Network Slicing: cutting the same net into several slices
One of the strengths that is underlined by 5G mobile networks is the ability to provide broadband and low latency connections, even when you have a large number of objects connected to the network. Each use case has its own requirements, and many of them can be orthogonal in terms of quality, allowing efficiency in infrastructure sharing scenarios.
For example, when considering an Autonomous Vehicle scenario, which requires low latency and high reliability for safety as a Virtual Reality use case, it requires a high throughput with relaxed reliability since losing a few pixels, frames, or lower resolution may be affordable. Those two quality parameters could be granted, at the same time, by the same physical infrastructure. On top of the same hardware a network slice can be built to support the Vehicle Network, with specific KPIs in terms of latency, and another network slice to provide large throughput for VR applications.
From an implementation perspective, Network Slicing will be possible thanks to the extensive use of the Software Defined Networking (SDN) paradigm, to control the many devices involved along with Network Functions Virtualization (NFV) to build logically dedicated functions for each of the slices.
As an example, Mobile network operators, can divide their network into smaller virtual subnetworks and connect them to each other. Since each partitioned virtual network provides independent network functions, the services and functions can be adapted to a customer needs. Slicing can be used to virtually partition the wireless access network, as well as the core components of the Evolved Packet Core (EPC) or the data centers. Characteristics of network slicing are flexibility, common infrastructure, isolation and a dedicated network.