A model-based approach for planning blockchain service provisioning

Recently, the blockchain-as-a-service paradigm arose, and many works have evaluated the performance issues related to it. However, not as much has been done regarding Dependability attributes, which have ever been a crucial topic on service provisioning, let it be either public or private infrastructures. This paper presents the blockchain provisioning planning tool (BPPT), a framework to evaluate the availability, deployment, and maintenance costs of Hyperledger Fabric-based applications over private computational infrastructures. The BPPT uses continuous time markov chain (CTMC) and reliability block diagram (RBD) models as an evaluation method of Hyperledger Fabric's environments and determines distributed applications' deployment feasibility and endorsement policies related to the platform. We also present case studies that may help those interested in paradigm changes to decide whether they should migrate from old to new technology. Some of the obtained results pointed-out that the AND endorsement, which requires that all nodes sign the authenticity of a transaction, has the highest deployment and maintenance costs, as well as the lowest availability values due to operational requirements, already an OR endorsement, which needs that at least one available node signs the transaction, provides the best relationship between the evaluated metrics. The KooN endorsement (that requires that K out of N nodes signs a transaction authenticity) is a more general model that supports analyzing midterm configurations, besides the two extreme configurations, that is, to AND and OR arrangements.