Enhancing service-level agreements using decentralized auctions and witnesses

Industrial applications (e.g., remote live event broadcasting and car-sharing) require high-quality cloud services to deliver business value. These cloud applications often rely on resources and services from different providers due to the distributed geographic location of data sources and diverse access policies. The generation and enforcement of the Service-Level Agreements (SLAs), in this case, are crucial for guaranteeing the quality of both application services and underlying cloud infrastructure. Traditional service allocation methods through auctions usually involve a centralized auctioneer to coordinate the auction procedure, which is expensive and suffers from a single point of failure. Recently, blockchain has emerged as a decentralized platform to support trustworthy online transactions in various scenarios. It leverages consensus mechanisms to agree on new data and cryptographic technologies to guarantee data integrity and immutability. Blockchain can provide a decentralized mechanism for service auctions and SLA automation; however, it is still challenging to select the most suitable provider/customer pairs and detect service violations in SLAs. To tackle these challenges, we propose a novel Auction and Witness Enhanced trustworthy SLA for Open, decentralized service MarkEtplaces (AWESOME) framework in this thesis. We aim to provide an efficient and trustworthy environment for generating SLAs and trading services. Especially, our integrated model includes a novel witness mechanism and supports interactions between untrusted service providers, customers, and witnesses to complete decentralized auctions and SLA enforcement on the blockchain. We present the cloud service as an example, and the proposed model can be easily extended to other service transaction scenarios.