The Cost of Executing Business Processes on Next-Generation Blockchains: The Case of Algorand

TL;DR

This paper analyzes the cost structure of executing business processes on Algorand, a next-generation blockchain, highlighting its potential for lower costs and discussing the unique challenges and future research directions.

Contribution

It introduces a BPMN choreography compiler supporting Algorand, compares its costs with Ethereum and cloud computing, and discusses the implications of Algorand's novel cost model.

Findings

Algorand offers lower transaction costs for process execution.

The cost model of Algorand differs significantly from Ethereum.

Future research is needed to fully understand and optimize blockchain-based process costs.

Abstract

Process (or workflow) execution on blockchain suffers from limited scalability; specifically, costs in the form of transactions fees are a major limitation for employing traditional public blockchain platforms in practice. Research, so far, has mainly focused on exploring first (Bitcoin) and second-generation (e.g., Ethereum) blockchains for business process enactment. However, since then, novel blockchain systems have been introduced - aimed at tackling many of the problems of previous-generation blockchains. We study such a system, Algorand, from a process execution perspective. Algorand promises low transaction fees and fast finality. However, Algorand's cost structure differs greatly from previous generation blockchains, rendering earlier cost models for blockchain-based process execution non-applicable. We discuss and contrast Algorand's novel cost structure with Ethereum's well-known cost model. To study the impact for process execution, we present a compiler for BPMN Choreographies, with an intermediary layer, which can support multi-platform output, and provide a translation to TEAL contracts, the smart contract language of Algorand. We compare the cost of executing processes on Algorand to previous work as well as traditional cloud computing. In short: they allow vast cost benefits. However, we note a multitude of future research challenges that remain in investigating and comparing such results.