Decision support for migrating application's functionality to FaaS

Abstract

We live in the era of cloud computing. Many companies move their legacy applications to cloud environments for the sake of improving their business and achieving more profitable success. Choosing the cloud service model is one essential step in the process of transition applications to a cloud environment due to the importance of its contribution to optimizing the applications performance and costs.

The variety of cloud service models bring along different aspects in terms of both user responsibilities and cost efficiency. Function as a Service (FaaS) is one of those models that looks appealing and convenient for users. This model provides users the ability to move parts of their applications to the cloud environment. The FaaS environment is responsible for managing, initializing and scaling the functions. Furthermore, the user is billed based on actual usage of the infrastructure. However, the interesting question is how one can choose the functions that fit the FaaS model, and what are the factors that could influence the selecting process.

In this work, we introduce a new approach to help developers and software architects in assessing the suitability of application functions for running in FaaS environment. This approach considers the factors that might influence the appropriateness of function for being deployed in FaaS model and provides users with a score representing the suitability of function for FaaS environment. We validate our approach by means of a prototypical implementation. This new approach does not consider the security concerns in the assessment process. But, this work could be considered as a starting point for automating the FaaS functions selection process.