Abstract

The cloud computing market has been in the center of attention for years where cloud providers strive to survive by either competition or cooperation. Some cloud providers choose to compete in the market that is dominated by few large providers and try to maximize their profit without sacrificing the service quality which leads to higher user ratings. Many research proposals tried to contribute to the cloud market competition. However, the majority of these proposals focus only on pricing mechanisms, neglecting thus the cloud service quality and users satisfaction. Meanwhile, cloud providers intend to form cloud federations to enhance their services quality and revenues. Nevertheless, traditional centralized cloud federations have strict challenges that might hinder the members' motivation to participate in, such as formation of stable coalitions with long-term commitments, participants' trustworthiness, shared revenue, and security of the managed data and services. For a stable and trustworthy federation, it is vital to avoid blind-trust on the claimed SLA guarantees from the members and monitor the quality of service considering the various characteristics of cloud services. This thesis aims to tackle the issues of cloud computing market from the two perspectives of competition and cooperation by: 1) modeling and solving the conflicting situation of revenue, user ratings and service quality, to improve the providers position in the market and increase the future users' demand; 2) proposing a user-centric game theoretical framework to allow the new and smaller cloud providers to have a share in the market and increase users satisfaction through providing high quality and added-value services; 3) motivating the cloud providers to adopt a coopetition behavior through a novel, fully distributed blockchain-based federation's structure that enables them to trade their computing resources through smart contracts; 4) introducing a new role of oracle as a verifier agent to monitor the quality of service and report to the smart contract agents deployed on the blockchain while optimizing the cost of using oracles; and 5) developing a Bayesian bandit learning oracles reliability mechanism to select the oracles smartly and optimize the cost and reliability of utilized oracles. All of the contributions are validated by simulations and implementations using real-world data.