![]() |
Professor Gautam Natarajan Title: Towards
Green Computing: Energy Management in Data
Centers |
October
19, 2007
Presentation: 1:30-2:30pm; Venue: ATRC 101
Discussion: 2:30-3:00pm; Venue: VOGT room (adjacent to ATRC 101)
ABSTRACT:
Data centers typically consist of
hundreds to thousands of servers that consume a phenomenal amount of electric
power. In the seminar we initially describe various research issues that need to
be addressed in this specific context as well as the role of industrial
engineers in a more generic context, namely, green computing. Then we present a
resource management problem (for optimally turning off servers, moving
applications, controlling server speeds, etc.) to reduce energy consumption in
data centers. We take advantage of the fact that user requests vary tremendously
both spatially and temporally. Using a statistical model to predict
user-request patterns and a queueing model to obtain performance
measures, we formulate and solve a mixed non-linear-integer programming
problem for resource management. One of the bi-products of the optimization is
that the system runs at very high loads with tremendous traffic variability and
long-range dependence. The talk would conclude by addressing effective traffic
models under such conditions.
BIOSKETCH:
N. Gautam is an associate professor in the department of Industrial and Systems Engineering with a courtesy appointment in the department of Electrical and Computer Engineering at Texas A&M University. Prior to joining Texas A&M University in 2005, he was on the Industrial Engineering faculty at Penn State University for eight years. He received his Ph.D. and M.S. in Operations Research from the University of North Carolina at Chapel Hill, and his B.Tech. in Mechanical Engineering from Indian Institute of Technology, Madras in 1993. Gautam's interests are in optimal design, control and performance evaluation of service systems, specifically computer-communication networks, transportation systems and information flows.
Link to Dr.Natarajan's personal Webpage