Wasp: Pushing Serverless Computing to the Limit
Current serverless and micro-service architectures rely on existing system software to preserve compatibility and minimize development effort. However, they also inherit the latencies and overheads that existing systems carry with them. Existing serverless architectures resort to process-level or language-based abstractions in order to achieve the latencies that a real-time, micro-service oriented application requires. While these abstractions offer certain levels of isolation, true full-system isolation is sometimes required.
Until now, achieving full-system isolation necessitated the use of heavy-weight virtual machines, raising the lower bound on response latencies. In this project, we aim to demonstrate that the lower bound on latencies in a fully virtualized environment is much lower than existing technology allows. We leverage this opportunity to eschew traditional OS interfaces and demonstrate a custom, microservice platform called Wasp.
- Coordinated Resource Management for Cloud-based Specialized Operating Systems
- Nautilus Aerokernel
- Towards a Practical Ecosystem of Specialized OS Kernels
- Multiverse: Easy Conversion of Runtime Systems Into OS Kernels via Automatic Hybridization
- Enabling Hybrid Parallel Runtimes Through Kernel and Virtualization Support