Trading off Latency Against Security in Open Energy Metering Infrastructures

Abstract

Embedded devices are expected to transform the landscape of networked services in many domains, among them smart homes and smart grid systems. The reliable and optimised operation of smart grids is dependent on reliable data provided by end nodes (e.g. smart meters), and assurance of secure communication across networks. Understanding whether advanced security building blocks have a role to play in forthcoming infrastructures needs a basic understanding of each potential building block with respect to resource usage and impact on timing. In this paper we study the performance penalty of asymmetric cryptography techniques used for protection of wirelessly transmitted data in a prototype smart metering system. The prototype system is built using hardware and software components from "Open Energy Monitor" project using a wireless data link between the metering device and the data collector device. We investigate the use of the Elliptic Curve Integrated Encryption Scheme (ECIES) in two versions -with standard building blocks and with added Elliptic Curve Digital Signature Algorithm (ECDSA) support. The use of the ECDSA allows the system to achieve the non-repudiation property. We compare those cryptographic techniques with the Advanced Encryption Standard in Galois Counter Mode (AES-GCM) technique in two versions -with 128 bit and 256 bit keys. Performance is compared in terms of execution time of (1) preparing data, (2) unpacking it, and (3) roundtrip time. We then discuss the implications of the measurements, where the roundtrip time of sending one measurement ranges from 378 ms in case of AES128-GCM to 16.3 sec using ECIES with ECDSA.Smart meter infrastructure security, Elliptic Curve Cryptography, Performance and latency trade-off