If communication service providers (CSPs) are going to take full advantage of 5G, they will need to anticipate a wave of new use cases that will redefine their digital service offerings.

 

In our previous post we talked about how DevOps is at the heart of 5G. In this follow-up blog we’ll take a closer look at why Communication Service Providers (CSPs) need a different approach when building their 5G capabilities. We will take a walk through 5G architecture, the challenges brought on by it, and consider different use cases that are made possible by 5G and which could be leveraged by Digital Service Providers (DSPs).

What is 5G?

5G is the next evolutionary step in mobile technology. For the end user 5G promises even higher download speeds and lower latency. As a quick comparison here are some numbers about different generations of mobile networks (see source.)

Network type

Average download speed (Mbps)

Max download speed (Mbps)

Latency (ms)

3G

0.5

3

65

4G

33

100+

40-50

5G

130-240

1000-10000 (theoretical)

20 (actual)

1 (theoretical)

 

Form the user equipment to the mobile radio network

How are these numbers achieved? Practically the same way as before. Mobile masts transmit radio signals to the end user equipment and provide connectivity, but the critical difference with 5G is that it is mostly done in a higher frequency spectrum. These new spectrums bring their own challenges. They can cover significantly smaller areas and can be more easily blocked by everyday objects, like buildings or the natural environment. To battle these shortcomings 5G New Radio (5G NR) is being introduced as a next generation air interface to improve the performance, scalability and efficiency of current mobile networks. 5G NR utilizes optimized OFDM-based (Orthogonal Frequency-Division Multiplexing) waveforms, multiple access and Dynamic Spectrum Sharing. Standardization bodies have still not closed on a single OFDM waveform to be used and will most likely settle on multiple solutions for different use cases.

Introduction of Massive MIMO is also a core component of such fast 5G networks. MIMO (Multiple-Input, Multiple-Output) is the transmission and reception of multiple data signals over the same radio channel. Traditionally, the antenna setups consist of 2 to 4 antennas, but for Massive MIMO vendors have introduced 96 or even 128 antenna setups. This can multiply the capacity of the connection without the need for additional spectrums.

On the radio access side we can already see the first problem statement of 5G architecture. Network density is significantly increased compared to previous generations of mobile networks. The number of mobile base stations and antennas may have to be increased by over 100x according to industry estimates. Such a number of network elements can only be efficiently operated and maintained if an end-to-end automation system is available and CSPs and vendors work in close cooperation to shorten either the mean time to deliver, or the mean time to recover, so that the end user experience is not compromised in failure scenarios. Automation and collaboration are the two basic DevOps principles which need to be deployed along with the radio networks.

Multi-access Edge Computing (MEC)

Multi-access Edge Computing, or less precisely Mobile Edge Computing, is a form of network architecture which greatly contributes to the speed and the latency requirements of 5G networks. MEC enables cloud computing and applications to run at the edge of the mobile network, effectively reducing the physical distance between user equipment and application, or between two user devices. In a 2015 white paper ETSI identifies MEC as a “key technology and architectural concept to enable the evolution to 5G”. MEC enables the local runtime for applications and thus the collection, storage, and analysis of the data, which makes it a key enabler for  e.g. IoT use cases.

The introduction of MEC architecture presents another challenge for CSPs in terms of the number of MEC locations, the number of application instances, and the overall cost of the MEC sites. However, the introduction of MEC is not just beneficial for CSPs but also for public cloud providers. We have already seen many collaborative initiatives from AWS, Microsoft, IBM and many others, to provide public edge cloud solutions.

The potential use of hybrid cloud solutions could enable new capabilities for CSPs, such as cloud bursting, or new procedures for archiving, releasing and deploying new software versions, or even handling disaster recovery.

In parallel with these capabilities CSPs will need to face the challenge of a diverse cloud infrastructure, software distribution, and provisioning on a significant increase in the number of sites. Such demands reemphasize the importance of key DevOps principles concerning automation and collaborative ecosystems, where CSP, public cloud provider and telecom vendor can contribute.

