Bloom Filter Based Metadata Placement And Management In Cloud Computing

As the data storage in this successful cloud era progress towards zeta bytes suitable query mapping structure for efficient data retrieval becomes a major issue. Metadata is an appropriate way to locate and identify the data from these huge storage servers as metadata file holds the information about a file stored in the cloud data servers in order to map the query to the exact location of the data server where the data resides. This paper proposes a novel metadata model along with a special mapping data structure in order to make data retrieval in cloud efficiently. With fluctuating load in cloud, the metadata becomes dynamic and hence handling metadata operations efficiently is an important aspect of the file system as they constitute up-to half of file system workloads and managing these dynamic metadata along with a suitable data structure is critical in scaling the overall performance. The proposed metadata model in this chapter explores a novel method of metadata management mechanism using a specially designed bloom filter called enhanced cloud bloom filter, a probabilistic data structure, which is used for efficient retrieval of data stored across various data storage servers located in geographically dispersed locations. This model also proposes a scalable and decentralized metadata placement schemes using Global Bloom Filter which scales up the metadata performance by judiciously distributing heavy management workloads among multiple metadata servers. The proposed model efficiently handles the dynamic metadata with effective update mechanism using Local Bloom Filter.