Mobile Computing Systems

Mobile wireless systems present a major computer systems design challenge. High performance for multimedia tasks is required while preserving battery life. Our research approach is to build and measure prototype systems, mostly based on the HP iPaq PDAs used in MIT Project Oxygen.

Energy-aware lossless data compression examines the tradeoffs between wireless transmission energy and the local compute energy needed to perform data compression. For our portable system (the iPAQ-like "Skiff"), sending a single bit of data requires the same energy as ~1000 ALU operations. If a data compressor can use less than 1000 ALU operations to shrink a file by a single bit, energy will be saved. However, we have found cases in which common compressors spend more energy to compress a file than they save by reducing wireless transfers. Our work measures several compression applications and algorithms; examines how they use energy; and makes suggestions (including an asymmetric scheme) that save over 57% of total system energy [1, 3, 7].

Bounded web slowdown is a technique to reduce the performance impact of 802.11 wireless card power down modes [2, 6].

A working handheld video-over-IP application [4] shows the potential of reconfigurable logic in accelerating handheld applications. (Video demonstration: MPEG 13MB).

A handheld power tracker is being developed to monitor both hardware and software sources of power dissipation during normal operation [5].

Publications

[1] "Energy Aware Lossless Data Compression", Kenneth C. Barr, S.M. Thesis, Massachusetts Institute of Technology, September 2002. (PDF)
[2] "Minimizing Energy for Wireless Web Access with Bounded Slowdown", Ronny Krashinsky and Hari Balakrishnan, MobiCom 2002, Atlanta, GA, September 2002. (PDF paper, PDF slides, PPT slides, simulation code)
[3] "Energy Aware Lossless Data Compression", Kenneth Barr and Krste Asanovic, Winner, Best Paper Award, First International Conference on Mobile Systems, Applications, and Services (MobiSys-2003), San Francisco, CA, May 2003. (PDF paper, PPT Slides)
[4] "Video Over IP: An Example Reconfigurable Computing Application for a Handheld Device", Elina Kamenetskaya, M.Eng. Thesis, Massachusetts Institute of Technology, May 2003. (PDF)
[5] "Eprof: An Energy Profiler for the iPAQ", Kelly Koskelin M.Eng. Thesis, Massachusetts Institute of Technology, February 2004. (PDF)
[6] "Minimizing Energy for Wireless Web Access with Bounded Slowdown", Ronny Krashinsky and Hari Balakrishnan, ACM/Kluwer Journal on Wireless Networks (WINET), Vol. 11, No. 1-2, pp. 135-148, January 2005. (Extended version of MobiCom 2002 paper) (PDF paper)
[7] "Energy-Aware Lossless Data Compression", Kenneth Barr and Krste Asanovic, ACM Transactions on Computer Systems,24(3):250-291, August 2006. (Extended version of MobiSys-2003 Paper) (PDF)

Funding

We gratefully thank the past and present sponsors of this work, including NSF, DARPA, HP, and Project Oxygen.