Control of Virtualized GPU Resource in Cloud Gaming

Jianguo Yao, Shanghai Jiao Tong University


As the virtualization technology for GPUs matures, cloud gaming has become an emerging application among cloud services. In addition to the poor default mechanisms of GPU resource sharing, the performance of cloud games is inevitably undermined by various runtime uncertainties such as rendering complex game scenarios. The question of how to handle the runtime uncertainties for GPU resource sharing remains unanswered. Here I describe a technique that controls virtualized GPU resource in cloud gaming. This technique interposes scheduling algorithms in the graphics API of the operating system, and hence the host graphic driver or the guest operating system remains unmodified. In order to guarantee the Service Level Agreement as well as maximizing GPU usage, it features feedback control that mitigates the impact of the runtime uncertainties on the system performance.

Presentation slides

Biography:  Dr. Jianguo Yao obtained his Ph.D. Degree from the Northwestern Polytechnical University in 2010, and was a joint-education Ph.D student at McGill University from 2007 to 2008. He was also a joint Postdoctoral Fellow at Ecole Polyechnique de Montreal and at McGill University from 2011 to 2012. In 2011, he worked briefly as an intern researcher at Bombardier Inc in Canada. Dr. Yao received the prestigious PBEEE/Quebec Merit Scholarship for Foreign Students from Quebec Fund for Research on Nature and Technology (FQRNT), Chinese State Scholarship to Study Abroad Research award from China Scholarship Council.
Currently, he is an Assistant Professor at the Shanghai Jiao Tong University. His research interests are cyber-physical systems, feedback control applications in clouds and power management in data centers. He has published more than 30 research papers in major peer-reviewed International journals and top-conference proceedings, such as IEEE Transactions on Parallel and Distributed Systems, IEEE Transaction on Service Computing, IEEE Transaction on Industrial Informatics, IEEE Transaction on Smart Grid, ACM Transactions on Architecture and Code Optimization, ACM Transactions on Embedded Computing Systems, ACM Transactions on Sensor Networks, HPDC, ICDCS, RTSS, INFOCOM, ACC etc. He is very active with the Controls and Clouds research community. He has participated in various conferences, and served as TPC members in the conferences including: INFOCOM 2014’2015, SIES 2013, RTCSA 2012’2013 and ICPADS 2012 etc.