Font Size: a A A

Design And Implementation Of Distributed Storage System Based On Master Slave Server

Posted on:2017-03-12Degree:MasterType:Thesis
Country:ChinaCandidate:R WangFull Text:PDF
GTID:2308330482481792Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the development of computer network, especially the vigorous development of Internet, the concept of traditional information system comes across great changes, the distributed information system based on the network has been widely used, playing an increasingly prominent role in the life of society. How to release and get information effectively, in a more wide and heterogeneous computing environment, has become a serious problem. With the rapid development of multimedia technology, audio, video, images and other non-structured documents are growing explosively. In order to facilitate people for these files’ storage, cloud storage came into being. Today’s cloud storage technology has improved, but as far as safety, consistency and fault-tolerance are concerned, there still exists room for further improvement.Based on a practical development of a distributed storage system VSCS(Very Secure Cloud System), We do the following work to solve a series of problems existing in today’s distributed software:1. On security issues, with the help of the open source software OpenSSL, we add SSL security mechanism, which makes data transmission more secure and reliable.2. On the issues of the cluster’s consistency, including the consistency of master server and database, databases, the master server and the slave server, we use the defense measures in an inconsistent state to ensure the stable operation of the entire cluster.3. On scalability issues, this paper uses two kinds of strategies to achieve scalability of the cluster, one is a method of configuration file, the other is a command line method.4. Due to the cluster establishment is heavily dependent on the network, operating system and other infrastructure, which greatly increased the instability of the entire cluster. When node failure occurs, with the help of cluster-monitor server, the system take emergency measures, replacing the bad slave server node, which makes users access files normally.
Keywords/Search Tags:distributed storage, load balance, consistency, security, exception handling, high concurrency, multi-thread
PDF Full Text Request
Related items