Data Technology industry has been utilizing the customary social databases for around 40 years. Be that as it may, in the latest years, there was a generous transformation in the IT business as far as business applications. Remain solitary applications have been supplanted with electronic applications, conferred servers with different proper servers and committed stockpiling with framework stockpiling. Lower expense, adaptability, the model of pay-as-you-go are the fundamental reasons, which caused the conveyed processing are transformed into reality. This is a standout amongst the hugest upsets in Information Technology, after the development of the Internet. Cloud databases, Big Table, Sherpa, and SimpleDB are getting the opportunity to be more natural to groups. They featured the hindrances of current social databases as far as convenience, adaptability, and provisioning. Cloud databases are basically utilized for data raised applications, for example, stockpiling and mining of gigantic information or business information. These applications are adaptable and multipurpose in nature. Various esteem based data organization applications, such as managing an account, online reservation, e-exchange and stock organization, and so on are delivered. Databases with the help of these sorts of uses need to incorporate four essential highlights: Atomicity, Consistency, Isolation, and Durability (ACID), in spite of the fact that utilizing these databases isn't basic for utilizing as a part of the cloud. The objective of this paper is to discover the points of interest and disservices of databases generally utilized in cloud frameworks and to survey the difficulties in creating cloud databases
ROMANet: Fine-Grained Reuse-Driven Off-Chip Memory Access Management and Data Organization for Deep Neural Network Accelerators
