Polyglot Middleware for Heterogeneous IoT Clouds: Unifying Python, Rust, and Web Assembly

Abstract

Abstract:  Heterogeneity across languages and runtimes impedes large scale IoT deployments, where firmware (C), gateway scripts (Python), and cloud microservices (Kotlin, Go) fracture developer workflows. This paper proposes OmniBus, a polyglot middleware grounded in WebAssembly (Wasm) as a lingua franca bridging Python data science, Rust systems modules, and browser based dashboards. OmniBus offers a triple target toolchain converting Python NumPy kernels and Rust crates into Wasm binaries, enabling uniform streaming pipelines via a graph oriented orchestrator. Zero copy memory mapping coupled with shared nothing concurrency lowers cross language call latency to 24?µs on x86 64 and 57?µs on ARM64 edge routers. Case studies—industrial image inspection, predictive maintenance, and adaptive lighting—illustrate a 31?% reduction in development time and 2.8× throughput vs. gRPC based polyglot baselines. Our findings argue that Wasm’s sandboxed byte code and deterministic execution semantics unlock seamless composability across the fog to cloud continuum.

Keywords: Web Assembly, polyglot middleware, fog computing, Python, Rust