Next-generation Web Applications with WebAssembly and TruffleWasm

TL;DR

This paper examines the architecture and performance of TruffleWasm, a WebAssembly implementation on GraalVM, highlighting its potential for efficient, polyglot web applications with near-native performance.

Contribution

It provides a detailed analysis of TruffleWasm's architecture and performance within the GraalVM ecosystem, advancing understanding of WebAssembly integration for web development.

Findings

TruffleWasm achieves near-native performance levels.

The architecture supports efficient polyglot web applications.

Performance tests demonstrate promising results in academic environments.

Abstract

In modern software development, the JavaScript ecosystem of various frameworks and libraries used to develop contemporary web applications presents many advantages. JavaScript is a widely known interpreted programming language, simple to learn and start development, and with numerous third-party libraries and extensions. However, with the rise of highly user-interactive websites and browser-based games, in some cases, JavaScripts executable engine could lack in performance. Therefore, developers could combine several other programming languages to create a polyglot user-interactive interoperable system to develop efficient modern web applications. The interoperability modules offer significant advantages but also present challenges in the execution due to high complexity and longer compilation times. This paper explores WebAssembly, a binary format compilation target with a low-level assembly-like language used for targeting from other programming languages. The binary format allows near-native performance level due to its compactness, as it prioritizes usage of low-level languages. Moreover, as a continuation of our previous research of the GraalVM ecosystem, we analyzed a guest language implementation of a WebAssembly based system, TruffleWasm, hosted on GraalVM and Truffle Java framework. This paper presents the architecture and review of the TruffleWasm within the GraalVM-based ecosystem as well as from performance test results within our academic environment.