Availability Modeling for Blockchain Provisioning in Private Clouds
J Dantas, P Silva, L Fiondella, C Melo, P Maciel
TL;DR
This paper tackles the dependability of private blockchain deployments by developing a two-stage availability analysis for Hyperledger Fabric in private clouds. It introduces a hierarchical method combining a Reliability Block Diagram and a Continuous-Time Markov Chain to capture hardware, container, and endorsement dynamics, and derives availability expressions including $A_{Server}$, $A_{KooN}$, and $A_{NoN}$. The study demonstrates feasibility through a case study with ten scenarios, showing how endorsement policies influence general availability and annual downtime. The approach provides a practical framework for stakeholders deciding on migration or deployment strategies in private infrastructures, though it lacks experimental validation.
Abstract
Blockchain technology has emerged, and many previous studies have assessed its performance issues. However, less attention has been paid to the dependability attributes, which have been a critical topic in service provisioning, considering public or private infrastructures. This paper introduces analytical models to assess the availability of private blockchain infrastructure for Hyperledger Fabric-based applications. Furthermore, a case study will be presented to demonstrate the feasibility of the proposed model, which may assist stakeholders in deciding whether to migrate from old to new technology. Some of the obtained results indicate that, unlike most conventional systems, general availability may decrease as new nodes are added to the environment. This phenomenon occurs due to the adopted endorsement policy, which determines the proportion of required nodes to sign the authenticity of a transaction.