These days it seems that every time we turn the news on, or look at a paper we hear something about 5G. And as soon as you look at industry specific news, journals or webinars about 5G you see network slicing being mentioned as a key component of making 5G able to support ultra-reliable low latency services (uRLLC).
But is network slicing so new?
In my opinion network slicing is the logical extension and consolidation of a number of current and new technologies. So in the case of network slicing, as King Solomon said:
“There is nothing new under the sun”
Good old VPNs
With Virtual Private Networks we have already had a basic building block for network slicing for nearly 20 years. When Microsoft’s Gurdeep Singh-Pall invented PPTP (or Point-to-Point Tunneling Protocol) in 1996, he did this by creating a virtual private network to allow users to have a secure Internet connection and work effectively and securely from home.
The main issue with using VPNs for network slicing is that it is very difficult to support deterministic services that require guarantees on delay, delay variation, packet loss, reliability etc. To start to address this issue, some basic quality of services tools were added, for example expedited forwarding (EF) and assured forwarding (AF) were added to Class of Service (CoS) queues to determine the order in which packets would be transported over the network. However, this is not enough for some business critical services, and fails to meet the needs of nearly all mission critical services.
So businesses and strategic industries with business critical and mission critical services requiring this level of determinism are either forced to build their own private networks OR request that the service provider build them a private network specifically for them. This is of course extremely expensive.
New technologies to the rescue
This is where we see a number of new technologies coming to the rescue.
Today’s networks cannot achieve this flexibility because they have extremely static architectures, with compute and storage at fixed points in the network, normally at the Service Edge or in the Core. However, with NFV compute and storage can be dynamically moved around the network to where and when it is required
So what network slicing does, is to use these new technologies to take VPNs to another level.
Network slicing allows virtual networks (slices) to be created across the network. Each slice is optimized to deliver a specific set of services, and may support one or many different users each with their own VPNs. So for example you can have a slice optimized for low latency services, a slice for high-bandwidth services, and so on.
So whilst the basic concept of network slicing is nothing new i.e. setting-up virtual networks to deliver services across a shared infrastructure. I do believe the amalgamation of virtual private networking with a number of new state of the art technologies and approaches does bring something revolutionary. With network slicing network, operators are able to create a dynamic, deterministic and assured network from a single set of network resources (physical and virtual). These resources can be programmed in an intelligent, autonomous way, to support services with completely different operational parameters and policies.
Given this, I anticipate that the reach of network slicing will go far beyond 5G new radio. With network slicing network operators finally have a set of tools that allows them to build a single transport infrastructure to cost efficiently enter new markets whilst addressing the current and evolving needs of all their current customers.
To learn more about ECI’s 5G Solutions, click here.