Title :
Three Ways to Improve the Efficiency of Virtual/Cloud Computing Lab
Author_Institution :
Sch. of Comput. Sci. & Software Eng., Univ. of Electron. & Sci. Technol. of China, Chengdu
Abstract :
Virtual computing lab (VCL) initiated at North Carolina State University (NCSU), is a virtualized, distributed and on-demand computing service architecture provided through Internet. As VCL related activities scale up to many adjacent states as well as to other countries including Canada and India, improving the system efficiency and utilization becomes an important consideration. In this paper, we consider the VCL consisting of m identical computing elements such as blades. Based on statistical and simulation analysis of real data from year 2004 to 2007, the VCL management system can be modeled by classical queueing systems. It is possible to use queueing techniques to reduce preloading time, control the service time distribution, dimension the size of servers etc. to improve VCL efficiency. In this paper, we analyze and show how to improve VCL system efficiency in three ways.
Keywords :
Internet; computer aided instruction; queueing theory; statistical distributions; student experiments; virtual machines; Erlang delay model; Internet; North Carolina State University; VCL management system; cloud computing lab; computer aided instruction; distributed computing service architecture; on-demand computing service architecture; queueing system; service time distribution; statistical analysis; virtual computing lab; Analytical models; Blades; Cloud computing; Computational modeling; Computer architecture; Data analysis; Distributed computing; Queueing analysis; Size control; Web and internet services; Erlang delay model; Virtual computing lab; system efficiency;
Conference_Titel :
Apperceiving Computing and Intelligence Analysis, 2008. ICACIA 2008. International Conference on
Conference_Location :
Chengdu
Print_ISBN :
978-1-4244-3427-5
Electronic_ISBN :
978-1-4244-3426-8
DOI :
10.1109/ICACIA.2008.4769995
