TY - GEN
T1 - Optimistic scheduling algorithm for mobile transactions based on reordering
AU - Kim, SungSuk
AU - Hwang, Chong Sun
AU - Yu, Heonchang
AU - Lee, Sang-Geun
PY - 2001
Y1 - 2001
N2 - In this paper, we propose transaction processing algorithms in the broadcast environment. We take an optimistic approach for mobile transactions because (a) it needs a small number of messages for maintaining transactional consistency, and (b) it can make good use of broadcasting facilities from the servers. The more data conflicts occur, however, the more mobile transactions may be aborted. Thus, we accept reordering technique to reduce the number of aborted transactions; that is, whenever any kind of conflict is found from broadcast information, the system determine the operation orders without violating transactional consistency, not just aborting the mobile transactions unconditionally. The proposed algorithms - O-Post algorithm for update transaction and O-Pre algorithm for read-only transaction - do not need much information from the server while resulting in serializable executions. Finally, we also evaluate the performance behavior through simulation study.
AB - In this paper, we propose transaction processing algorithms in the broadcast environment. We take an optimistic approach for mobile transactions because (a) it needs a small number of messages for maintaining transactional consistency, and (b) it can make good use of broadcasting facilities from the servers. The more data conflicts occur, however, the more mobile transactions may be aborted. Thus, we accept reordering technique to reduce the number of aborted transactions; that is, whenever any kind of conflict is found from broadcast information, the system determine the operation orders without violating transactional consistency, not just aborting the mobile transactions unconditionally. The proposed algorithms - O-Post algorithm for update transaction and O-Pre algorithm for read-only transaction - do not need much information from the server while resulting in serializable executions. Finally, we also evaluate the performance behavior through simulation study.
UR - http://www.scopus.com/inward/record.url?scp=84944053454&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84944053454&partnerID=8YFLogxK
U2 - 10.1007/3-540-44498-x_9
DO - 10.1007/3-540-44498-x_9
M3 - Conference contribution
AN - SCOPUS:84944053454
SN - 3540414541
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 105
EP - 117
BT - Mobile Data Management - 2nd International Conference, MDM 2001, Proceedings
A2 - Tan, Kian-Lee
A2 - Franklin, Michael J.
A2 - Lui, John Chi-Shing
PB - Springer Verlag
T2 - 2nd International Conference on Mobile Data Management, MDM 2001
Y2 - 8 January 2001 through 10 January 2001
ER -