An era of big data demands datacenters, which house the computing infrastructure that translates raw data into valuable information. This book defines datacenters broadly, as large distributed systems that perform parallel computation for diverse users. These systems exist in multiple forms-private and public-and are built at multiple scales. Datacenter design and management is multifaceted, requiring the simultaneous pursuit of multiple objectives. Performance, efficiency, and fairness are first-order design and management objectives, each which can be viewed from several perspectives. This book surveys datacenter research from a computer architect's perspective, addressing challenges in applications, design, management, server simulation, and system simulation. This perspective complements the rich bodies of work in datacenters as a warehouse-scale system, which study the implications for infrastructure that encloses computing equipment, and in datacenters as a distributed systems, which employ abstract details in processor and memory subsystems. This book is written for first- or second-year graduate students in computer architecture and may be helpful for those in computer systems. The goal of this book is to prepare computer architects for datacenter-oriented research by describing prevalent perspectives and the state-of-the-art.
Benjamin C. Lee is the Nortel Networks Associate Professor of Electrical Engineering and Computer Science at Duke University. His research focuses on power-efficient architectures and emerging technologies for high-performance computer systems. He is also interested in the economics and public policy of computation. At Berkeley, he developed auto-tuning frameworks for sparse linear algebra. At Harvard, he introduced statistical machine learning for processor design. At Duke, Dr. Lee leads the System Integration Architecture Laboratory, which studies datacenter design and management with economic mechanisms and game theory.
Benjamin C. Lee is the Nortel Networks Associate Professor of Electrical Engineering and Computer Science at Duke University. His research focuses on power-efficient architectures and emerging technologies for high-performance computer systems. He is also interested in the economics and public policy of computation. At Berkeley, he developed auto-tuning frameworks for sparse linear algebra. At Harvard, he introduced statistical machine learning for processor design. At Duke, Dr. Lee leads the System Integration Architecture Laboratory, which studies datacenter design and management with economic mechanisms and game theory.
Dr. Lee received his B.S. in Electrical Engineering and Computer Science from the University of California at Berkeley, his Ph.D. in Computer Science at Harvard University, and his post-doctorate in Electrical Engineering at Stanford University. He has held visiting positions at Microsoft Research, Intel Labs, and Lawrence Livermore National Lab. Dr. Lee has received the NSF CAREER Award and the NSF Computing Innovation Fellowship. His research has been honored twice as Top Picks by IEEE Micro Magazine and twice as Research Highlights by the Communications of the ACM.
