Serverless Software Engineering – and How to Get There.

Abstract

AbstractServerless computing is on the rise but developing software to exploit this space involves a deep rethink of software architecture, deployment, and operation (perhaps also, software development processes and team structures). Central to this revolution, we find a compelling argument for distributed, services-based software architectures. But converting a large, established monolith architecture system to microservices is non-trivial and fraught with both cost and risk. For the many firms with established software systems, this architectural system conversion might be considered the first stop-off on the journey to serverless computing. In tandem, software deployment and production monitoring also require reinvention. The focus of this paper involves an examination of the advantages of microservices architectures, include techniques for migrating from monolith architectures. Through application of a Multivocal Literature Review (MLR), we find that migrating from a monolith architecture to a microservices architecture is risky and non-trivial, but that there are techniques that can be employed to support the transition. We find also that monoliths have their advantages which might be overlooked to some extent in the race to serverless computing.KeywordsServerless software engineeringMonolithMicroservicesMigration