Adaptive data placement for improving performance of online social network services in a multicloud environment

The existing online social network (OSN) services in amultiple-cloud (Multicloud) environment use replications to store user data for improving the service performance. However, it not only generates tremendous traffic for synchronization between data but also stores considerable redundant data, thus causing large storage costs. In addition, it does not provide dynamic load balancing considering the resource status of each cloud.As a result, it cannot cope with the degradation of performance caused by the resource contention. We introduce an adaptive data placement algorithm without the replications for improving the performance of the OSN services in the Multicloud environment. Our approach is designed to avoid server overhead using data balancing technique,which locates data from a cloud to another according to the amount of traffic. To provide acceptable latency delay, it also considers the relationship between users and the distance between user and cloud when transferring data. To validate our approach, we experimented with actual users' locations and times of use collected from OSN services. Our findings indicate that this approach can reduce the resource contention by an average ofmore than 59%, reduce storage volume to at least 50%, andmaintain the latency delay under 50ms.