Every now and again, we hear a new technical term that requires a cursory Google search to make sure we are current with the times. Sometimes new terms are just recycling old concepts. Right now “Network Slicing” is en vogue, promising to help enable the evolution of modern networks.
At its core, it involves the ability to run multiple virtual networks across a shared physical infrastructure – essentially a separation of data plane and control plane. We have seen these before, for example in Software Defined Networking (SDN) and Network Functions Virtualization (NFV), both of which are inextricably linked to network slicing. Although often used concurrently in 5G discussions, network slicing is, in fact, an architecture paradigm that can exist outside of 5G and provide immense value for service providers in terms of efficient implementation of value-added services that can be monetized as revenue.
Dedicated Virtual Networks
Network slicing aims to isolate specific application traffic into a dedicated virtual network, whereby each slice carries specific application traffic such as IoT Telemetry or Automotive. Having an isolated virtual network enables different use cases to have unique network characteristics to a diverse end-user community. It also provides an opportunity to match allocated resources for the slice to expected usage patterns and specific value-points of the end-user services.
An example of a slicing application is telemetry sensors. Telemetry sensors that are required to send data every 12 hours may settle for high latency values, low bandwidth, and centralized compute services in the cloud. An industrial IoT which controls the manufacturing floor will require low latency and local compute with high-bandwidth at the far edge. In this case, building a network which is required to share both of the services will not be efficient and will create unbalanced costs. Imagine if you had to build an 8 lane high-way and let bicyclists ride it occasionally!
Cyclists vs. Autos
But with network slicing, the service provider can offer a different connectivity based on a dedicated slice, which ensures the service offerings do not overlap. In our example, that would be one road crafted only for cyclists and a different road for autos. Having a dedicated slice (e.g. road) can keep costs and expected revenues better aligned; for example, we assume in our analogy that a truck driver will pay more compared to a motorcyclist.
With respect to slicing implemented around the topic of network security in particular, service providers can offer a security posture that gives them the best chance to keep costs in check while keeping the network safe and affordable to operate.
For more information on this, come hear Radware’s Eyal Yaron speak at 5G World Congress in ExCeL, London on June 12. Details for Eyal’s panel can be found here.