Containers! Containers everywhere!

Applications are also changing in 5G. As new architectures like MEC are introduced the aggregated edge and the core network elements need to change too. Beyond the obvious architectural requirements and standardized features, other software quality characteristics like operability, compatibility and maintainability have to be more dominant. The sheer number of applications and the frequency of new service introduction, driven by the digital service consumer, forces telecom vendors and CSPs to reassess the current VNF architectures and push forward with implementation of cloud-native network functions (CNFs). Containerization and CNFs are part of a growing trend that seeks to answer most of these challenges by standardizing interfaces, providing a consistent runtime environment, and easing the application lifecycle management. In return, containers and CNFs have their own challenges. We already have fractured infrastructure environments and lower maturity in e.g. security tooling, as well as  problems with stateful containers. Version and release management, monitoring and analytics are also areas where significant improvement should be undertaken in coming years.

As we have seen in the poll responses during a recent Light Reading webinar, the vast majority of respondents expect to have half of their workloads running as microservices in containers by year 2025. However, do not be fooled into thinking there is ample time for implementing the operating environment and automation because  the implementation of such systems can easily take years.

All about the data

In a 5G network where billions of connected devices are expected everything generates data and effective network automation should try to exploit it. From Consumer Data Platforms (CDP), Network or Service Operations Center (NOC and SOC), KPIs and metrics, to application performance indicators and logs, all data should be collected into appropriately designed data lakes. This enormous amount of data should be put into operational use by various machine learning and even AI algorithms. These ML and AI solutions could transform fairly simple, scripted, event-driven networks into real agile networks capable of self optimization and self healing.

5G use cases

According to these calculations a heavy internet user household of 4 can be satisfied with less than 200 Mbps bandwidth, so what are the use cases for demand greater than that?

3GPP defined 3 basic use cases as part of its SMARTER (Study on New Services and Markets Technology Enablers) project. These are Enhanced Mobile Broadband (eMBB), Ultra Reliable Low Latency Communications (URLLC) and Massive Machine Type Communications (mMTC). Enabler technologies for such use cases are Gigabit LTE (GiLTE), Massive MIMO, Millimeter Wave (mmWave), spectrum sharing technologies, MEC, Fixed-Wireless Access and Narrowband IoT.

Enhanced Mobile Broadband (eMBB)

eMBB use cases are data-intensive use cases which require high bandwidth. Such cases include cloud and UHD, 8K video streaming, immersive gaming (including AR and VR) gaming, video analytics, immersive event experience, and telemedicine. 

Ultra Reliable Low Latency Communications (URLLC)

URLLC or Mission-Critical Control use cases require extremely low latency with high reliability, availability and security. Such cases can be AR enabled maintenance, AR  supported tourism, autonomous vehicles or vehicles to everything cases (V2X), product line automation, remote operations, telemedicine or any Tactile Internet application.

Massive Machine Type Communications (mMTC)

mMTC use cases are cases where low energy devices with low data volume connect on a massive scale such as in smart cities, as autonomous vehicle assistants, remote robot controls, or port/airport operations. Enhancement of Narrowband IoT is needed to utilize mMTC to its full extent, such as voice support, location support, device mobility and over-the-air firmware updates. Also efficient uplink transmission with multiple access is needed in the form of WAN-managed multi-hop mesh architecture. 

Digitalization redefining the role of CSPs

“As telcos undergo digital transformation, they evolve from communications services providers (CSPs) to digital services providers (DSPs),” says Microsoft’s Rainer Kellerhals. But to successfully capture the benefits of a 5G network CSPs need to define their digital services - and the use cases they want to provide and monetize for their end users - during the process of digital transformation. The real business case of 5G lies not within the new network architecture and higher download speed for subscribers, but in the digital service offerings that would distinguish CSPs from “dumb” pipelines.

To capture this potential we are offering a facilitated collaborative session of service design that brings together influencers of multiple business branches, provided jointly by Eficode’s 5G DevOps and design teams.