TY - JOUR
T1 - Mobile agent based adaptive scheduling mechanism in peer to peer grid computing
AU - Choi, Sung Jin
AU - Baik, Maeng Soon
AU - Hwang, Chong Sun
AU - Gil, Joon Min
AU - Yu, Heon Chang
PY - 2005
Y1 - 2005
N2 - In a peer to peer grid computing environment, volunteers are exposed to failures such as crash and link failures. In addition, since volunteers can dynamically join and leave executions and they are not dedicated only to a peer to peer grid computing, the executions of volunteers are stopped or suspended more frequently than in a grid computing environment. These failures result in the delay and blocking of the executions of tasks and even partial or entire loss of the executions. In addition, these failures make it difficult for a volunteer server to schedule tasks and manage the allocated tasks as well as volunteers. Existing peer to peer grid computing systems, however, do not deal with these failures in scheduling mechanisms. Moreover, since existing scheduling mechanisms are performed only by a volunteer server in a centralized way, there is a high overhead. To solve these problems, we propose a mobile agent based adaptive scheduling mechanism (MAASM). We implemented MAASM in Korea@Home and ODDUGI mobile agent system. The MAASM reduces the overhead of volunteer server by using mobile agents in scheduling procedure in a distributed way. In addition, it tolerates the various failures(especially, volunteer autonomy failures) which frequently occur in a peer to peer grid computing environment. Consequently, MAASM guarantees reliable and continuous executions in spite of the failures, so it decreases total execution time.
AB - In a peer to peer grid computing environment, volunteers are exposed to failures such as crash and link failures. In addition, since volunteers can dynamically join and leave executions and they are not dedicated only to a peer to peer grid computing, the executions of volunteers are stopped or suspended more frequently than in a grid computing environment. These failures result in the delay and blocking of the executions of tasks and even partial or entire loss of the executions. In addition, these failures make it difficult for a volunteer server to schedule tasks and manage the allocated tasks as well as volunteers. Existing peer to peer grid computing systems, however, do not deal with these failures in scheduling mechanisms. Moreover, since existing scheduling mechanisms are performed only by a volunteer server in a centralized way, there is a high overhead. To solve these problems, we propose a mobile agent based adaptive scheduling mechanism (MAASM). We implemented MAASM in Korea@Home and ODDUGI mobile agent system. The MAASM reduces the overhead of volunteer server by using mobile agents in scheduling procedure in a distributed way. In addition, it tolerates the various failures(especially, volunteer autonomy failures) which frequently occur in a peer to peer grid computing environment. Consequently, MAASM guarantees reliable and continuous executions in spite of the failures, so it decreases total execution time.
UR - http://www.scopus.com/inward/record.url?scp=24944524167&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=24944524167&partnerID=8YFLogxK
U2 - 10.1007/11424925_98
DO - 10.1007/11424925_98
M3 - Conference article
AN - SCOPUS:24944524167
SN - 0302-9743
VL - 3483
SP - 936
EP - 947
JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
IS - IV
T2 - International Conference on Computational Science and Its Applications - ICCSA 2005
Y2 - 9 May 2005 through 12 May 2005
ER -