Green Clusters

• Thesis title in Czech: Green Clusters
• Thesis title in English: Green Clusters
• Key words: Map-Reduce, Energy-efficient, ARM, FPGA, Big-Data
• English key words: Map-Reduce, Energy-efficient, ARM, FPGA, Big-Data
• Academic year of topic announcement: 2010/2011
• Thesis type: diploma thesis
• Thesis language: angličtina
• Department: Department of Distributed and Dependable Systems (32-KDSS)
• Supervisor: RNDr. Leo Galamboš, Ph.D.
• Author: hidden - assigned and confirmed by the Study Dept.
• Date of registration: 11.11.2010
• Date of assignment: 11.11.2010
• Date and time of defence: 03.02.2015 11:30
• Date of electronic submission: 04.12.2014
• Date of submission of printed version: 05.12.2014
• Date of proceeded defence: 03.02.2015
• Opponents: doc. RNDr. Martin Kruliš, Ph.D.

Guidelines

The thesis discusses distributed computing over a large data sets, e.g. MapReduce. This is often implemented with a great amount of computers which tend to consume ludicrous amount of energy.

The main objective of this thesis is to study if it's possible to save energy while maintaining the same computing performance. To achieve this task, the thesis discusses the possibility of replacing conventional computers with a grid of ARM/MIPS or possibly FPGA devices.

A grid of ARM/MIPS devices would be overly expensive, so the QEMU emulator may help to reduce the cost of aparature. The final power consumption approximation will be obtained on a small amount of real hardware devices.

The thesis will also study the possibilities to improve and optimise the MapReduce algorithm on particular hardware.

References

Jeffrey Dean and Sanjay Ghemawat (2004) MapReduce: Simplified Data Processing on Large Clusters. OSDI'04: Sixth Symposium on Operating System Design and Implementation, San Francisco, CA, December, 2004.
MIPS64 (2010) http://www.mips.com/products/architectures/mips64/
Clive Maxfield (2004) "The Design Warrior's Guide to FPGAs".Published by Elsevier, 2004. ISBN 0750676043, 9780750676045.