TY - GEN
T1 - A Static Uniform Job Assignment Algorithm to Workers in User-PC Computing System
AU - Zhou, Xudong
AU - Funabiki, Nobuo
AU - Sugawara, Shinji
AU - Htet, Hein
AU - Kamoyedji, Ariel
N1 - Publisher Copyright:
© 2022 ACM.
PY - 2022/7/29
Y1 - 2022/7/29
N2 - Currently, we are studying the User-PC computing system (UPC) as a low-cost high-performance distributed computing platform following the master-worker model. It uses idling resources of personal computers (PCs) of members in a group. Adopting Docker, it allows a worker PC to execute various jobs. Previously, we proposed algorithms of assigning non-uniform jobs to workers in the UPC system, where job programs are much different from each other. However, some applications need to execute a lot of uniform jobs that use the same program but the slightly different input data. Then, the total CPU time becomes nearly linear to the number of jobs. In this paper, we propose a static uniform job assignment algorithm to workers in the UPC system. To minimize the maximum makespan among the workers, linear equations are derived to find the optimal assignment such that the CPU time to complete the assigned jobs becomes equal between the workers. For evaluations, Android programming learning assistance system (APLAS) that has been developed in our group is selected as the typical application, where the software testing program should execute with various source codes from students. We generated the assignment of 578 jobs to six workers using the proposal, and executed them in the UPC system. The reduction of results confirmed the effectiveness of our proposal in the scenarios of running uniform jobs.
AB - Currently, we are studying the User-PC computing system (UPC) as a low-cost high-performance distributed computing platform following the master-worker model. It uses idling resources of personal computers (PCs) of members in a group. Adopting Docker, it allows a worker PC to execute various jobs. Previously, we proposed algorithms of assigning non-uniform jobs to workers in the UPC system, where job programs are much different from each other. However, some applications need to execute a lot of uniform jobs that use the same program but the slightly different input data. Then, the total CPU time becomes nearly linear to the number of jobs. In this paper, we propose a static uniform job assignment algorithm to workers in the UPC system. To minimize the maximum makespan among the workers, linear equations are derived to find the optimal assignment such that the CPU time to complete the assigned jobs becomes equal between the workers. For evaluations, Android programming learning assistance system (APLAS) that has been developed in our group is selected as the typical application, where the software testing program should execute with various source codes from students. We generated the assignment of 578 jobs to six workers using the proposal, and executed them in the UPC system. The reduction of results confirmed the effectiveness of our proposal in the scenarios of running uniform jobs.
KW - distributed computing platform
KW - linear equations
KW - static assignment
KW - uniform job
UR - http://www.scopus.com/inward/record.url?scp=85140738737&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85140738737&partnerID=8YFLogxK
U2 - 10.1145/3556223.3556227
DO - 10.1145/3556223.3556227
M3 - Conference contribution
AN - SCOPUS:85140738737
T3 - ACM International Conference Proceeding Series
SP - 24
EP - 30
BT - Proceedings of the 10th International Conference on Computer and Communications Management, ICCCM 2022
PB - Association for Computing Machinery
T2 - 10th International Conference on Computer and Communications Management, ICCCM 2022
Y2 - 29 July 2022 through 31 July 2022
ER -