Enforcing trustworthy cloud SLA with witnesses: A game theory-based model using smart contracts

Area of Science:

• Computer Science
• Distributed Systems
• Blockchain Technology

Background:

• Traditional cloud Service Level Agreements (SLAs) suffer from trust issues between customers and providers.
• Verifying off-chain Service Level Objective (SLO) violations before on-chain recording remains a challenge in current blockchain explorations.

Purpose of the Study:

• To propose a novel witness model using smart contracts to address the trust deficit in cloud SLA enforcement.
• To ensure the truthful reporting of SLO violations through a game-theoretic incentive mechanism.

Main Methods:

• Implementation of a "Witness" role with a carefully designed payoff function incentivizing honest reporting.
• Analysis and proof of witness honesty using the Nash Equilibrium principle from game theory.
• Development of an unbiased selection algorithm for random and independent witness selection to prevent collusion.
• Introduction of an auditing mechanism with quantitative indicators to detect malicious witness behaviors.

Main Results:

• The proposed witness model effectively solves the trust issue in cloud SLA enforcement.
• The game-theoretic design ensures witnesses are incentivized to report truthfully.
• The unbiased selection and auditing mechanisms successfully prevent collusion and detect malicious witnesses.
• Experimental studies on the Ethereum blockchain confirm the model's feasibility and performance, aligning with design expectations regarding transaction fees.

Conclusions:

• The blockchain-based witness model provides a feasible and trustworthy solution for enforcing cloud SLAs.
• Incentive mechanisms and robust auditing are crucial for reliable decentralized systems.
• This approach enhances transparency and accountability in cloud computing environments.