A Self-Optimized Storage for Distributed Data as a Service

In this paper we propose an automated partial replication storage technique for the cloud, which we call ssCloud. We associate this technique with a dual direction load balancing technique using the cloud servers. We show that by having dual direction load balancing and an automated partial replication of the storage we enhance the download speed from the cloud and reduce the cost of storing multiple replicas on multiple cloud nodes. This is done by having two cloud interfaces, one for the download requests from the clients (load balancing module) and another for uploading and splitting files into blocks and saving them on the different cloud servers (FileController). We also compare our approach to other similar approaches and describe the benefits of using ssCloud.