Dedicated private networks vs. network slicing: How do they compare?
We see dedicated private networks and network slicing as two (of many) different ways of building a ‘private network’. In this article, we explore what each of these deployment models are and how they compare against each other.
Dedicated private networks and slicing: Under the same umbrella
There are multiple ways of building a private cellular network, each of them varies in terms of the level of physical and logical isolation of (or integration with) the public network and ownership of the different network functions. This results in a wide range of services under the banner of ‘private networks’ with different combinations of spectrum, radio and core networking from public mobile operators, enterprises themselves and/or other third parties.
In our report ‘Private networks: Lessons so far and what next’, we summarise these into three main models of deploying private cellular networks and explore these in more detail:
- Dedicated private network
- Network slicing
- Partially integrated hybrid network
This article will purely focus on the first two and assess the differences between the two. We will explore the partially integrated hybrid network model in a separate article.
Dedicated private network
This is an independent network that is completely separate from the public mobile network. We define a private cellular network using the following characteristics:
- On-premise network designed to cover a specific location, site or premise
- Uses dedicated spectrum for 4G LTE or 5G technologies, which can be unlicensed or licensed spectrum (e.g. can be leased by a mobile network operator or a third party)
- Has dedicated software (operating functions e.g. radio, core, management) and hardware within the customer premise
- Hardware could be enterprise’s own dedicated infrastructure, leased dedicated capabilities from an MNO or a third party, or dedicated assets under a managed services contract
Tasks involved in the overall lifecycle, from planning and design, deployment, integration and ongoing management, can be delivered by different types of partners. This includes traditional mobile network operators (MNOs), new emerging specialist players (some examples here), systems integrators and technology vendors.
Standalone dedicated private networks are often deployed in the following situations:
- When complete physical and logical isolation is required for security and privacy reasons, where there is highly sensitive data or operations
- For remote and rural locations/deployments e.g. in oil, gas and mining
- When the highest availability, reliability and even lowest latency is required for mission critical use cases
Dedicated private networks tend to have a higher deployment cost than the other two models and require specific engineering skills (e.g. RF planning).
This refers to the ability for an MNO to virtually segment its network to create logical functionally-discrete end-to-end networks (network slices) over common infrastructure. Each slice can have the functionality of a complete network, with specific network layer characteristics and operational parameters that can be tailored to specific needs. This essentially allows MNOs to segment the network to create ‘private slices’ that have different SLA agreements or layers for different groups of stakeholders and use cases. According to 3GPP, network slices can each have “different requirements on functionality (e.g. priority, charging, policy control, security and mobility), differences in performance requirements (e.g. latency, mobility, availability, reliability and data rates), or they can serve only specific users”1.
Most activity around network slicing at the moment is still focused on static slices, i.e. slices with characteristics that are individually designed and pre-defined in advance, which is typically supported by non-standalone 5G. As we move to a fully virtualised and cloud-native 5G standalone core, MNOs can slowly move to more dynamic network slicing where the MNO and even third parties (e.g. an enterprise customer) can create their own network slice instances and configure them in real-time to changing requirements and demands.
Although we see network slicing falling under the private cellular networking menu, it doesn’t necessarily mean that it directly competes with what dedicated private networks will deliver. We see potentially complementary models whereby an enterprise deploys a dedicated private network on their site and leverages a network slice for any assets that move outside of the site. For example, an airport may have a network slice that supports certain use cases for the passenger terminals but may deploy a dedicated private network specifically for the maintenance hangars.
However, the number of total slices that MNOs will make available and dedicate to a customer or use case is still unclear. From a technical perspective, this depends on the overall network capacity but from a commercial perspective, this also depends on operators’ willingness and enterprise demand.
What does each model mean for a telco vs. an enterprise customer?
We summarise the key benefits and differences between dedicated private networks and network slicing in the diagram and points below.
- Technology readiness: Dedicated private networks can be deployed using current technologies (LTE or 5G) whereas network slicing is still nascent. Deployments with static network slices are slowly emerging but full dynamic network slicing is not yet widely available or fully mature.
- Cost of the solution: The expectation is that network slicing will be lower than the cost of deploying dedicated private networks given the use of shared infrastructure.
- Level of isolation: Dedicated private networks provide complete isolation for the enterprise as all network elements are dedicated to the customer – this does not rely on any public network infrastructure or resources and therefore can provide higher security and reliability. Network slicing however is still done over shared infrastructure.
- Need for the enterprise to acquire spectrum: There would be no need for the enterprise to acquire spectrum in the network slicing model whereas the dedicated private network model may require dedicated licensed spectrum (depending on the country’s spectrum regime or licensing model)
- Degree of control (for the enterprise): Should the enterprise want it, it would have full control over the management elements of a dedicated private network. However, in future with dynamic network slicing, the enterprise could potentially have the ability to configure their own network slices to meet changing requirements.
- Role of the telco operator: In the network slicing model, the telco would be the primary end-to-end provider. In the dedicated private network model, the role of the telco isn’t as clear cut. Different tasks can be conducted by the traditional MNOs, new emerging players, systems integrators, managed services providers or even the enterprise themselves, particularly in scenarios where the enterprise has access to its own spectrum.
For more information or details on the points above, please see our reports below:
- Private networks: Lessons so far and what next
- Navigating the private cellular maze: When, where and how?
- 5G network slicing: How to secure the opportunity
Author: Yesmean is a Principal Consultant at STL Partners, specialising in helping network service providers and other technology partners commercialise private network solutions.