Reliable 5G RAN Slicing: Bridging the Gap between Theory and Practice - RMGS

Description

An important challenge faced by Cloud and internet service providers (CSPs and ISPs) is how to provide efficiently and reliably customers with networking solutions that support their pre-specified quality of service (QoS) requirements, and this requires the reliable and cost-effective provision of virtual networks to individual corporate customers called slices. Advances in 5G technology have led to the introduction of radio access network (RAN) slicing that enables the dynamic allocation of resources that include network resources, as well as storage and computation resources, to individual customers in a way that meets their individual QoS requirements based on their individual applications and types of traffic. RAN slicing in 5G supports these diverse services and facilitates the efficient reassignment of resources from one virtual RAN (vRAN), or slice, to another. RAN slicing is an important part of end-to-end network slicing which includes the wireline (optical) network in addition to the RAN (wireless) and provides end-to-end networking solutions. RAN slicing has been considered an important research topic in industry and academia and there are many publications on this topic.However, despite the wealth of knowledge and solutions that have been generated, bridging the gap between the published literature and industry impact is still a challenge. Many existing publications provide orchestration solutions and algorithms that address problems related to specific models and related theories, while the industry finds it difficult to apply these solutions in their networks. The important aspects of resilience and reliability which are also key customer requirements did not receive sufficient attention. Consideration of these aspects requires the allocation of additional resources to slices, such that if failures occur, the adverse effect on the user’s QoS is mitigated or even avoided. There are many publications that provide optimal solutions for RAN slicing where the aim is to minimize cost subject to meeting QoS requirements without allowing for redundancy of resources to be used if a failure occurs, or if a disaster occurs that leads to simultaneous multiple failures. A key contribution of this proposed project is a novel algorithm that will lead to the optimal solution for RAN provisioning and slicing at minimal cost meeting QoS requirements, but the difference with existing work is that the QoS requirements will include network resilience requirements. In other words, we will provide solutions for efficient allocation of resources, so that in case of a failure, in most cases, the users’ connections are automatically restored using the redundant resources, so they remain intact and no interruptions are caused to the users.Another important practical feature of RAN slicing is the consideration of multiple layers as the internet operates in multiple layers, and considering the complexity and heterogeneity of networking demands of future services, this is likely to continue. Other practical challenges faced by the industry that will be addressed in this project are the simplicity of the solutions and ease of deployment, scalability, and running time that are compatible with their customers’ QoSrequirements. Constant interaction between researchers and industry is required to better understand the real-world needs to develop solutions that are compatible with these needs.