Affiliations: RMIT University, School of Computer Science and
Information Technology, GPO Box 3476V, Melbourne, Vic 3001 Australia. Tel.:
+614 139 996 35; Fax: +613 966 216 17; E-mail: {kwonlai, zahirt,
pbertok}@cs.rmit.edu.au
Note: [] Corresponding author
Abstract: Traffic and access delay can be reduced in a mobile network by
caching data objects at network nodes near the clients. Traditional caching
techniques, however, are unsuitable in this environment because they do not
account for the changing location of the users. To deal with this problem,
cache relocation techniques can be applied to dynamically relocate data objects
so they remain close to the moving clients. Existing relocation techniques rely
heavily on path prediction. Unfortunately, the inaccuracy of path prediction
can lead to high relocation overhead and poor response time. This paper
presents an analytical study of the effects of client mobility on network cache
performance. Two new cache relocation techniques are proposed to deal with the
issue of poor path prediction and reduce the overhead of existing cache
relocation schemes. The first technique, 2PR, compensates for poor path
prediction by temporarily moving data objects to a common parent node prior to
a handover. Objects are moved to the correct destination once the client's new
location has been confirmed. The second technique, ROLP, reduces the traffic
overhead associated with cache relocation by ensuring duplicate objects are not
relocated and relocation of objects are performed only from the nearest node to
the destination. Test results show that 2PR reduces the query delay experienced
by mobile clients by 60 to 83% after handovers, while ROLP reduces the
overhead of cache relocation by between 33 to 65% compared to existing
schemes.
Keywords: User mobility, cache relocation, transparent handovers
