Estimating transaction cost for cloud-based private ethereum blockchains

Blockchain technology is increasingly being used by several companies in the most varied sectors of the economy. The possibility of having decentralized applications (DApps) allows for the emergence of technological innovations such as cryptocurrencies and decentralized asset tracking applications. Many of these DApps are deployed in the cloud with Infrastructure as a Service (IaaS) payment model, in which the payment is made according to the use of the service. However, it is not simple to estimate the cloud infrastructure costs that a DApp will consume. Furthermore, correctly estimating infrastructure costs is essential to analyze the viability and develop business models for enterprise DApps. This work presents an experience report on estimating the cloud infrastructure cost for an enterprise DApp. To do that, we deployed a private Ethereum DApp, using Proof-of-Authority consensus algorithm, with several different configurations of Amazon Web Services (AWS) EC2 instances and blockchain parameters. We benchmark the transaction processing capacity, CPU and disk usage in each configuration, estimating their maximum capacity and costs. We shared our methodology to measure and estimate those costs and our insights on best configuration practices for reducing costs of deploying enterprise DApps in the cloud.