Energy and Performance Comparison of Cryptocurrency Mining for Embedded Devices

Abstract

Blockchains are composed of a network of computing entities which maintain an immutable distributed ledger thus facilitating auditability and accountability of transactions. The process of Cryptocurrency mining involves miners verifying transactions and adding them to the ledger. While Cryp-tocurrency mining prevents random nodes from creating and appending blocks, it incurs a negative impact on the embedded devices due to their resource-constrained nature. Thus, a need for profiling the power consumption of cryptocurrency mining platforms for embedded devices becomes necessary. In this work, we comparatively study the performance and power consumption of a suite of commonly used cryptocurrency mining platforms and analyze energy-performance trade-offs. In particular, our analysis is based on the power consumption during mining, changes in the power behavior and their impact on the overall system. Experiments conducted using real devices indicate that mining incurs a 2-fold increase in power consumption compared to those without mining. Our insights can be used to further investigate the impact of cryptocurrency malware on power consumption as well as design algorithms for energy efficient mining.