Newsletter

4/14/09 - The Spring 2009 (PDF) edition of the CAC newsletter is available.

Newsletter Archive

 

The University of Florida , the University of Arizona and Rutgers, the State University of New Jersey , have established a national research center for autonomic computing (CAC).

This center is funded by the Industry/University Cooperative Research Center program of the National Science Foundation, CAC members from industry and government, and university matching funds.

Speaker Bios and Abstracts - Greencomputing Workshop 2009

Distributed Energy Adaptive Computing

Dr. Krishna Kant, Intel Research

Abstract: Fueled by burgeoning online services, power and thermal issues are becoming a substantial issue in terms of cost and environmental impact both on the server (or data center) side and client side. In this talk, we shall motivate an approach that puts power/thermal issues at the heart of distributed computing, and strives to dynamically optimize energy use, heat dissipation and energy supply in order to deliver an acceptable user experience. The talk shall discuss how such an approach can enhance the sustainability of computing and will lay out the challenges in realizing the vision.

Speaker Bio: Dr. Krishna Kant has been with Intel Corp since 1997 where he has worked in a variety of research areas including traffic characterization, security/robustness in the Internet, data center networking, utility computing, and power control of computer systems. He is currently on a visiting appointment with the National Science Foundation. From 1991 to 1997, he was with Telcordia Technologies (formerly Bellcore) and worked on SS7 signaling and congestion control. Prior to this, he was an Associate Professor of Computer Science at Penn State University. He is the author of the graduate text book .Introduction to Computer System Performance Modeling., McGraw Hill 1992. He received his Ph.D. degree in Computer Science from University of Texas at Dallas in 1981.

Power/Performance Tradeoffs and Resource Management Issues in High-Performance and Embedded Systems

Margaret Martonosi, Princeton University

Abstract: Chip multiprocessors and Systems-on-Chip have emerged as dominant research focuses in both the general purpose computer architecture and embedded systems research communities. Some of the most prominent challenges that exist for parallel systems have to do with balancing resources across the many concurrent applications or users. CMP resource management problems span a wide range, include managing the per-application cache capacity and bandwidth usage, apportioning power/thermal envelopes between application, and others. This talk will describe my research group's recent work on CMP/SoC resource management. I will close by offering a vision for how hardware-software interfaces should be adjusted to afford better leverage on resource management problems, thereby offering energy-efficient computing at high-performance for a range of scales and domains.

Speaker Bio: Margaret Martonosi is currently Professor of Electrical Engineering at Princeton University, where she has been on the faculty since 1994. She also holds an affiliated faculty appointment in Princeton CS. Martonosi's research interests are in computer architecture and the hardware/software interface, with particular focus on power-efficient systems and mobile computing. In the field of processor architecture, she has done extensive work on power modeling and management and on memory hierarchy performance and energy. This has included the development of the Wattch power modeling tool, the first architecture level power modeling infrastructure for superscalar processors. Her memory hierarchy work has included early performance-oriented studies, as well as more recent work on energy-aware memory hierarchies. In the field of mobile computing and sensor networks, Martonosi lad the Princeton ZebraNet project, which included two real-world deployments of tracking collars on Zebras in Central Kenya. She is now co-leader of the Sarana project, which is building software interfaces for collaborative computing among mobile devices. Martonosi is co-author on over 100 refereed publications and inventor on five granted US patents. Martonosi completed her Ph.D. at Stanford University, and also holds a Master's degree from Stanford and a bachelor's degree from Cornell University, all in Electrical Engineering.

Statistical Multiplexing Techniques for Improved Power Provisioning in Data Centers

Bhuvan Urgaonkar, Penn State University

Abstract: Current capacity planning techniques based on heavy over-provisioning of power infrastructure hurt (i) the operational costs of data centers as well as (ii) the computational work they can support. We explore a combination of statistical multiplexing techniques to improve the utilization of the power hierarchy within a data center. At the highest level of the power hierarchy, we employ controlled under-provisioning and over-booking of power needs of hosted workloads. At the lower levels, we introduce the novel notion of soft fuses to flexibly distribute provisioned power among hosted workloads based on their needs. Our techniques are built upon a measurement-driven profiling and prediction framework to characterize key statistical properties of the power needs of hosted workloads and their aggregates. We characterize the gains in terms of the amount of computational work (CPU cycles) per provisioned unit of power---Computation per Provisioned Watt (CPW). Our technique is able to double the CPW offered by a Power Distribution Unit (PDU) running the e-commerce benchmark TPC-W compared to conventional provisioning practices. Over-booking the PDU by 10% based on tails of power profiles yields a further improvement of 20%. Reactive techniques implemented on our Xen VMM-based servers dynamically modulate CPU DVFS states to ensure power draw below the limits posed by soft fuses. Finally, information captured in our profiles also provide ways of controlling application performance degradation despite overbooking.

