Tajana Simunic Rosing
Computer Science Department
University of California, San Diego
Energy consumption has become a major problem for data centers. In this talk I give an overview of the approaches we have developed at UCSD to significantly improve the energy efficiency in such large scale computing systems by increasing the amount of computation done along with decreasing the amount of power needed to complete it. We integrated SLA management capabilities along with symbiotic workload scheduling into Xen VM scheduler to achieve measured 2x or higher increases in energy efficiency while meeting the SLA requirements. This is done by running a mix of SLA sensitive workloads and batch jobs per server in our fully instrumented data center container. To further decrease the energy costs of running these workload mixes, we developed novel power, thermal and cooling management techniques. The clustering of jobs into fewer servers also may lead to spikes in power consumption. We propose that distributed batteries be used in order to shave these peak power levels. Our distributed per-server UPS architectures stores energy during low activity periods and uses this energy during power spikes. This allows the installation of 24% more servers within the same power budget leading to improved energy efficiency. Lastly, we evaluate and design novel techniques to manage energy for globally distributed data centers by utilizing green energy prediction along with green energy aware wide area routing. We compare our design with state-of-the-art immediate green energy usage which supplements brown energy at the cost of terminating or rescheduling jobs whenever green energy availability is low. Prediction yields in 5x fewer job terminations and improves green energy usage efficiency by 2x in average compared to the instantaneous use.
Tajana Simunic Rosing is currently an Associate Professor in Computer Science Department at UCSD, Director of System Energy Efficiency Lab and leader of Multiscale Systems Center Thrust on Energy Efficient Large Scale Systems. Her research interests are energy efficient computing, embedded and wireless systems. Tajana’s work on event driven dynamic power management laid the mathematical foundations for the engineering problem, devised a globally optimal solution and more importantly defined the framework for future researchers to approach these kinds of problems in embedded system design. Her recent results demonstrate the importance of joint power and thermal management in multicore server systems in order to minimize the overall energy cost. Furthermore, she developed a novel class of proactive thermal management policies that can lower the incidence of hot spots in multicore processors with no performance impact. Her current work is focused on developing energy efficient scheduling policies for virtualized server environments and on energy efficiency in population area healthcare networks. From 1998 until 2005 she was a full time research scientist at HP Labs while also leading research efforts at Stanford University. She finished her PhD in EE in 2001 at Stanford, concurrently with finishing her Masters in Engineering Management. Her PhD topic was dynamic management of power consumption. Prior to pursuing the PhD, she worked as a senior design engineer at Altera Corporation. She obtained the MS in EE from University of Arizona. She has served at a number of Technical Paper Committees, and is currently an Associate Editor of IEEE Transactions on Mobile Computing. In the past she has been an Associate Editor of IEEE Transactions on Circuits and Systems.