)
June 7, 2011 – “Agilent Funds Synthetic Biology Research by Densmore”
Speaking up about research interests during freshman year might make students a little nervous, but doing so could also take them deep into a Puerto Rican rainforest and provide additional opportunities throughout their undergraduate career.
This was and is the case for Kangping Hu (ECE ’13), who presented his Undergraduate Research Opportunities Program (UROP) poster paper, “Mode Conversion of NAU Launched Whistler Wave Over Arecibo, Puerto Rico,” in October at the George Sherman Union (GSU). His research took first place out of 219 participants at the 15th Annual Undergraduate Research Symposium.
Hu’s work explores the effects of wave-particle interactions in the ionosphere. The ionosphere has become increasingly important as the use of wireless communications rises, because certain ionospheres’ density irregularities can result in signal distortion.
By studying this, Hu says he hopes to “investigate the spectral broadening effects in the ionosphere for communications and remote sensing purposes.”
Hu took advantage of having Professor Min-Chang Lee (ECE) as his advisor soon after he came to BU. Grants Lee received even funded some of Hu’s research with him over the last three summers.
“I have been tutoring Kangping in his research over the past three years, beginning his freshman year in 2009, with my other undergraduate and graduate students,” said Lee. “He is both intelligent and a hard worker.”
Over his last winter and summer breaks, Hu flew to Puerto Rico to conduct research in the rainforest at the Arecibo Observatory, home of the world’s largest radio telescope.
“We did the research in the observatory using a dish that was in the heart of the jungle,” said Hu. “We spent a week there setting up experiments in the middle of the night from 8 p.m. ’til 5 a.m. in order to remove the sun as a variable.”
His work didn’t end after the information was collected. Last summer, Hu analyzed the data during the UROP summer research session. “With this analysis, we made more developments for the presentation,” he explained.
He is currently applying his accumulated research to write his Senior Honors thesis. After graduation, Hu plans to continue his academic career and get a graduate degree in electrical engineering.
-Sneha Dasgupta (COM ’13)
In today’s world, it’s easier to find a college student carrying a laptop, smartphone, or iPod than to spot one without.
As people become increasingly dependent on electronic devices, more personal information, including credit card and bank account passwords, go on these handheld machines as well.
Security is now of great concern since many of these digital pieces require only a four digit passcode to sign in – but what if these tablets could recognize the swipe of your finger?
That’s the idea behind research by Boston University’s Professor Janusz Konrad (ECE) and Associate Professor Prakash Ishwar (ECE), who are collaborating with Professor Nasir Memon and Associate Professor Katherine Isbister, both from the Polytechnic Institute of New York.
“We hope the swipes can be unique enough, by linking several of them to form a word for example, to verify user identity,” said Konrad.
To support the project, the National Science Foundation recently awarded the teams nearly $800,000, with $400,000 of that going to Konrad and Ishwar.
Konrad, Ishwar, and their research team will focus their side of the project on examining how hand and body gestures can be used for authentication.
“For example, hand or body movements could be used to open a door instead of a card swipe if the camera recognized an individual’s gestures,” said Konrad.
Ishwar added that swipes and hand gestures are revocable, meaning that you can update the gesture you use to authenticate yourself. This is different from fingerprints and iris scans which, if compromised, cannot be replaced.
“There are endless choices for personalizing gestures that are easy to remember and reproduce, reliable, and pleasant to work with,” said Ishwar.
Though Ishwar and Konrad have long been exploring the possibility of using cameras in action recognition, they didn’t think their results would be good enough for authentication until they met Memon when he visited BU as an Institute of Electrical and Electronics Engineers (IEEE) Distinguished Lecturer.
“He had already worked on the swipes idea, but when he saw our gesture results, it all clicked together,” said Konrad.
Initial tests using a Microsoft Kinect camera have shown promise, Konrad added, and he and Ishwar are already looking toward other places where their technology might be used.
“One of the challenges will be extending this to regular surveillance cameras often deployed at entrances to buildings,” said Konrad.
If successful, the project has the chance not only to reduce breach-related costs but also increase our sense of security.
-Rachel Harrington (firstname.lastname@example.org)
When we think about light, what often comes to mind are the beams that keep the world from being a dark and dreary place or the electricity and bulbs that keep our homes bright.
In all these cases, light’s energy is coming from the center of its beams, but sometimes light can possess orbital angular momentum. When this is the case, those beams, called optical vortices, allow no energy in the center and essentially look like donuts. The energy, rather than moving in a straight line along the beam path, spins around the donut.
In May 2011, a team headed by Associate Professor Siddharth Ramachandran (ECE) that includes Dr. Steve Golowich of MIT Lincoln Laboratory and Dr. Poul Kristensen of OFS-Fitel received a grant from the Defense Advanced Research Project Agency (DARPA) to study optical vortices and whether optical fibers can be used to generate and transmit them.
Previously it was thought that optical vortices, while exotic and interesting, have little use because of their instability in fibers, but recent work by Ramachandran and Kristensen suggests that novel optical fiber structures can help transmit these beams, which opens up the prospect of using them for telecommunications.
Pleased with their initial results, DARPA has awarded a Phase II grant worth approximately $800K to Ramachandran, Golowich, and Kristensen to continue this project.
