MapReduce is emerged as a prominent programming model for data-intensive computation. In this work, we study power-aware MapReduce scheduling in the speed scaling setting first introduced by Yao et al. [FOCS 1995]. We focus on the minimization of the total weighted completion time of a set of MapReduce jobs under a given budget of energy. Using a linear programming relaxation of our problem, we...

MapReduce is an emerging programming paradigm for data parallel applications proposed by Google to simplify large-scale data processing. MapReduce implementation consists of map function that processes input key/value pairs to generate intermediate key/value pairs and reduce function that merges and converts intermediate key/value pairs into final results. The reduce function can only start pro...

We use Hadoop, an open-source implementation of Google’s distributed file system and the MapReduce framework for distributed data processing, on modestly-sized compute clusters to evaluate its efficacy for standard machine learning tasks. We show benchmark performance on searching and sorting tasks to investigate the effects of various system configurations. We also distinguish classes of machi...

Spot market provides the ideal mechanism to leverage idle CPU resources and smooth out the computation demands. Unfortunately, few applications can take advantage of spot market because they cannot handle sudden terminations. We describe Spot Cloud MapReduce, the first MapReduce implementation that can fully take advantage of a spot market. Even if a massive number of nodes are terminated regul...