Hyped as the technology that will transform the world, 5G is moving past the buzzword stage with first implementations coming to life in 2019. Nations are racing to 5G with such fervor that it now became one of the hottest hot-button geopolitical issues.
With latency as low as 1 ms and speeds of up to 4 Gbps, as well as a wider range of frequency bands and enhanced capacity, 5G will be able to accommodate innovative use cases and much greater numbers of connected devices, driving overall growth for Internet of Things (IoT).
In addition to the speed and capacity improvements, a key enabling technology that will allow operators and the society to reap the full potential of powerful new 5G connectivity is network slicing — a novel architecture which enables the creation of multiple virtual networks on top of a common shared physical infrastructure. A huge variety of use cases is envisaged for 5G and it’s the network slicing that will allow granular levels of quality of service for specific use cases to make addressing such variety commercially feasible.
Let’s step back a bit.
So far, cellular networks have been optimized for mobile phones. That was by far the largest use case. However, with the arrival of 5G, the expectation is that cellular phones will become only one of a variety of devices with different characteristics and needs, all connected through 5G.
In September 2015 the Radio-communications division of the International Telecommunications Union (ITU-R), the industry’s standard-setting body, published Recommendation ITU-R M.2083-0  in which it defines the framework and overall objectives of the future development of international mobile telecommunications for 2020 and beyond. In it, ITU-R defined following three usage scenarios for 5G:
- enhanced Mobile Broadband (eMBB): provides higher speeds for applications such as web browsing, streaming and video conferencing;
- Ultra-reliable and Low-latency communications (URLLC): enables mission-critical applications, industrial automation, new medical applications, and autonomous driving that require very short network traversal time; and
- massive Machine Type Communications (mMTC): extends LTE (Long Term Evolution) IoT (Internet of Things) capabilities to support a huge number of devices with enhanced coverage and long battery life.
These use cases will all have different network requirements in terms of latency, mobility, power consumption, security, policy control, reliability, etc. For example:
|5G Use Case||Example||Requirements|
|eMBB||UHD video, gaming||High capacity, video caching|
|mMTC or massive IoT (mIoT)||Large networks of sensors for smart cities, agriculture, etc.)||Large number of connections – over 1 million per km2. Mostly immobile devices|
|URLLC or mission-critical IoT||Autonomous transportation, smart manufacturing, smart grid, public safety, etc.||Low latency (ITS 5ms, motion control 1 ms), high reliability|
Building different physical networks to serve different types of use cases would, of course, not make any commercial sense.
3GPP (3rd Generation Partnership Project) — the international group that governs cellular standards, including 5G — has prescribed network slicing as the approach for the above mentioned scenarios. It is elaborated in 3GPP’s technical specification TS 28.801 . According to the specification network slicing is about transforming the network/system from a static “one size fits all” paradigm, to a new paradigm where logical networks/partitions are created, with appropriate isolation, resources and optimized topology to serve a particular purpose or service category. It means utilizing advanced systems in order to create virtual distinctions between different uses of the ‘same’ 5G network in order to better serve the requirements of those uses and maximize network efficiency.
In other words, network slicing allows operators to “slice” one physical network into multiple, virtual, end-to-end (E2E) networks across device, access, transport and core networks. Each slice logically isolated with fault and security issues contained withing a slice. Each slice with dedicated resources, such as network bandwidth or Quality of Service (QoS), configured for different types of services with different characteristics and requirements.
The concept as defined by 3GPP will redefine how telecommunications sector conducts its business. The partitioning or slicing with appropriate resources and optimization is expected to broaden the horizon of telecommunications in many vertical segments such as automotive, energy, finance, health, manufacturing, utilities, etc. to the extent we have not seen so far. By being able to individually service particular needs of specific industries, telcos could transform from a “dumb” connectivity provider to a critical partner in digitization efforts for a variety of industries.
Network slicing requires an end-to-end approach (defined in the next section) to ensure that performance requirements of various services are met with certainty through the same infrastructure while maintaining harmonization among telecom and non-telecom players. And at the end, the expectation is that the dynamic sharing of the same infrastructure by different tenants will reduce the total cost of ownership and increase profit margins for the stakeholders.
Network slicing is an end-to-end logical instance of a network having tailored network slices (partitions) for individual services. The customization is enabled through a combination of SDN (software defined networking), NFV (network function virtualization), automation, service provisioning and orchestration. SDN separates the control plane (routing) and user plane (forwarding) to optimize the performance of the network while NFV aims to virtualize all physical network resources. The orchestration at the end allows end-to-end management of a slice during its lifecycle [3-5].
