Fonte de entropia baseada em RSSI bluetooth com validação estatística e integração via API REST

Abstract

Cryptography is one of the main mechanisms to ensure computational security today. For its proper functioning, it depends on the generation of random numbers, which in turn requires reliable sources of entropy. Although operating systems such as Linux already provide native entropy collection mechanisms, adding complementary sources can increase the robustness and resilience of the systems. In this context, this work proposes an alternative entropy source based on the variation of the Received Signal Strength Indicator (RSSI) of Bluetooth signals, using the Bluerandom approach as reference. The implementation was developed in Python, making the entropy data available through a REST API. Thus, simpler devices can obtain randomness through HTTPS requests. To evaluate the quality of the generated entropy, the ent, rngtest, and dieharder statistical test suites were applied. The results showed that, by using the collected entropy buffer as a seed for a PRNG, it was possible to generate sequences with high statistical quality, aligned with the standards required in modern cryptographic applications.