Research On Reliability Of Information System In Cloud Environment

With the rapid development of science and technology, it has entered the Internet+ era. Information processing from all walks of life is associated with the Internet and information system, and every aspect of our life is inseparable with the Internet. Data volume increases sharply, therefore it is called “big data”. In the new environment of large amount of data and complex data structure, cloud computing emerged at the right moment. In cloud computing, configurable computing resource pool can be shared through network. It's fast, flexible,resource sharing, and service can be measured. Hadoop, an open source distributed computing platform, is now the mainstream of cloud computing platform. It can develop and run large data program. In cloud computing mode,large scale service and multiple type increase the unreliable factors of information system that may lead to system failure. Cloud computing system which has numerous users is large and widely used. A failure can cause severe losses, so it is very urgent to improve the reliability of information system in cloud.This paper analyzed the reliability of the information system in cloud environment based on Hadoop platform. It improved the architecture of Hadoop Distributed File System(HDFS) and proposed job scheduler based on reliability,the specific research contents are as follows:1. To elaborate the core technology of Hadoop: HDFS and Map Reduce, to analyze the executing process and scheduling mechanism, to study the scheduling algorithms of Hadoop and compare merits and demerits.2.There's only one Name Node in HDFS cluster which may cause single node failure. To improve the HDFS structure to solve the problem. To propose the selection mechanism of Master based on Paxos algorithm and set up Name Node cluster.3.To analyze the reliability of node and link in cluster, to set up scheduling model, to put forward Job Scheduler Based on Reliability of Node and Link so the reliability is improved.4. To set up the experiment platform and perform an experiment of the improved HDFS structure and JSBRNL algorithm. It proves that the reliability of information system is improved.