Modern software systems' business requirements are pushing systems architecture towards the cloud. It is not uncommon to do so by extending applications' functionality to include a web-API, and then deploy the server on a virtual machine (running server editions of e.g Ubuntu or Windows). Then a "frontend" program is written, often in a different programming language, to call the application remotely. However, cloud native applications can make radically more efficient choices in all regards.
The topic of this thesis, is to investigate advanced options for cloud native web application developers. And could include
- Investigating languages like Scala or OCaml, that compile to java-script, such that the front- and backend programs can be written in the same language.
- Investigate options for writing backend programs as operating systems modules, such as MirageOS, where the compiler may specialise the virtual machine to be a runtime for the applications' backend.
- Design a measurement method for comparing historical ways of doing remote procedure calls (REST, gRPC, GraphQL, etc.). And use the comparison to try and design novel ways of calling procedures remotely.
In either case, leaning from an investigation is all about demonstrating differences and evaluating them. As a vessel for demonstration, we have bought `academemes.com`, which will host a social network platform for sharing academic memes that relate to academic papers. The student is free to suggest features for the page, that demonstrate their findings.