An Experimental Evaluation of the Adaptive Replica-assisted Migration for Parallel Storage Systems

Abstract

We have proposed a method replica-assisted migration that use temporally replica data in other storage nodes during executing data migration to keep both qualities of service accesses and data migration duration. However, we have not evaluated the method with real workload including update operations. In this paper, we first improve the method and evaluate an efficiency of the method under file server workloads. Base on the result of it, we also propose Adaptive replica-assisted migration for reduce migration duration.

Comparison of replica assisting and speed adjustment for service-aware horizontal data migration on autonomous disks

“…Our previous work [6] shows that one of our method to provide migration source node selection worked well with the workload, while the other of our method to forward partial accesses to replicas made no sense because of high skew access distribution and high update load. We consider using the concept of neighbor WAL method for eliminating quantities of load on replicas caused by write operations.…”

“…When each node has finished migrating the specified quantity of data, it sets its r i to zero. Detailed information of algorithm for calculating r i is de- scribed in [6]. Access forwarding to nodes storing replicas decreases the hit ratios of cache memory on each storage node because both the primary and replica data of other nodes may be stored in the same cache and will compete for cache space.…”

“…We have also proposed adaptive Replica-assisted Migration (RM), which reduces and distributes the load caused by data migration by using replica data for data redundancy [6]. The method reduces the amount of transferred data and provides a source node selection by using replica data location.…”