Automatic Discovery and Selection of Services in Multi-PaaS Environments

Automatic Discovery and Selection of Services in Multi-PaaS Environments

Abstract : Over the past couple of years, a new paradigm has emerged which is referred to as DevOps. It is a methodology to efficiently manage the relationship between development and operations in order to automate applications lifecycle. Spurred by its popularity, it is used today to manage applications in the PaaS level of the Cloud. However, it becomes very challenging when it comes to deploying an application in multi-PaaS environments. The first challenge is to discover and select services taking into account the application requirements and on the PaaS capabilities. Indeed, PaaS providers do not use the same mechanisms to describe and expose their services. Added to that, there is no standard way to describe application requirements. To tackle these anomalies, we propose an automatic and declarative approach to discover and select services offered by PaaS providers. It enables developers to express their requirements and PaaS providers to expose their offers in manifests. To do so, a matching algorithm selects the most appropriate offer in terms of PaaS capabilities to deploy the application. An offer may involve either a single or multi-PaaS provider(s). The key ingredients of our solution are threefold : (1) manifests to describe application requirements and the offers, (2) an ontology to remove semantic ambiguities in PaaS providers capabilities, and (3) a matching algorithm to elect the most appropriate offer to the application.The solution is proposed as a REST API and is delivered with aWeb client.

Keywords : Cloud computing, Multi-PaaS, PaaS service discovery, PaaS services selection, Semantic Web, Ontologies

Reference : Rami Sellami, Stéphane Mouton, "Automatic Discovery and Selection of Services in Multi-PaaS Environments", WEBIST 2018 - Proceedings of the 14th International Conference on Web Information Systems and Technologies, Seville, Spain, 18-20 September, 2018

Voir en ligne : scitepress