Application Based Caching in Fog Computing to Improve Quality of Service

Abstract

Fog Computing is an emergent network paradigm that arises as a response to the prevalence of Internet of Things (IoT). By the use of fog computing, cloud is extended close to end users to reduce latency, traffic load, and needed bandwidth. Efficient data caching presents the core of fog computing. Moreover, low quality caching techniques may represent an additional burden on network resources in case of high miss ratio. As an emergent paradigm, fog computing raises the demand on efficient caching techniques, these techniques must be compatible with IoT and the wide variety of its applications. In this paper, a new caching approach is proposed, referred to as Application Based Caching for Fog computing, abbreviated as ABCFOG. The proposed approach considers the type of application as the main caching prediction criteria. ABCFOG has been tested under various case studies including three types of applications. It is discussed in details before it has been evaluated by simulation using NS-2 Network Simulator. Three evaluation parameters are measured, hit ratio, response time and bandwidth. Results show that ABCFOG has improved caching with at least 30% in response time and hit ratio. However, an additional cost of bandwidth is needed for such improvement.