Performance Evaluation of Function Composition in Middlewares supporting FaaS for Serverless computing

The concept of Serverless Computing is a new and exciting aspect of cloud computing that involves the deployment of small pieces of software applications and services as serverless functions. Serverless computing architecture enables the cloud provider to fully manage the execution of a server's code, eliminating the need for customers to develop and deploy the traditional underlying infrastructure required for running applications and programs. Even though big tech companies are extensively utilizing serverless computing in their products and investing billions on this novel but affirmed technology, it is affected by various problems still considered an open field in research. In fact, by definition, FaaS architectures are geographically dislocated and consequently subject to event propagation delays that can significantly degrade the overall system performance. What is generally done, is to reduce as much as possible cumulative delays especially if attributable to the infrastructure itself that could determine a greater or lesser competitiveness on the market. The background idea, which becomes the leit motiv throughout this work, is to develop and assess the performance, and thus the validity, of a Message-Oriented Middleware-centric serverless platform architecture promising to enable advanced analytics capabilities and better overall performance, without renouncing the essential characteristic of scalability in the context of distributed systems. Experiments in emulated conditions show that applying the MOM coordination co-locality principle improves the end-to-end delay and data processing performance.