A Group Based Replication Mechanism to Reduce the Wastage of Processing Cycles in Volunteer Computing

作者：Khan Muhammad Khalid^*; Hyder Syed Irfan; Ahmed Ghayas Uddin; Begum Saira; Aamir Muhammad

来源：Wireless Personal Communications, 2014, 76(3): 591-601.

DOI：10.1007/s11277-014-1727-x

• Desktop Grids
• Fault Tolerance
• Replication
• Scheduling

Volunteer Computing is an exciting discipline for high throughput applications but due to inherent resource volatility, several issues arises such as tasks failure due to host / CPU unavailability, results validation etc. Replication is a solution to most of these problems but excessive replication can cause wastage of precious processing cycles. In this paper, we have extended our work of group based scheduling mechanism by proposing three replication policies i.e. proactive, reactive and hybrid used for Silver, Platinum and Gold groups respectively. The measures used for making the volunteer grouping are: collective impact of CPU & RAM, spot checking and task completion history. We evaluated the proposed mechanism over a network of 900 nodes having varied resources and behavior. The results are promising and show that the wastage of processing cycles can greatly be reduced by applying the suitable replication policies for each group.

• 出版日期2014-6