EasiPCC: Popularity-aware Collaborative Caching for Web Requests in Low-Duty-Cycle Sensor Networks

The lightweight RESTful protocols, such as CoAP, have been proposed to provide web service on resource constrained devices. However, these devices normally operate in battery and work in sleeping mode most time to prolong lifetime. Consequently it is difficult to maintain persistent end-to-end connection and leads to the high access delay for web requests. In this paper, we propose a collaborative caching scheme to solve this problem. Initially we define sleep probability based on the feature of low-duty-cycle networks to decide what data to cache. Then we introduce the definition of node popularity, which focuses nodes that are on the request forwarding public paths and forward most web requests. A centralized algorithm is presented based on the theory of gravity center to find which nodes to cache. Finally, we propose a cache discovery method. Experiments show that our caching scheme significantly improves the performance of web data access in low-duty-cycle sensor networks.