Enabling self-adaptability of small scale and large scale security systems using dynamic partial reconfiguration

Abstract

The application areas of field programmable gate arrays (FPGAs) are increasing due to its hardware acceleration and reprogrammable features. From large-scale computation systems like cloud, aerospace, and defence to small-scale computation systems like home automation and mobile phones, the dynamic partial reconfiguration property is found to be attractive to design adaptive systems for self-reconfiguration and self-healing. The article presents two self- adaptive security systems for small scale as well as for large-scale systems. The security system is designed to include encryption accelerators and hash code generation accelerators. The security system designed for small-scale systems saves space and power using hardware adaptation by loading or creating only the required accelerator during execution of the application. It uses light weight cryptographic algorithms. A second design for large-scale systems focuses on getting more throughput by allocating more resources to the required accelerator at runtime. Two designs are created for hardware adaptation based on the accelerator requirement at runtime. Proposed adaptive design for small-scale systems achieved 68.50% decrease in resource consumption and design for large-scale systems achieved 25.93% more throughput than the best existing implementations.