Each network slice may consist of multiple network elements from the access to the core as well as application servers (end-to-end). Each slice may have its own protocols, quality of service and security settings. For example, a separate slice could be defined for each eMBB, URLLC and mMTC services on the same infrastructure. The eMBB slice will require higher bandwidth than URLLC and mMTC and thus more radio resources. URLLC is highly sensitive to latency requiring short round trip time for applications such as self-driving, remote surgery. The mMTC slice on the other hand requires the least bandwidth but requires a large capacity to cater for millions of devices .
Therefore, network slicing requires an end-to-end design that necessitates direct involvement of many industries beyond telecommunications. The network slicing issues cannot be resolved by a single SDO (standard development organization) such as 3GPP or industry group such as 5GAA (5G Automotive Association). Hence, they need to work together in finding the right solutions. For example, 3GPP and 5GAA are working together to develop solutions for future mobility and transportation services. Similarly, GSMA (GSM Association) and NGMN (Next Generation Mobile Networks) are defining business drivers, concepts, and high-level requirements in collaboration with operators. There are many other organizations and groups that are involved in these processes. Details of many of them could be found in GSMA “Network Slicing – Use Cases Requirements” whitepaper .
5G along with Network Slicing can be used in a number of vertical industries. The list in its entirety would be too long. Let’s look into just few uses cases.
The energy sector has some very specific requirements that can be met more efficiently with 5G than with preceding generations. One such instance is high speed and reliable communications that are needed between power grid substations and control center. This could be met by a network slice that is well equipped to provide ultra-reliable low latency communications. Similarly, power usage information for millions of devices (i.e. mMTC) could be made available more efficiently by using smart metering slice.
The healthcare use cases are varied. Communications at hospitals / clinics, assisted living in rehabs / care homes, simple personal health monitors and remote surgery are just some of the varied cases where network slicing could be utilized.
At this stage, most of the requirements for the health and wellness are not stringent except during life threatening emergencies and remote surgery operations. For these two circumstances, applicability of network slicing need to be further analyzed.
Augmented / Virtual Reality (AR/VR)
AR/VR perhaps may not be considered vertical industry in itself as it applies to many industries. AR as the term suggests augments or alters one’s ongoing perception of a real-world environment by means of computer generated views applied to single or multiple sensory modes including auditory, visual and haptic. VR on the other hand is a technology that completely replaces the user’s real-world environment with a simulated one.
To achieve audio-video interaction in AR/VR environments , 5G network slices need to support required motion-to-photon, motion-to-sound, mouth-to-ear delay, latency, etc. thresholds. These are stringent challenges which are much tougher than required for existing voice, video and data services.
5G with network slicing could optimize numerous public safety functions where they could get more granular levels of quality of service depending on the application. First responders, for instance, could prioritize specific use cases within those applications – for example, firefighters fighting a blaze could temporarily prioritize voice communication as well as data collection from sensors on firefighters’ gear or within a smart building over other less critical local IoT applications.
C-V2X (Cellular vehicle to everything) is one of the prominent use cases of 5G and likely for network slicing. It is an umbrella term for 3GPP defined V2X services [8-9], which was initially promoted for LTE-based V2X. Specifications for C-V2X for 5G are still under development and as part of 3GPP Release 16 expected to be completed at the end of 2019.
3GPP along with 5GAA formed in 2016 are working directly with telecom and automotive industries on gathering relevant use cases and requirements. Infotainment, telematics, road safety features, and remote steering are some of the use cases that fit in this category.
To support these use cases, C-V2X will require 5G systems to provide URLLC between vehicles, between vehicles and network, and seamless service continuity even when the vehicle moving at high speed. 5G and network slicing will have geographical limitations (at least in the initially years) and thus realistic requirements should be set forth.
Beside many other elements, the success of network slicing and related use cases require the use of artificial intelligence and machine learning. The required speed and granularity demand that 5G network slicing must be effectively managed, making today’s manual and simplistic automation less practical.
A lot has been said and done, however the industry is still struggling to make solid business cases for 5G and network slicing. This can be eased out by making slicing a part of the investment case for 5G and developing a long term roadmap. The cost of deploying network slicing depending on the application / use-case may vary, but bottom line is that it has to be minimal in comparison to an overall investment case for 5G.
