Agent-Based Fog Computing: Gossiping, Reasoning, and Bargaining
Kwang Mong Sim · Letters of the IEEE Computer Society · 2018
Fog computingis a layered model that consists offog nodesresiding betweencloud data centersandInternet of Things(IoT) devices. With a large number of highly distributedfog nodesandIoTdevices,decentralized and computationally efficient techniquesforresource discovery and selectionare necessities of fog computing. Since consumers are charged for using fog and cloud resources,economically efficient resource pricing mechanismsare crucial. This work devises: 1) anagent-based fog computing model, 2) agossip-based resource discovery algorithmfor propagating resource requests, 3) areasoning techniquefor determining similarities between user requirements and resource profiles, and 4) abargaining mechanismfor dynamically pricing vertically arranged fog and cloud resources. Mathematical proofs validate that the gossip algorithm and reasoning technique arecomputationally efficient, and the bargaining mechanism iseconomically efficient. Since agents can determine resource similaritiesindependentlyand propagate requestsautonomouslywithout centralized control,agent-based resource discovery and selection approaches are decentralized techniques.