Interview with STL partners

Zeetta Networks Q+A: network slicing and splicing with private 5G networks

Article from Matt Bamforth first published on STL Partners’  website on 4th January 2021.

Zeetta Networks is providing innovative capabilities in the provision of integrated 5G/Wi-Fi network management solutions, in their active work in the 5G-ENCODE project in manufacturing. We spoke to their CEO and Co-Founder Vassilis Seferidis about the work that Zeetta Networks is doing on network slicing and splicing in the context of enterprise private 5G networks.

How would you define network splicing and slicing and what has changed that enables you to do this?

The technology of network slicing and splicing is very much related to network virtualisation and software defined networking. Network slicing is a concept that has been around since the inception of 5G. If you think more broadly, beyond the 3GPP definition, network slicing is all about taking a physical network and splitting it into different logical sub networks to satisfy bespoke networking demands of groups of people, devices or applications. The primary is that you can support many different use cases, each one with its own specific performance requirements. This is key for 5G networks that need to support a wide range of applications from low latency industrial controls to mobile broadband and high-density IoT deployments.

Network splicing can be seen as the opposite of that, where two or more physical or virtual sub-networks are merged to construct the ‘aggregate topology’, i.e. a new virtual network that combines the network nodes and the services of all the underlying sub-networks. For example, this could include the sub network of a private LTE/5G mobile deployment and the sub network of Wi-Fi in a factory, and to be able to see them in a single pane of glass.

Network splicing can be also seen as the bundling of different network services together.

So, if you are a customer and you want to receive two or three network services, then this can be achieved through ‘splicing’ the individual services into a new bundle which we call “network mode”.

This is a very useful function for two main reasons: First, it provides a single-pane-of-glass of the entire network so it is easier to see the interdependency between different parts of the network and how they affect one another. Second, it simplifies the provision and management of complex, end-to-end services across the whole network.

We combine a much wider variety of services (e.g. low-latency controls, IoT deployments, AR/VR applications, etc) which are delivered as a spliced offering. Each one of those services is a network slice for me, because you are delivering a specific network of ability to a user, which you can also bundle together for a combined, ‘spliced’ offering.

What does hybrid networking mean to you and why are these network splicing and slicing capabilities relevant to today’s hybrid networking world?

One of the key benefits of NFV/SDN is the ability to support technology- and vendor-agnostic solutions because they separate the software from the hardware. This means that regardless of the particularities of the underlying hardware, I can still abstract any network element (physical or virtual) and represent it in my computer as a network model. That’s exactly the value proposition here, because we now have a way to abstract the physical complexity, and to support different vendors and technologies such as LTE/5G and Wi-Fi, Bluetooth, ZigBee under a single orchestration platform.

The other important result is that we can now easily combine different networks domains such as networks that are private as well as public, including public and private cloud. In both cases, we are talking about extending the control and management of different network administrative domains, under a single orchestration software layer.

The idea here is that network orchestration and automation can be achieved using network slicing and splicing whereby we define separate network slices in each administrative domain and then we join them together with network splicing to allow the quality of service (QoS) to continue uninterrupted as we move from one domain to the other.

What work is Zeetta Networks currently doing around network splicing and network slicing for the 5G-ENCODE project?

Our NetOS^® software is an orchestration software that allows you to do slicing and splicing in a single domain, for example, if you take a private mobile network you are able to define network slices to support particular applications for example, a network that allows you to control a machinery in an industrial setting within strict latency requirements. Our NetOS^® platform currently operates in a single domain, and supports our three products: Visualise, Optimise and Automate that simplify the orchestration of network services in a single network domain.

What we are working on for the 5G-ENCODE project is to extend this capability in a multi-domain environment and more specifically in extending the orchestration of services across private and public 5G networks. This is important where you have many private or public 5G networks next to each other and you want to, define and manage services across the different (public and private) domains. So, we are developing what we call ‘Multi -Domain Orchestrator’ (MDO), which is basically extending the capability of slicing and splicing into a multi-domain environment. Our MDO software allows you to splice different network slices from different domains into a unified, larger network slice, to enable the seamless delivery of services across the domains. In the 5G- ENCODE project we demonstrate this functionality in a demanding industrial setting.

Who are the different enterprise stakeholders that would benefit from this solution?

The main benefit that network splicing brings to the enterprise is that you have a much easier way to monitor the network through a single-pane-of-glass view of the entire network. Moreover, the fact that you can now use a very simple graphical user interface to provision services over the aggregate topology, hugely simplifies network operations. This means that the enterprise is able to control and manage their network, without requiring expert knowledge on either the Wi-Fi or the LTE/5G mobile technologies, which is key because there are very few people who know how to configure devices and provision services on both a Wi-Fi as well as a mobile network. So having the IT manager of an enterprise managing both networks is a big task unless they use a software suite like Zeetta’s that simplifies the whole visualisation and lifecycle management of services across the entire network.

