When we use computers, we expect them to react in real time and not make us wait for them to perform what we have asked them to do. Now that more computing is going to the cloud, computers are using shared processors and common platforms, creating new challenges in the response times.
Addressing such challenges is particularly critical for real-time systems, or computing systems that require real-time performance guarantees, such as active safety features in automobiles designed to prevent cars from accidents.
Chenyang Lu, the Fullgraf Professor, and Christopher Gill, both professors of computer science in the Washington University in St. Louis School of Engineering & Applied Science, have been working address these challenges for the past several years. To continue this work, they have received a three-year, $610,330 grant from the Office of Naval Research to study dynamic real-time virtualization and cloud computing.
Virtualization allows many independently developed systems to be used as virtual machines sharing a common, virtualized computing platform without reconfiguring operating systems or applications. They allow for larger uses, such as analyzing "big data," but do not offer guarantees of real-time performance, so applications may be delayed.
In this work, Lu and Gill plan to develop the platform for cloud computing environments to manage and coordinate real-time virtual machines, which allows for consolidation and flexibility of allocating resources. The platform would give other researchers and users a new way to evaluate their research results. They also plan to strengthen the capabilities of the platform by applying it to cloud computing through dealing with realistic situations and establishing expectations for the systems' behavior.
"Today's cloud systems are slow and unpredictable," Lu says. "If you send a job to a cloud computing platform, sometimes it comes back quickly, and sometimes it comes back slowly because it's sharing the machine you're submitting it to with other virtual machines. This becomes a problem when you want a real-time response predictably, such as using the Internet of Things to control physical environments and systems. Anytime you want to control something, from automobiles to manufacturing plants to smart cities, it needs to be a more predictable, real-time performance."
For example, he said, cities that use traffic sensors to optimize traffic via the cloud need a real-time response. If the system cannot keep up, traffic lights get off schedule and problems ensue.
In 2013, Lu and Gill developed an open-source, real-time scheduling software called RT-Xen, for a popular virtualization software, known as Xen, used by cloud computing companies.
The School of Engineering & Applied Science at Washington University in St. Louis focuses intellectual efforts through a new convergence paradigm and builds on strengths, particularly as applied to medicine and health, energy and environment, entrepreneurship and security. With 88 tenured/tenure-track and 40 additional full-time faculty, 1,300 undergraduate students, more than 900 graduate students and more than 23,000 alumni, we are working to leverage our partnerships with academic and industry partners — across disciplines and across the world — to contribute to solving the greatest global challenges of the 21st century.