Load balancing in in-memory key-value stores for response time minimization

Abstract

In-memory key-value stores (IMKVS) have now turned into a mainstream technology in order to meet with demanding temporal application requirements under heavy loads. This work examines the factors that affect the load distribution in IMKVS clusters as well as migration policies that cure the problem of unbalanced loads as a means to provide response time guarantees. Experiments are conducted in a Redis deployment under various settings. The results show that the key distribution and key length are contributing factors to the load balancing problem and impact the cluster’s response times. On the contrary, key popularity and query volume seem to have minor or no effect at all.