Impact of SRAM architectures - implementing ChaCha20 encryption algorithm using In-memory compute design choices on sustainability

Abstract

As our dependency on technology increases, the use of electronic devices also rises, leading to a greater environmental impact that must be monitored. In today’s world, where data security is critical, especially in IoT devices, it is essential to implement encryption algorithms that ensure security [1] and support sustainability. This work introduces two key metrics to evaluate sustainability: embodied carbon footprint and operational carbon footprint. Together, these metrics offer a comprehensive way to assess the environmental impact of design decisions. This study analyzes three different SRAM architectures ( 10T [2], 9T, and 6T ) implementing the ChaCha20 encryption algorithm. The implementation based on 10T bitcell [2] is extended for low power and lesser area implementations.This work proposes two new implementations based on 9T and 6T bitcells which is influenced from the 10T implementation. The evaluation includes traditional metrics such as power, performance, area, and newly introduced sustain- ability metrics. The security of each design against side-channel attacks is assessed using Welch’s t-test. Among the three implementations, the 6T bitcell design shows the highest operational footprint due to its longer execution time and the lowest embodied footprint due to its smaller area. While analyzing embodied and operational footprints separately is important, combining both is essential for making informed and sustainable design choices. The design choice should be sustainable and maintain strong protection against side-channel threats.