Improving Read/Write Performance for Key-value Storage System by Automatic Adjustment of Consistency Level

Distributed Key-value database is designed for storing, retrieving, managing associative arrays and data is replicated across different nodes for high availability, and no single point of failure. In such systems, Apache Cassandra is a peer-to-peer architecture which any user can connect to any node in any data center and can read and write data anywhere. Most of the systems usually select a fixed number of replicas of read/write requests in key-value storage. When the more replicas a read request chooses, it may increase the response time and reduce the system performance. Therefore, the encoded data is written to a Distributed Hash Table (DHT) and the decoded data are retrieved by automatically adjusting the read and write consistency level. The proposed approach tends to achieve the read/write performance of client requests for key-value storage system by automatically select the minimum number of consistent replicas of defining the consistency level