Loading...

Resource Allocation for Big Data Analysis in Clouds

Nabavinejad, Morteza | 2017

493 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 50893 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Goudarzi, Maziar
  7. Abstract:
  8. Forecasts predict that the volume of digital data will increase by 300 times in 2020 compared with 2005. This significant growth further emphasizes the importance of Big Data as well as Big Data Processing. MapReduce and its open source implementation Hadoop are prevailing frameworks for implementing Big Data Analytics and applications. Because of inherently huge amount of data and computational requirements of Big Data applications, acquisition of large amount of computational resources is necessary. However, managing in-house clusters to respond the computational requirements is costly such that small- and middle-sized companies either cannot afford it, or find cloud-based solutions economically more attractive. Consequently increasingly more companies are moving towards the on-demand resources available in the cloud. The cloud computing concept provides the opportunity to employ the required computational resources in the form of VMs and pay for them based on the “pay-as-you-go” pricing model. The cloud providers have even launched dedicated services such as Amazon Elastic MapReduce (EMR) service to satisfy the growing demand of computational resources for MapReduce applications. While deploying Big Data applications on cloud platform brings fascinating opportunities, there are concerns that need to be addressed. Big Data challegnes such as data skew, data variety and Cloud challenges such as VM interference need to be tackled to maximize the benefit of Cloud resources for Big Data. In this research, we propose new performance estimation and resource allocation methods to conquer aforementioned challegnes and increase the jobs performance in Cloud
  9. Keywords:
  10. Energy Management ; Cloud Computing ; Big Data ; Optimum Allocation ; Resources Allocation

 Digital Object List

 Bookmark

No TOC