FAML: Fog Descriptor Language for Fog Service Development and Deployments

Abstract

Cloud-centric services are experiencing major challenges due to the tsunami of computations and data streaming. Some of these challenges are bandwidth, latency, uninterrupted requests, continuous data streaming by the end devices, and computation dependability. Fog computing is a new computation paradigm that attempts to address delay and latency challenges in cloud services, mainly by offloading computations to computing nodes near the consumers. One of the main challenges to this computation model adoption is the lack of standardized software development kits (SDKs), tools, and run-time environment. This paper introduces a new descriptor and annotation language that can be used to develop an offloadable fog service from an already existing cloud service. FAML is an abbreviation for Fog Annotation and Meta Language; a language that can beused to develop a fog service. The FAML compiler has been implemented to work with Golang cloud service implementations. The generated fog services have been offloaded into fog nodes in the same local network of the service consumer as part of the experiment's evaluation. The paper presents the usage of the proposed language in three different service development use cases: image enhancement service, weather forecast service, and bulk update service.