Securing User’s Data in HDFS

摘要

With th e advent of Technology and in creasing growth in volume of data the business are finding th e cloud as suitable option to host their data. However, putting the sensitive data on third-party infrastru cture poses several security risks to th eir data utilizing the advantages of the Clouds. Though there are many risks and concerns are involved in cloud computing according man y surveys conducted by different organizations the prime concern of clients when opting for cloud solution is the security of th eir data. Key issue is to protect important data from unauthorized access by adversaries in case the confidentiality of data is broken by internal or external attacks on the cloud hosting those data. HDFS is the file system su itable for storing and processing large volume of data using MapReduce model. When public cloud is based on the Hadoop which uses HDFS to store data, the data are stored in plain text and by default the transport of data is also insecure wh en client submit the data to storage servers on cloud. Requirement here is design and implement a prototype to secure th e HDFS to harn ess is with security features so that it can be deployed in public cloud to provide storage and computing services. We have proposed and implemented secureHDFS by incorporating Elliptic Curve Integrated Encryption which provides data confidentiality as well as integrity in Hadoop. Experiments were carried out to analyze th e performance with respect to other hybrid encryption schemes