RSD: Rate-Based Sync Deferment for Personal Cloud Storage Services

Abstract

Cloud storage services, such as Dropbox, Google Drive, and OneDrive, to cite a few, are becoming an increasingly “vital” tool in our everyday life. Unluckily, these services can incur large network overhead in different usage scenarios. To reduce it, these systems utilize several techniques, such as source-based deduplication, chunking, delta compression, and so on. One of these techniques is sync deferment, which relies on the packing of updates to intentionally defer the synchronization process for some time, and increase the volume of useful data per overhead byte. The scientific literature has shown this technique to be very helpful, though there are still some limitations on current solutions. To resolve them, we present here a new adaptive sync deferment method, which is comparable to the current state of the art in terms of network overhead, but is also able to minimize the file synchronization time up to 12 times.