Network Slicing is a way to run multiple networks on one infrastructure. In theory, it could mean that telcos could run mobile broadband and IoT networks over the same physical network, and also offer greater agility within those networks. We explore the theory and its challenges, examine use cases, what would be needed to make it happen, and look at five business case scenarios.
The Network Slicing research project was sponsored by HPE. This report and the analysis it contains were independently produced by STL Partners.
Service providers continue to face a decline in revenue
STL Partners has written for some time about the significant pressure faced by communications service providers (CSPs), both from operator rivals and players in adjacent sectors. Traditional telecoms revenue streams such as voice and messaging are shrinking, and as a result operator growth is slowing. Figure 1 shows that the average year-on-year revenue growth rate for 68 major telecoms groups worldwide has fallen since at least 2010.
Figure 1: 68 major telecoms groups – aggregate telecoms revenue, 2009-16
Source: Company accounts; STL Partners analysis
Much of this decline is fuelled by the impact of new competition: digital players such as Google, Facebook (including Whatsapp), Microsoft (including Skype and Skype-for-business) and Netflix, who are equipped to provide their own digital services, including voice- and messaging-enabled applications, without the headache of maintaining capital-intensive network infrastructure. It is now widely acknowledged that voice minutes and SMS bundles will continue to decline as a revenue stream as other players can offer the same, or better, capabilities ‘over-the-top’ to consumers and organisations for much less or free.
Data is not enough to ensure future growth
Of course, in order to use these new digital services, organisations and consumers do need network connectivity and, as a result, data consumption levels have shot up. Currently, the only players able to offer data connectivity are the communications service providers themselves, and therefore many have pointed to data as the primary source of new revenues which might offset the gap left by the decline in voice and messaging. In developed markets, in particular, some operators hope that it may be possible to ‘premiumise’ data services and drive higher average revenues per user (ARPUs). We do not believe that the evidence supports this and anticipate that plummeting data connectivity rates ($/MB) will neutralise growth in volumes resulting in low or no net growth in revenues.
In many developed markets, intense competition and strict regulation restricts the ability of operators to resist data price decline and squeeze more out of customers. Figure 5, for example, shows that despite mobile data consumption in the United Kingdom growing 243% between 2013 and 2015, ARPUs actually fell 4.5% over the period. More data, it is clear, does not automatically translate into more money.
Figure 5: UK mobile ARPUs and data volumes, 2013-15
Source: STL Partners, Ofcom
In Figure 6 below, we show our revenue forecast for a telecoms operator offering converged fixed and mobile telecoms services to both enterprise and consumer customers in a developed market. In this conservative estimate, data revenues grow slightly, but not enough to offset voice and messaging revenues falling by half.
Figure 6: Forecast revenues for converged telco in advanced market
Source: STL Partners analysis
It is STL Partners’ belief that the path to sustainable telecoms growth lies not just in better monetising connectivity, but rather in telcos developing new capabilities of their own, continuously innovating and launching new products and services that more readily meet the needs of their customer base. It is only by doing so, and by leveraging new technology and network assets where possible, that telcos will be able to truly compete with digital players. In essence, communications service providers must either evolve to overcome commoditisation or to embrace it. Either way, they cannot continue business as usual.
Virtualisation and slicing: enablers for change?
STL Partners has written previously about Telco Cloud, a concept in which telcos redefine themselves by adopting cloud business platforms and practices (similar to internet and content players), alongside virtualisation of their core assets. This could lead to increased service agility, and the ability to create new, network-integrated services. In turn, this could drive new revenue growth.
Network virtualisation is still at an early stage, but its adoption is increasingly seen as inevitable. Operators worldwide are already deploying NFV/SDN technology, some setting ambitious virtualisation targets over time. The forthcoming 5G standards, as well as IoT technologies, are being developed with virtualisation in mind, and technology vendors are increasingly evolving their software offerings. If managed effectively, virtualisation could be the catalyst for the transformation towards the digital service provider.
One way in which virtualisation might enable this change is through the concept of ‘network slicing’, under which network operators would be able to operate multiple logically separate virtual networks over a single network infrastructure. This paper examines what network slicing might look like in practise, and what that could mean for CSPs.