Speaker Bio: Bhuvan Urgaonkar is currently an assistant professor of computer science and engineering at the Pennsylvania State University. He received his B.Tech (honors) in computer science and engineering at the Indian Institute of Technology Kharagpur in 1999, M.S. in computer science at the University of Massachusetts in 2002, and his Ph.D. in computer science at the University of Massachusetts in 2005. His current research interests are in the areas of operating system virtualization, enterprise storage systems, power-efficient computing platforms, and workload characterization and generation for enterprise applications. He has co-authored papers that won the best student paper awards at the Sixteenth Annual Meeting of the IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS 2008) and the International Conference on Autonomic Computing (ICAC) 2005. Urgaonkar is a member of the ACM, IEEE, and USENIX.

Modeling Tradeoffs between Costs, Reliability and Air Emissions for Computer Data Centers

David W. Coit, Frank A. Felder, Hatice Tekiner, Rutgers University

Abstract: Computer data centers account for about 5 percent of the U.S. electricity consumption and are the fastest growing users of electricity. Data centers must be designed and operated to achieve cost, reliability and, now with growing concerns with global climate change, reduction of greenhouse gases. This presentation uses a comprehensive and detailed model of a natural gas supply system and electric power system from generation, transmission to distribution to analyze the tradeoffs between costs, reliability and air emissions. It presents the result of adding data center electric loads and, under different scenarios, computes the relative tradeoffs along these three dimensions. It also allows the determination of a data center.s carbon footprint related to its electricity consumption. By having a comprehensive and integrative model, it enables the evaluation of data center needs, not just within the fence of the center, but system wide, opening up additional ways of improving the performance of data centers.

Speaker Bio: David Coit is an Associate Professor in the Department of Industrial & Systems Engineering, Rutgers University. He conducts research in systems reliability optimization, energy systems reliability and multiple objective optimization. His research has been funded by the National Science Foundation (NSF), power utilities and industry. He previously worked for IIT Research Institute (IITRI) for over 10 years, and he has degrees from University of Pittsburgh, Cornell University, and Rensselaer Polytechnic Institute.

Speaker Bio: Frank Felder is the Director of the Center for Energy, Economic and Environmental Policy at the Blousttein School of Planning and Public Policy, Rutgers University. His is also a member of the Bloustein School faculty. He conducts applied energy research for the State of New Jersey and on topics related to the economics, reliability and environmental implications of the electric power industry.

Network Optimization for Efficient Energy Usage

Lisa Zhang, Bell Labs

Abstract: In this talk we give a brief survey on how energy consumption depends on processing speed for a variety of network devices. Current approaches for energy saving includes speed scaling and power down. The former adjusts processor speed to reflect the traffic load and the latter toggles the processor between on and off modes for energy savings. We present a number of routing and scheduling problems motivated by these approaches and the energy data. We also discuss initial results for some of these problems. This work is joint with Matthew Andrews and Antonio Fernandez.

Speaker Bio: Lisa Zhang is a member of technical staff in the Algorithms Research Group at Bell Labs in Murray Hill, NJ. She received her B.A., summa cum laude, in mathematics from Wellesley College in 1993 and her Ph.D. in Theory of Computing from the Massachusetts Institute of Technology in 1997. Her research area is algorithm design and analysis. Her research broadly concerns algorithmic and complexity issues of networking, with a focus on design and optimization, routing and scheduling protocols, and stability and Quality-of-Service analyses. She twice won the Bell Labs President's Gold Award and the Lucent Chairman's Award.

Green Radio The case for more efficient cellular base-stations

Peter M Grant, Regius Professor of Engineering, the University of Edinburgh

Abstract: This presentation will discuss the power drain or efficiency of mobile terminals and base-stations to define the issues with current cellular systems operation, particularly in base-station or access point designs. Indications will be given for the total power consumption of UK cellular GSM and 3G networks. These issues are set to increase with the move from predominantly speech and text messaging to the roll out of mobile broadband, with its much higher data rate transmission requirements and the urgent requirement for lower cost per delivered bit. The presentation will conclude by indicating the typical approaches that are about to be researched to alleviate the high power drain to achieve green radio system design credentials with significantly reduced CO_2 emissions. This presentation is set in the context of the Mobile VCE, an industry-UK government funded research programme conducted in 4 UK Universities in collaboration with 11 international industrial sponsor and programme monitoring companies.

Speaker Bio: Prof Peter Grant has been on staff at Edinburgh since 1971 after working in the UK for Plessey and Hughes for 5 years. He was appointed as the first head to form and integrate the School of Engineering at University of Edinburgh, leading it from 2002 - 2008. Before that he served as Head of Electronics from 1999 - 2002.

Peter Grant has three "doctorates", a PhD from the University of Edinburgh in 1975, an honorary DEng (Doctor of Engineering) from the Heriot-Watt University in Edinburgh in 2006 and another honorary DEng from Napier University, Edinburgh in 2007.

He holds five Fellowships from: IEEE, IEE/IET, Royal Academy of Engineering, Royal Society of Edinburgh and he was elected one of the first four fellows of the European Association for Speech, Signal and Image Processing (EURASIP), having previously served there as President 2000-2002. He was also awarded in 2004 the 82nd IEE Faraday Medal.

Professor Grant was in 2007 appointed to be the 8th Regius Professor of Engineering at The University of Edinburgh. "Regius" i.e. regal chair appointments are conferred by the Queen of Great Britain (as a title only without any financial support)!