Exploiting Dynamic Resource Allocation for Efficient Parallel Data Processing in the Cloud

About Project:

In recent years ad hoc parallel data processing has emerged to be one of the killer applications for Infrastructure-as-a-Service (IaaS) clouds. Major Cloud computing companies have started to integrate frameworks for parallel data processing in their product portfolio, making it easy for customers to access these services and to deploy their programs. However, the processing frameworks which are currently used have been designed for static, homogeneous cluster setups and disregard the particular nature of a cloud. Consequently, the allocated compute resources may be inadequate for big parts of the submitted job and unnecessarily increase processing time and cost. In this paper, we discuss the opportunities and challenges for efficient parallel data processing in clouds and present our research project Nephele. Nephele is the first data processing framework to explicitly exploit the dynamic resource allocation offered by today’s IaaS clouds for both, task scheduling and execution. Particular tasks of a processing job can be assigned to different types of virtual machines which are automatically instantiated and terminated during the job execution. Based on this new framework, we perform extended evaluations of MapReduce-inspired processing jobs on an IaaS cloud system and compare the results to the popular data processing framework Hadoop.

Today a growing number of companies have to process huge amounts of data in a cost-efficient manner. Classic representatives for these companies are operators of Internet search engines, like Google, Yahoo, or Micros ft. The vast amount of data they have to deal with every day has made traditional database solutions prohibitively expensive . Instead, these companies have popularized an architectural paradigm based on a large number of commodity servers. Problems like processing crawled documents or regenerating a web index are split into several independent subtasks, distributed among the available nodes, and computed in parallel. In order to simplify the development of distributed applications on top of such architectures, many of these companies have also built customized data processing frameworks. Examples are Google’s MapReduce , Microsoft’s Dryad , or Yahoo!’s Map-ReduceMerge . They can be classified by terms like high-throughput computing (HTC) or many-task computing (MTC), depending on the amount of data and the number of tasks involved in the computation . Although these systems differ in design, their programming models share similar objectives, namely hiding the hassle of parallel programming, fault tolerance,and execution optimizations from the developer. Developers can typically continue to write sequential programs. The processing framework then takes care of distributing the program among the available nodes and executes each instance of the program on the appropriate fragment of data. For companies that only have to process large amounts of data occasionally running their own data center is obviously not an option. Instead, Cloud computing has emerged as a promising approach to rent a large IT infrastructure on a short-term pay-per-usage basis. Operators of so-called IaaS clouds, like Amazon EC2 , let their customers allocate, access, and control a set of virtual machines (VMs) which run inside their data centers and only charge them for the period of time the machines are allocated. The VMs are typically offered in different types, each type with its own characteristics (number of CPU cores, amount of main memory, etc.) and cost.