Slicing: a vision for fundamental transformation
Defining slicing is not about the ‘what’, it’s the ‘how’
Network slicing is a term that has been discussed quietly in the industry for some time, but it has gained prominence more recently in parallel with the industry’s developing new 5G standards. Slicing has recently become the focus of a public disagreement between industry players involved in driving 5G standards. In essence, one group of operators and vendors are keen on accelerating New Radio (NR) standards in 5G, whereas another group see this as potentially undermining future standards in end-to-end slicing. A related debate also exists within operators between the core network and radio access teams, but that is neither new, nor surprising. These debates are not about slicing, since most parties appear to broadly agree on its potential, but more about how 5G will be introduced: as an evolution of 4G or as a completely new network.
A few considerations
In recent years, network slicing has also gained prominence as a way of creating unified 5G networks, which cover multiple very-different use-cases with a single infrastructure. Turning a necessity into a virtue, this technical “fix” is now being seen as a possible basis for extra capabilities and new services. However, many of the benefits could – and should – be achievable before 5G.
While network-slicing can in theory extend all the way through core networks and down to the radio connection, it is still subject to the laws of physics: if there is no coverage, poor RF propagation, or limited overall capacity, there is a hard limit to what performance can be guaranteed. There are also boundaries at the device, 3rd-party server/cloud interface, or where other networks interconnect, which mean that “end-to-end control” doesn’t always mean an entire system.
It’s important not to fall into the trap of thinking that because we have a slicing “hammer” that all problems start to look like “nails”. Telcos have many other approaches to future service creation and revenue expansion, that lie outside the core network. Content partnerships, vertical-industry solutions, in-home automation and new forms of connectivity all offer opportunities. If network-slicing does not reach its aspirations, there are still plenty of other options for the industry to prosper.
Independently of the 5G debate, slicing can be considered part of a wider trend (in both fixed and wireless networks) towards a more software-centric infrastructure leading to more flexible networks. As more network resources become virtual (rather than physical), operators could readily allocate resources to a particular ‘network slice.’ Hence, slicing is arguably really about the orchestration of operator assets and how an operator is able to effectively manage its network.
This vision affirms that the ‘one size fits all’ model will not applicable for the future where a diverse set of requirements will need to addressed with more customised services: from (enhanced) mobile broadband (eMBB), to ultra-low latency types (uRLLC), to low-power machine-type communications for IoT devices (mMTC).
Figure 7: With slicing, networks can be adapted to customers and applications
Taking the work done by industry organisations, such as The Next Generation Mobile Networks (NGMN) Alliance , 5G Americas and the Open Networking Foundation (ONF) into consideration, STL Partners has developed the following definition for network slicing as the basis for this paper:
‘Network slicing is a mechanism to create and dynamically manage functionally-discrete virtualised networks over a common infrastructure’
- Executive Summary
- Slicing: a vision for fundamental transformation
- Defining slicing is not about the ‘what’, it’s the ‘how’
- How slicing could enable growth
- New services from network slicing
- Evidence of the demand for slicing
- Examples of new services
- The slicing business models
- So, where is the money?
- Scenarios for the telco of the future
- The scenarios imply different business models and ways of making money…
- How slicing might work in practice
- Key challenges to achieving slicing
- Early 5G trials and proofs of concept
- The evolution to slicing
- A tricky transition with major obstacles to address
- Figure 1: Benefits of network slicing
- Figure 2: How might (operator) assets translate into demand for slices?
- Figure 3: ‘External’ slicing business models
- Figure 4: 68 major telecoms groups – aggregate telecoms revenue, 2009-16
- Figure 5: UK mobile ARPUs and data volumes, 2013-15
- Figure 6: Forecast revenues for converged telco in advanced market
- Figure 7: With slicing, networks can be adapted to customers and applications
- Figure 8: Diagram of slicing
- Figure 9: Network slicing compared with existing technologies and services
- Figure 10: Potential benefits of network slicing for network operators
- Figure 11: Google Chrome’s release channels – a model for network development?
- Figure 12: How operating models could change under network slicing
- Figure 13: How might (operator) assets translate into demand for slices?
- Figure 14: Example 1 – Emergency Services VMNO
- Figure 15: Example 2 – Low Power IoT Service
- Figure 16: Example 3 – Pop-up Network
- Figure 17: Example 4 – Global Streaming Service
- Figure 18: Example 5 – Smart Meters
- Figure 19: Example 6 – Renewable Energy
- Figure 20: Example 7 – Mining
- Figure 21: Slicing Business Models
- Figure 22: Mapping out the scenarios
- Figure 23: Where will revenues come from?
- Figure 24: Traditional telco cost structure and operating model is set up to operate networks not innovate in services
- Figure 25: Under the slicing scenarios, the cost structures shift accordingly
- Figure 26: Challenges identified from interview programme
- Figure 27: Phases of network transformation for slicing future