“We are delighted to have received this highly competitive award, for it provides the necessary resources to investigate much more complex optical fiber structures which we expect to use to generate very high order orbital angular momentum states,” said Ramachandran. “This has never been done, but if the investigations are successful, we expect the results to help increase the data carrying capacity of optical fibers significantly.”
Added Ramachandran, learning more about optical vortices could lead to advances in fields outside telecommunications, such as DNA sorting, high-resolution imaging, and nonlinear and quantum optics.
Already, Ramachandran and his team have published several reports showing the use of these beams to perform quantum encryption for secure networks and for studying spin-orbit interactions in quantum systems.
At the European Conference on Optical Communication in September, the team demonstrated a significant breakthrough that led to a four-fold data capacity enhancement over a single wavelength channel with fibers that use these exotic beams. With their new funding, the researchers hope to continue making progress on this work.
If you’d like to hear more about Ramachandran’s work on optical vortices, he will be giving a talk on the subject at the Future of Light Symposium on November 29. Learn more.
-Rachel Harrington (email@example.com)
Today, people with chronic conditions like diabetes are benefitting from real-time monitoring devices like miniaturized implants, home monitoring equipment, and smartphone applications. Unfortunately, even though tracking a person’s symptoms and vitals has improved, hospitals and their medical teams are not ready to benefit from possessing so much personalized health data.
Boston University’s Professor Ioannis Paschalidis (ECE/SE) and Dr. William Adams (BMC) have teamed up with MIT’s Professor Dimitris Bertsimas to develop algorithms that can systematically process all patient data in hospital electronic medical records and personalized health records. These algorithms will be designed to classify patients based on the risk of developing an acute condition that would require hospitalization. Such information can then be used to drive preventive actions.
“What motivated us to start this particular project is the recognition that the US health care system is extremely inefficient as it is geared toward treating acute conditions,” said Paschalidis. “There are, we believe, tremendous opportunities for preventing the occurrence of these conditions and the expensive hospitalizations they cause.”
To support their work, the National Science Foundation (NSF) has awarded Paschalidis (PI), Adams (Co-PI), and Bertsimas (Co-PI) a five-year, nearly $2 million grant for the project.
By focusing on disease prevention and keeping patients out of the hospital, their work has the potential to improve a healthcare system that is often considered to be very expensive and highly inefficient.
“To that end, the meaningful use of electronic health records is seen as a key to improving efficiency,” the team wrote in their proposal.
The research will utilize Paschalidis’s expertise in data models, optimization and decision theory, but it is truly a collaborative project. Adams, for example, will work with physicians to get feedback on the outcome of the algorithms.
“The main challenge is going to be the adoption of the techniques we develop by physicians in particular and the healthcare system in general,” Paschalidis said. “Dr. Adams will serve as our ambassador to that community.”
Adams said that the Boston Medical Center has spent more than ten years “developing a robust and rich clinical informatics infrastructure for clinical care and research” and welcomes this new partnership.
“Translational science involves collaborative efforts between traditionally independent scientists,” Adams said. “This project is innovative and important in that it brings together mathematicians, engineers, clinicians, and informaticians to better understand and improve healthcare.”
As healthcare costs increase, their research may prove not only to be timely but also life-changing and cost-effective.
-Rachel Harrington (firstname.lastname@example.org)
New life forms reporting to robots reporting to humans may seem like something out of a science fiction movie, but this may be closer to reality than you think.
With recent Office of Naval Research, DARPA, Agilent, and NSF funding, researchers from Boston University, alongside those from Harvard, MIT, UC Berkeley, Northeastern, and the University of Pennsylvania, are working on projects that combine humans, robots, and genetic engineering. The work has the potential to alert humans of harmful bacteria, create “assembly lines” of genetic parts, and create specialized cells to fight a host of diseases.
Boston University’s laboratory, the Cross-disciplinary Integration of Design Automation Research (CIDAR) group, plays a central role in these projects. As members of the lab work on software and experimental “design drivers” to help synthetic biologists work more efficiently, they focus on the following research areas: specification, design, assembly, and community outreach.
Synthetic biology merges technology and biology to solve emerging problems pertaining to areas such as energy, health, food, and the environment. The CIDAR group aims to use computation techniques, largely derived from electrical engineering and computer science, and apply those to experimentally verified synthetic biological systems.
With students in electrical and computer science; bioinformatics; biomedical engineering; and molecular cell biology and biochemistry, CIDAR researchers yield nontraditional results thanks to the cross-disciplinary nature of the group. The work is particularly useful to students interested in electronic design automation, synthetic biology, or professional software development and engineering.
The CIDAR laboratory also brings computationalists and experimentalists together. Computationalists often focus on theoretically interesting problems, which alone does not directly make the engineering of today’s biological systems a reality. Experimentalists, on the other hand, spend time designing ad-hoc, piecewise software that computationalists can develop better by taking a holistic approach. By bringing the two types of research styles together, Assistant Professor Douglas Densmore (ECE), who leads the CIDAR lab, hopes to come up with new solutions to apply toward synthetic biology.
Learn more about CIDAR in this video overview with Densmore.
If you’re interested in joining CIDAR, see open available positions.
-Sneha Dasgupta (COM ’13)
“Life is Suite”