Energy Consumption of TLS, Searchable Encryption and Fully Homomorphic Encryption

TL;DR

The paper investigates the environmental footprint of cryptographic PETs by measuring energy costs of TLS, Searchable Encryption, and Fully Homomorphic Encryption. It combines Life-Cycle Assessment with software-based energy measurements using $RAPL$-based tooling (CodeCarbon) to quantify PET overheads on a single hardware platform. Key findings show energy overheads from $2\times$ for TLS to $10\times$ for SE and up to $10^5\times$ for FHE, highlighting vast efficiency gaps between mature and emerging PETs. The work provides a simple, reproducible methodology and discusses implications for sustainable PET design, including hardware acceleration and decentralization as potential mitigations.

Abstract

Privacy-enhancing technologies (PETs) have attracted significant attention in response to privacy regulations, driving the development of applications that prioritize user data protection. At the same time, the information and communication technology (ICT) sector faces growing pressure to reduce its environmental footprint, particularly its energy consumption. While numerous studies have assessed the energy consumption of ICT applications, the environmental impact of cryptographic PETs remains largely unexplored. This work investigates this question by measuring the energy consumption increase induced by three PETs compared to their non-private counterparts: TLS, Searchable Encryption, and Fully Homomorphic Encryption (FHE). These technologies were chosen for two reasons. First, they cover different maturity levels -- from the widely deployed TLS protocol to the emerging FHE schemes -- allowing us to examine the influence of maturity on energy consumption. Second, they each have well-established applications in industry: web browsing, encrypted databases, and privacy-preserving machine learning. Our results reveal highly variable energy consumption increases, ranging from 2x for TLS to 10x for Searchable Encryption and 100,000x for FHE. Our experiments demonstrate a simple and reproducible methodology, based on existing open-source software, to quantify the energy costs of PETs. They also highlight the wide spectrum of energy demands across technologies, underscoring the importance of further research on sustainable PET design. Finally, we discuss orthogonal research directions, such as hardware acceleration, to outline promising directions toward sustainable PETs.

Figures (5)

• Figure 1: Average energy consumption of encrypted and plaintext ML inference using various classification models (100 samples, 30 features).
• Figure 2: Average energy consumption of encrypted and plaintext ML inference for varying number of features and samples.
• Figure 3: Energy consumption of encrypted and plaintext ML training of a logistic regression for varying numbers of features and samples
• Figure 4: Average energy consumption of encrypted and plaintext queries (SWiSSSE vs. Redis) for varying database sizes (1000 queries).
• Figure 5: Average energy consumption of HTTP and HTTPS requests on five different websites (1000 requests).