EALARM: Enhanced Autonomic Load-Aware Resource Management for P2P Key-Value Storage in Cloud

In the cloud environment, load balancing is a fundamental technique for optimizing utilization and achieving sharing of computing infrastructures. However, it becomes impractical to tackle load balancing issue manually as the enlargement of cloud cluster. In this paper, we thus put forward EALARM, a novel model for dynamical load balancing on P2P key-value storage system in cloud. This model measures the computing resource utilization on physical nodes and proceeds data migration in the unit of virtual servers from overloaded physical machines to under-loaded machines. If the above step fails, the system will then scale up or down automatically. In order to coordinate the physical and virtual machines, we build the physical layer in a hierarchical tree structure on the logical layer of distributed DHT. Experiment results indicate that EALARM model enhances the performance of previous ALARM [13].