Network Slicing tends to divert from the current one size fits all policy where all the applications are served primarily in a best effort mode through a single network. Evolving from today’ single network style to perhaps dedicated network (for customers willing to pay a premium) for eMBB, URLLC and mMTC and to fully sliced model requires deeper analysis as well as a new mindset.
As pointed out by Bell Labs Consulting in its white paper  more connections can be enabled either by dedicated networks or network slices. The study further shows that as the number of connections grow so does the cost for both scenarios. However, dedicated networks incur substantially higher CAPEX (capital expenditure) and OPEX (operating expenditure) than the network slicing model. The primary reason – the infrastructure to the most part is not shared among the different connections in the dedicated networks model. The future networks may very well be comprised of hundreds of customer segments with tailored QoS (quality of service) and security requirements and thus in these circumstances dedicated networks’ model simply becomes unrealistic.
The OPEX is the largest component in the TCO (total cost of ownership) for today’s networks and is similarly expected to be for future sliced-based networks. The practice of running OE (operational excellence) programs is quite common where the two areas of focus are automation and reduction in manpower. These practices will be improved with network slicing. But to make these programs more successful we will require increasing amounts of automation in order to minimize and replace manpower. A similar conclusion was reached in the study which determined that OPEX is the largest component of ownership costs of today’s networks. The study showed that the components of OPEX which cost the most were the ones that required the most human intervention, such as performance and fault management. Network slicing’s high usage of AI and increased automation will thus mean that there will be a reduction in manpower requirements. This will lead to restructuring and layoffs in the industry.
On a positive note, network slicing will increase profits. In particular, the efficiency gains may mean that in some estimates a one percent increase in customers that require tailored connections may result in a three percentage point increase in revenue .
From a bird’s eye view, network slicing is not expected to cause any shifts in the current regulatory regime. The principles and practices of net neutrality, illegal content, quality of service, and cross-border data transfers are not expected to change.
One angle that may be worthwhile to look into is today’s flat taxation regime on mobile broadband (data). With network slicing, the data is segmented and thus different taxation schemes can be applied to different consumer segments, though with much difficulty. For example, URLLC segments may be taxed higher than eMBB due to level of criticality. If that is considered and applied, it will ease the burden on the poorest populations. We know that mobile broadband costs represent over a third of the average annual income of low-income populations in many developing countries [3, 11].
Network slicing brings a number of technical challenges, as well as business and organizational challenges [12-13]. 5G is a matter of months away, and yet serious challenges exist when it comes to making a business case, standardizing, management of network slices, partnering with industry verticals, and others. If 5G is to flourish, these challenges will need to be resolved fast.
The above mentioned transformation of telcos from a “dumb” connectivity provider to a digitization partner for various verticals, enabled by network slicing and multi-access edge computing, is the best way for telcos to increase their profitability. That, however, will require from telcos to undertake an unprecedented organizational change transforming almost every aspect of their operations – from go-to-market and sales approaches that would become akin to professional services’ or system integrators’ approaches, to rapid adoption of automation and AI for network management. Many will fail in this quest. As a result, the telecommunications landscape will likely look very different in ten years.
More specific to network slicing are the challenges due to slicing taking place on a shared infrastructure. This prompts challenges in terms of QoS, handovers between networks, latency, spectrum, roaming, and security, just to name a few critical elements. Thus implementation has two key challenges, namely isolation and resource management. Without proper isolation, slices may not be able to perform adequately, however if slices are assigned dedicated resources these may lead to over-provisioning and therefore reduction of financial benefits of network slicing. Resource management mechanisms are needed to strike a balance for the implementation of dedicated and shared resources. Furthermore:
- Spectrum is a costly, scarce and shared resource. Take for instance, that for a critical surgery an operator has to allocate end to end resources and maintaining best in class in latency, security and QoS for a long period of time. This sort of dedicated service is costly.
- Roaming is another element that needs attention. Customers expect same level of service no matter where they are, at all the times. However, during roaming it will be difficult to meet the demands particularly if the standards are not established. 3GPP has a study item on this particular topic that it will address in its Release 16 . Rel-16 is expected to be completed in June 2020.
- Security becomes critical and challenging particularly when slice crosses the border of telecom world. Different infrastructures will have different security levels and policies since those are managed and administered by both telecom and non-telecom players.
- Business models requires development on per slice / service basis to meet the dynamic demands and traffic variations. These need to become diverse and flexible to match the services.
