Efficient service auto-discovery for next generation network slicing architecture

Abstract

The evolution towards the Fifth Generation (5G) mobile networks will be characterized by ubiquitous connectivity, superior user experience, stringent reliability requirements and efficient management of a massive number of devices. These characteristics will lead to a growth of diverse services and business models, which will increase the service complexity. In order to meet the needs of future networks, cloud and virtualized technologies will be used. Network slicing is among the emerging virtualization concepts: the network resources are logically partitioned in sets, called network slices, which are properly selected in order to meet and support the communication service requirements for specific use cases. In a service centric 5G network, the foremost requirements are (a) distribution of offered services, and (b) flexible and automated discovery of a service by users. In this paper, we propose an efficient dynamic service discovery framework in the context of network slicing, based on the recently introduced concept of CloudCasting protocol. The proposed solution provides several benefits as service isolation, enhanced scalability, prompt reaction to mobility, natural accommodation of heterogeneous connectivity, and lightweight signaling to establish service discovery and distribution.