��������

Cloud computing has revolutionized how we interact with our electronic devices — but as an increasing number of devices demand more and more bandwidth, its limitations are becoming more apparent, too.

As the Internet of Things (IoT) transmits a growing amount of data over every network without requiring human-to-human or human-to-computer interaction, using the cloud for storage and boosted processing power can lead to long communication latencies, inadequate mobile service, difficulties when scaling up and the inability to provide large-scale personalized services.

Yifan Zhang — an assistant professor in the Department of Computer Science at ��������’s Thomas J. Watson School of Engineering and Applied Science — is seeking solutions to these problems, and he recently received a to fund his research. He is the third computer science faculty member to receive the honor this year.

Zhang’s winning proposal, “Enabling Edge-hosted Private Services via Unikernel-based Lightweight Virtualization,” may sound dense to non-computer scientists. The two basic premises are easy to grasp, though.

Edge computing puts processing power at the edge of a local network, which is closer than asking the cloud to shoulder the bulk of what users demand from their technology.

“The computation delay between the user device and the edge server is a lot shorter than between the user device and the cloud server,” Zhang said. “It also helps to enable a bunch of new services that I call edge-based private services.”

Unikernel-based programming breaks down the functions of traditional operating systems into smaller pieces, allowing a device to choose which functions it needs from moment to moment and therefore run more efficiently.

“Unikernels have a very tiny OS image and very small runtime overheads,” Zhang said. “If you compare it to a generic OS, you have memory, you have the network, you have the CPU and all of the needed resources.

“For unikernels, maybe you don’t need a file system or drivers. If you include those subsystems, that would make the OS image large and the runtimes slow. Unikernels compile just the applications you want to run with the necessary functionalities.”

With the NSF CAREER funding, Zhang plans to “address the challenges in a step-by-step manner,” building on research he has already done during his six years at ��������. If successful, he sees many benefits for the future of computers and mobile devices.

“As far as users are concerned, you can have more applications and more innovative services that are not currently available because of resource constraints from using the cloud,” he said. “Also, the communication latency between users and the cloud is long, while the edge is very short. You can get your result quickly.”

This year marked Zhang’s third time applying for a CAREER Award, which supports early-career faculty who have the potential to serve as academic role models in research and education.

“Every time I’ve submitted, I’ve received very useful feedback from the reviewers,” he said. Then I’ve improved the proposal and brought in some new preliminary research results. I was quite confident this time that I had a chance. It’s a huge relief for me, and also very exciting.”