EdgeStore: Towards an Edge-Based Distributed Storage System for Emergency Response

With the pervasive influence of rich sensing technologies and internet connectivity, Social Sensing based Edge Computing (SSEC) is gradually becoming the forefront of context-aware sensing applications in smart cities. These applications generate huge volumes of data which require a sustainable storage platform. While storing sensing data into the cloud can be an immense benefit, cloud-based distributed storage requires constant internet connectivity, which deems the platform unusable in emergency situations with disrupted internet connectivity. This paper introduces the idea of utilizing privately owned edge devices for a decentralized distributed data storage during emergency situations. However, persuading the edge devices to cooperate and participate in sharing storage brings about an unprecedented set of challenges. The first challenge lies in the inconsistent availability of the storage system as the edge devices may not be inclined to share storage resource. The second challenge lies in configuring all the edge devices to integrate with the storage system in a scalable manner. In this paper, we present EdgeStore, an edge-based distributed storage system that attempts to address the above challenges by introducing a game-theoretic resource incentivizing framework. We implement a prototype of EdgeStore on top of the Linux Ext4 file system using the FUSE interface. The results on two real world case studies demonstrate that EdgeStore provides significant performance improvements over the state-of-the-art distributed storage systems in terms of throughput, energy consumption and latency.