Bhushan P. Jain

Purlieus: locality-aware resource allocation for MapReduce in a cloud

We present Purlieus, a MapReduce resource allocation system aimed at enhancing the performance of MapReduce jobs in the cloud. Purlieus provisions virtual MapReduce clusters in a locality-aware manner enabling MapReduce virtual machines (VMs) access to input data and importantly, intermediate data from local or close-by physical machines. We demonstrate how this locality-awareness during both map and reduce phases of the job not only improves runtime performance of individual jobs but also has an additional advantage of reducing network traffic generated in the cloud data center. This is accomplished using a novel coupling of, otherwise independent, data and VM placement steps. We conduct a detailed evaluation of Purlieus and demonstrate significant savings in network traffic and almost 50% reduction in job execution times for a variety of workloads.

Methodology for Usage of Emerging Disk to Ameliorate Hybrid Storage Clouds

With the dramatic evolution of various greener disk and memory technologies, helped in rapid establishment of Hybrid storage environment comprising of heterogeneous storage units. In order to provide an efficient and optimised storage solution there exists a necessity of adapting smarter changes in the management stack that enables compact sensing, processing, decision making capability based on the importance of data and disk life span predicted on operational workloads. This paper reviews system architecture for two faces of RAS cloud features namely disaster recovery planning, disk breakage prediction independent of disk technology and host aware data tier based on disk life span.

Optimization of disk replacement as a large domain objective function

There are several motivations for trying to find hardware infrastructure which can be replaced. Most administrators believe in owning a diversified portfolio of disks, tapes to replace older ones, a strategy that has been widely accepted and practiced. When attempting to create an efficient portfolio of disks to be replaced there are numerous factors to consider set of fitness heuristics over a population of disks, the goal here is to find a portfolio that has a high probability of failure and at the same time will ensure savings in power consumption and carbon emission. Given the fundamental financial and price information of power consumed, cooling required, we attempt to use GA (Genetic Algorithm) and a hybrid PSO-EO (Particle Swarm Optimization) algorithm to identify disks that are likely to outperform the existing ones by preventing failure and having excess returns and compare these two approaches.