Show simple item record

dc.contributor.advisor Perera, I
dc.contributor.author Nelson, SD
dc.date.accessioned 2018
dc.date.available 2018
dc.date.issued 2018
dc.identifier.uri http://dl.lib.mrt.ac.lk/handle/123/15798
dc.description.abstract Node management techniques especially designed for the modern systems are becoming more and more advanced and complex with the research involvement from the industry and institutions today. Almost all type industries with the computing systems are trending towards the distributed environments now. In this context node management plays a very significant and important role which helps to fulfill the main aspect of the distributed-ness. Clustering is the process of managing, maintaining and binding different nodes together sharing common set of configurations to work together for achieving a common goal. In this project, we design and implement an approach towards a lightweight caching framework for a distributed environment. Here, we evaluate several protocols and choose WebSockets protocol for implementing the node management of the cluster. In our approach, each node maintains node server and node client set in a multithreaded environment to cater the node management. We have detailed out the design and implementation aspects on chapter 4. Then we demonstrate the performance achievements of the caching framework by using it on an application, having small to medium scale cluster. Then through a simulated use-case of the caching cluster framework, the system stability is monitored while up scaling the system step by step. Finally, the above-mentioned use case is extended to a simulated experimental evaluation making a comparison with a reference framework. For that experiment, a reference framework which based on TCP protocol is implemented with the help of Netty framework. We have concluded by highlighting the significance of our caching cluster framework by the outstanding behaviors and features such as efficiency, lightweight-ness, and scalability, stability on scaling and very low system overhead. en_US
dc.language.iso en en_US
dc.subject COMPUTER SCIENCE AND ENGINEERING-Dissertations en_US
dc.subject CLUSTER MANAGEMENT en_US
dc.subject NODE MANAGEMENT en_US
dc.title A Lightweight caching framework en_US
dc.type Thesis-Full-text en_US
dc.identifier.faculty Engineering en_US
dc.identifier.degree MSc in Computer Science and Engineering en_US
dc.identifier.department Department of Computer Science & Engineering en_US
dc.date.accept 2018
dc.identifier.accno TH3790 en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record