Aspects of event-driven cloud-native application development

Abstract

Managing and configuring servers have been challenging burdens. Therefore, serverless computing has been introduced to overcome the complexities of managing servers, get rid of operations and handle different terms such as scalability and high availability. Codes or binaries in serverless computing are executed upon direct invocation or as a response to events in a highly scalable manner, which makes it most appropriate for building event-driven applications.An example of such serverless computing services provider, is OpenWhisk the open-sourced project by IBM. To receive events and react to them by invoking or running OpenWhisk actions (codes or Docker containers), OpenWhisk provides an ecosystem of packages of services to enable, facilitate, ease the usage of the services and subscribe to their events. While OpenWhisk provides different powerful means and tools to interact with events, it lacks a number of important services (event sources) packages within the OpenWhisk ecosystem which are necessary to ease and facilitate subscribing to receive their events and using the different functionalities of the services. This thesis enriches the OpenWhisk ecosystem by integrating and enabling more services. A use-case-based approach is chosen to select services to be integrated and enabled. The proposed use-case is an Early Warning System (EWS) used to warn the public for disasters, possible incidents and helping rescue survivors. In this approach, diversity of the integrated services in terms of domains and vendors are guaranteed to avoid vendor lock-in and provide flexibility in the available services. The integrated services were then categorized based on taxonomic categories using the domain of services to ease out finding and organizing packages.