A further point is around security and compliance, which can be a challenge for an enterprise. Having that single pane of glass helps with the pre-approval process, knowing exactly which devices you are being connected to and identify quickly the source of any potential threats even before they becoming problems.

These are benefits brought to the network owner, but Zeetta’s orchestration suite will most likely be used by someone like a Managed Service Provider (MSP), who maintains the network on behalf of the enterprise. An enterprise will typically have an MSP who is overseeing their network, troubleshooting and provision services and, they work very closely with the IT team to manage the entire network. Our software tools enable an MSP to manage, at any one time, many deployments, of heterogeneous networks across multiple sites. With our orchestration products you simplify the network operations both for the IT manager in an enterprise so they can manage the network themselves, regardless or its complexity (in terms or vendor or technology), and for the MSP because they can now support clients with a much wider heterogeneity in their networks. So, the same tool saves time for the MSPs, and save money for their clients.

The other element is the automation and programmability when your network requires frequent modifications. If you are in a small office environment where you configure your Wi-Fi during installation and then you leave it untouched until you replace it with a newer version in few years’ time, then this isn’t going to be particularly applicable. However, if you operate a complex network with many sub-systems, technologies and vendors, such as a stadium or a multi-purpose venue, where one day you have a corporate event, and the next day you have a football match or a music concert, this is the kind of environment where you want to provide customisable connectivity in a flexible way. You need the agility to dynamically change the network configuration according to the ever-changing requirements of your business.

What do you see as the advantages of deploying a private cellular network over public cellular?

Private cellular networks offer some very good benefits, those are well documented. In our 5G-ENCODE project, we are working with the National Composite Centre (NCC) in the UK, to design and deploy a private 5G network to support few key industrial use cases. NCC is a place where different manufacturers will come to develop their composite prototypes. For example, Airbus will be coming to design, develop and test a new aeroplane wing in their own isolated ‘cell’ enclosure. Quite often, they will find themselves working alongside with a competitor, e.g. Boeing, who may be occupying a cell  next to theirs. So, there is a strong requirement of security and isolation of the network environment for each client. Each client wants to have full control of their data (measurements, analytics etc.), they want accurate and deterministic control of the wireless access network and they want all their operations to be kept safe and isolated. Since security is of paramount importance they can’t rely on Wi-Fi to satisfy their requirements. Wi-Fi is also not suitable for the low latency application either.

Public cellular networks are also not suitable for ultra-low-latency application since every time you travel 300 miles you are wasting one millisecond. This means that if your data centre or your mobile core is more than 300 miles away, then you have already lost two milliseconds by simply going back and forth. So, it is clear that it is physically impossible to support ultra-low latency applications through a traditional public mobile network architecture.

What are the benefits of private 5G over private LTE then?

The new type of applications like network slicing and splicing you can potentially do with 4G as well, but not to the same extent that you can do it with 5G. 5G is designed from the ground up to be cloud native and to support such technologies whereas LTE needs to be modified with proprietary extensions in order to support these functions. This software-based architecture, can be applied equally well in a public cloud, or on-prem in a private cloud to support low latency and enhanced security applications.

Network slicing is part of the core 5G capability and will become increasingly important with the roll out of private and public 5G standalone deployments.

Zeetta’s expectation is that in the next five years, network orchestration will be performed using network slicing and splicing so that network slices will become the fundamental services that you could buy from your ISV provider and deliver not only connectivity, but other added-value services.

What needs to happen to accelerate the adoption of private 5G networks?

We strongly believe that the wide adoption of private 5G networks depends on the simplification of the deployment and management of these networks with software tools that orchestrate and automate network operations.

It is very likely that future deployments would be multi-vendor environments because both enterprises and services providers don’t like to tie themselves to particular vendors. This is a very important issue and it is related to the global discussions around the diversification of the telecommunications supply chain. You The market and national regulators demand a variety of suppliers and technologies which work well together. But in order to achieve that you need to have the right software tools that operate across vendors and technologies that simplify these deployments. It’s no good to have a very complex network that nobody can use and your IT manager has to go for training to understand how to deploy it. It has to be deployed in a much easier way. We want to give companies the tools that will allow this to happen, to make the 5G private networks as simple to deploy as Wi-Fi.

What role do you see telecoms operators playing in accelerating the adoption of private cellular networks?

Network operators can play two roles. The first very important role is for them to become like the MSPs of larger enterprises. They can help enterprises to build and then become the managers of enterprises’ private cellular networks. That’s quite straightforward, and they’re doing it already in many ways.

The other role is that they can extend their reach into private networks by allowing a slice of their public network to be attached to a slice of the private network in a neutral hosting environment. This will allow telco operators to gain market share in the indoors wireless access networks which so far has been dominated by Wi-Fi. To me, that’s the greater, game-changing opportunity.

Author: Matt Bamforth is a consultant, specialising in edge computing, telco cloud and 5G