Evaluation of Rust and Java gRPC Services for Automatic Speech Recognition

For Automatic Speech Recognition (ASR) systems to effectively translate audio to text, high-performance and low-latency backend services are required. The performance of gRPC services built in Java and Rust, two well-liked languages with different advantages, is compared in this study. Garbage collection causes possible latency problems for Java, a programming language well-known for its simplicity of use and maturity, while Rust provides memory safety without compromising speed. We test both solutions for different loads and durations in order to undertake a set of performance evaluations. The findings show that Rust exhibits more consistent performance over longer times, but Java performs better in short-duration testing. These results offer vital information for creating high-performing ASR systems.