Potential bottleneck issues might not be technical or regulatory. The level of change for telcos and the industry verticals wishing to benefit from 5G, as well as the amounts involved, is such that it will likely generate issues related to bureaucratic red tape, business ethics, even geopolitics. This will be particularly problematic when it comes to involving critical infrastructure vertical markets, and others that might be to a significant extent controlled by governments. In some developing countries, add corruption to the list of key issues.
The industry is already at the crossroads, because on one hand the technology is making speedy headway while on the other there will be a smaller workforce required due to automation and artificial intelligence. Network slicing has the perfect ingredients to further deteriorate the situation.
Network slicing for 5G era is still shaping up, with many concerns and issues still remaining unsolved. It’s a new, complex, paradigm shifting technology, but the one without which it would be hard to justify the massive global investment in 5G infrastructure. Network slicing requires further due diligence; new business models have to be developed that drive innovative partnerships between telecom and non-telecom players; business cases have to adjusted; standardization of services and handovers across various industry players have to be renegotiated in the much more detail than ever before; demanding SLAs (service level agreements) need to be agreed between operators and vertical markets; etc. All that needs to happen now for 5G to flourish.
- ITU-R 2015. Recommendation ITU-R M.2083-0 – IMT Vision – Framework and overall objectives of the future development of IMT for 2020 and beyond. https://www.itu.int/dms_pubrec/itu-r/rec/m/R-REC-M.2083-0-201509-I!!PDF-E.pdf
- 3GPP TS 28.801 (V15.1.0) 2018. Telecommunication management; Study on management and orchestration of network slicing for next generation network. Technical Specification (Release 15), Technical Specification Group Services and System Aspects, 3GPP, January. https://portal.3gpp.org/desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=3091
- Asif, S.Z., 2018. 5G Mobile Communications: Concepts and Technologies. CRC Press, USA.
- Nokia 2018. Unleashing the economic potential of network slicing. https://onestore.nokia.com/asset/202089
- Tittel, E. 2019. SDN vs. NFV: What’s the difference? Cisco. https://www.cisco.com/c/en/us/solutions/software-defined-networking/sdn-vs-nfv.html
- Vyakaranam, N. and Krishna, D. 2018. 5G: Network As A Service – How 5G enables the telecom operators to lease out their network. https://netmanias.com/en/post/blog/13311/5g/5g-network-as-a-service-how-5g-enables-the-telecom-operators-to-lease-out-their-network
- GSMA 2018. Network Slicing – Use Cases Requirements. https://www.gsma.com/futurenetworks/wp-content/uploads/2018/07/Network-Slicing-Use-Case-Requirements-fixed.pdf
- 5G Americas 2016. V2X Cellular Solutions. http://www.5gamericas.org/files/2914/7769/1296/5GA_V2X_Report_FINAL_for_upload.pdf
- 3GPP TSG-RAN Meeting #73, RP-161894 2016. LTE-based V2X Services, September 19-22, New Orleans, LA, USA. https://portal.3gpp.org/ngppapp/CreateTdoc.aspx?mode=view&contributionId=730345
- Nokia Bell Labs Consulting 2019. Future X Network cost economics. https://www.bell-labs.com/new-5g-whitepaper-download/
- GSMA 2016. Digital inclusion and mobile sector taxation 2016. https://www.gsma.com/mobilefordevelopment/wp-content/uploads/2016/07/Digital-Inclusion-and-Mobile-Sector-Taxation-2016.pdf
- Ezhirpavai, R. 2018. 6 reasons why network slicing challenges 5G’s progress (Reader Forum). Aricent. https://www.rcrwireless.com/20180530/wireless/6-reasons-why-network-slicing-challenges-5gs-progress-reader-forum-Tag10
- Grimaldo, S. W. and Rudd, S. 2017. Network slicing the key that unlocks 5G revenue potential – where 5G meets SDN/NFV. Strategy Analytics. https://www.strategyanalytics.com/access-services/service-providers/networks-and-service-platforms/complimentary/report-detail/network-slicing-the-key-that-unlocks-5g-revenue-potential-where-5g-meets-sdn-nfv
- 5G Americas 2018. Wireless Technology Evolution – Transition from 4G to 5G – 3GPP Releases 14 to 16. http://www.5gamericas.org/files/8015/4024/0611/3GPP_Rel_14-16_10.22-final_for_upload.pdf