Enabling the IoT paradigm through multi-tenancy supported by scalable data acquisition layer

Abstract

© 2016, Institut Mines-Télécom and Springer-Verlag France. The Internet of things (IoT) is composed of a large number of end systems. Data gathered from these end systems are processed in the cloud environment. This paper delineates multi-tenancy as a prerequisite for enabling cost effective IoT solutions using any cloud platform. In the IoT world, a large number of end systems imply a large number of users who should have open access to specific subsets of data. As multi-tenancy includes sharing the same application instance, security is a key requirement when engineering new software as a service (SaaS) applications. This paper describes multi-tenant implementation approaches and inspects security aspects and data isolation. It proposes a solution for data layer scalability in order to achieve the desired performances. Furthermore, it proposes an SOA solution to provide a secure multi-tenant web service with scalable data layers. We demonstrate that the proposed approach results in some performance loss, which is not acceptable for performance critical applications but is acceptable for services developed for tenants themselves. We experimentally confirm our data scalability theory based on an actor model and developed a cost estimator. Finally, we conclude that application level multi-tenancy is necessary from a cost efficiency point of view in a scenario involving a large number of users. The solution was tested in a private cloud environment.