WebAssembly serverless functions in AWS Lambda | Cloud Native Computing Foundation

A recent survey by DataDog found that most AWS Lambda serverless functions are written in JavaScript and Python.

However, it is also well known that those high-level languages are very slow. In fact, according to a paper published on Science , up to 60,000 times slower than the same programs written in C or C++.

Hence, while JavaScript and Python are great for simple functions, they are not suitable for computational intensive tasks, such as image, video, audio, and natural language processing, which are increasingly common in modern applications.

WebAssembly, on the other hand, achieves performance on par with C/C++ compiled native binaries (NaCl) while still preserving portability, safety, and manageability associated with high-level language runtimes. WasmEdge is one of the fastest WebAssembly runtime on the market today.

Our vision for the future is to run WebAssembly as an alternative lightweight runtime side-by-side with Docker and microVMs in cloud-native infrastructure. WebAssembly offers much higher performance and consumes much fewer resources than Docker-like containers or microVMs. But for now, AWS Lambda and many other platforms only support running WebAssembly inside a microVM. Nonetheless, running WebAssembly functions inside a microVM still offers many advantages over running containerized NaCl programs.

For starters, WebAssembly provides fine-grained runtime isolation for individual functions. A microservice could have multiple functions and support services running inside a microVM. WebAssembly can make the microservice more secure and more stable .

Second, the WebAssembly bytecode is portable . Even inside a container, NaCl is still dependent on the underlying CPU, OS, and dynamic libraries installed on the OS. WebAssembly bytecode applications are platform-independent. Developers can write once and deploy in any cloud, any container, and any hardware platform.

Third, WebAssembly apps are easy to deploy and manage. They have much fewer platform dependencies and complexities compared with NaCl dynamic libraries and executables.

Finally, WebAssembly is polyglot. C/C++, Rust, Swift, Kotlin programs can all be easily compiled into WebAssembly. Even JavaScript is supported on WebAssembly. The WasmEdge Tensorflow API provides the most ergonomic way to execute TensorFlow models in the Rust programming language.