Title :
Improvement on dynamic migration technology of virtual machine based on Xen
Author :
Sun Mingsong ; Ren Wenwen
Author_Institution :
Network Inf. Center, Harbin Univ. of Sci. & Technol., Harbin, China
fDate :
June 28 2013-July 1 2013
Abstract :
Virtualization technology has been very popular in recent years since it can provide good server consolidation, independence and manageability. The most valuable and striking aspect is the dynamic migration for virtual machine. Aiming at longer time and memory pages repeated retransmission in the process of memory pre-copy, the traditional dynamic migration framework for virtual machine was improved, probability prediction module and memory compression module was added to the framework in this paper. A working set calculating algorithm based on dirty page probability prediction was proposed. By compressing memory pages in the migration field, the amount of data migration is reduced and the network burden is lowered. Probability prediction algorithm and memory compression algorithm was put forward, data structures of the two algorithms and design ideas was described. Experimental results show that the framework can effectively improve migration performance.
Keywords :
data compression; probability; virtual machines; virtualisation; Xen; data migration; data structures; dirty page probability prediction; dynamic migration technology; memory compression algorithm; memory precopy process; virtual machine; virtualization technology; Benchmark testing; Dynamic scheduling; Oils; Optimization; Servers; Sun; Switches; Xen; dynamic migration framework; dynamic migration of virtual machine; optimization algorithm; virtualization;
Conference_Titel :
Strategic Technology (IFOST), 2013 8th International Forum on
Conference_Location :
Ulaanbaatar
Print_ISBN :
978-1-4799-0931-5
DOI :
10.1109/IFOST.2013.6616868