Stable federated fog formation: An evolutionary game theoretical approach

• Post by: Ahmad Hammoud
• 26 June 2021
• Comments off

Abstract

Instability within fog federations is considered a serious problem that degrades the performance of the provided services. The latter may affect the service availability due to fog providers withdrawing their resources. It may either lead to failures for some users’ invocations, or to an increase in the number of tasks inside the servers’ processing queue. Such a critical problem strips the fog paradigm from its main characteristic, the low latency factor. As far as we are aware, no work in the literature has addressed the problem of encountering unstable fog federations. Their main concerns were increasing the providers’ payoff regardless of their behavior. To address the aforementioned limitation, this work studies the stability of the federations through modeling the problem as an evolutionary game-theoretical model. Moreover, it devises a decentralized algorithm that implants the Replicator Dynamics model within which expresses the evolutionary dynamics. Experiments are conducted using EUA Datasets to simulate our algorithm and to show that it leads to an evolutionarily stable strategy over time, which stabilizes the federations and improves the Quality-of-Service for the users.

Authors’ Notes

The contributions of this paper are:

• Adopting an evolutionary game mechanism that simulates the dynamicity of the fog providers, in terms of rational and irrational decision making. To the best of our knowledge, no previous work has ever addressed the dynamic strategies that encounter such a paradigm.
• Forming the initial set of fog federations using the k-means clustering technique. Such a technique allows forming federations based on providers’ similarities. In our algorithm, we use the location of the providers to join neighboring fog providers altogether.
• Advancing a latency-aware greedy approach for placing services on the available fog nodes within the federation.
• Devising a decentralized algorithm for fog providers that leads to stabilizing the federations through reaching the evolutionarily stable strategy.