A Load Balancing Scheme for Distributed Key-Value Caching System in Cloud Environment

Distributed key-value caching system has been deployed in many kinds of clouds. The effect of load balancing between each node is crucial in key-value caching system in clouds. Invalidation is a feature of the data in the key-value caching system, and making some cache invalid is efficient instead of adjusting the load location when load balancing. It is worth studying that how to utilize the feature reasonable to reach maximum load balancing. Aiming at this point, a new cache-invalidation-scope model in key-value caching system is proposed. Combined with greedy algorithm, the scheme provides a better load balancing algorithm for different load cases (CLB). CLB algorithm utilizes entropy and the scope of invalid cache invalid as the evaluation basis of load balancing effect. Compared with existing algorithm, CLB load balancing algorithm lifts up the performance of key-value caching system more than one times.