On the same day as Qualcomm’s first public demonstration of millimeter wave (mmWave) - a technology central to 5G - Ericsson announced its prediction of 150 million 5G subscriptions by the end of 2021. This intriguing contrast of 5G reality versus hype suggests it’s time for a 5G reality check.
While there have been many 5G demonstrations to date, Qualcomm’s was interesting because it showed how mmWave will support high-speed mobile connections in high-frequency spectrum. The demo, a TDD prototype system running in the 28-GHz band, included a base station with 128 antennas and a device with 16 antennas. It used intelligent beamforming and beam tracking to maintain a connection as a researcher slowly moved the device. With data speeds near 500 Mbps downstream and 80 Mbps upstream in a 226-MHz channel, it showed how 5G technology will make mobile broadband services viable in high-frequency spectrum.
While the demo was a prototype system, Qualcomm made it clear at the event that it would bring 5G chipsets to market in time to support commercial deployment of 5G services in 2020. It seems the entire industry is working to this timeline.
See Also
5G Insights February 2016
Three market segments
Amid an increasing number of 5G demos by Qualcomm and other vendors, consensus has emerged about key aspects of the 5G standard. 5G is now designed to serve three different market segments. These are enhanced mobile broadband, massive Internet of Things (IoT) connectivity, and ultra-reliable communications. Key features of the different segments include:
- Extreme mobile broadband: High speeds measured in Gbps, latency down to 1 millisecond (ms) end-to-end, and the use of high-frequency spectrum bands above 6 GHz.
- Massive IoT: Low power consumption, low cost, and the use of low-frequency spectrum bands to provide broad and in-building coverage.
- Ultra-reliable communications: High reliability, high availability, and low latency down to 1ms end-to-end.
Key design features
In order to serve these different segments, 5G is designed to be flexible and scalable at every level - from the physical layer of equipment through to the network architecture. Some key design features:
- Unified air interface: A new 5G air interface is being developed to work across a huge range of spectrum bands and channel sizes and support a variety of different use cases with different requirements across key performance metrics like data speeds, power consumption, latency, and system reliability. This air interface relies heavily on OFDM (the basis of LTE and Wi-Fi) but is a new design with flexibility and scalability to meet new requirements. New wireless technologies - including mmWave and massive MIMO - are key to the new 5G air interface.
- Flexible network architecture: 5G relies on NFV and SDN to create a more distributed and flexible network architecture. The aim is to use virtualization to provide different devices and services using “network slices” containing the requisite network features and functionality. The aim of 5G is to deliver different levels of performance - such as throughput and latency - across a common infrastructure.
- Multi-connectivity framework: 5G will build on existing features such as carrier aggregation and LTE Direct to allow deeper integration of different networks and devices. Devices will connect simultaneously to multiple networks and will aggregate spectrum and services opportunistically. Devices will also become key parts of the network and connect directly to other devices to expand coverage and improve reliability.
The flexibility of the new system will support a variety of new business models. For example: once network features or “slices” are customized to specific devices and/or services, different network slices will carry different costs. By extension, services like Netflix that demand higher levels of speed and quality from the network must pay a premium - as long as this complies with net neutrality regulations. This could help operators fulfill their longstanding ambition of getting OTT players to help pay for networks as all devices and services will effectively be going through the middle (TTM) of 5G networks via their own customized network slices.
Beware 5G hype...
Companies across the mobile industry are hard at work developing 5G technology, but it’s still three years until the initial 5G standard is finalized, and five years until 5G services launch. But we’re already seeing 5G reality clouded by 5G hype - for example, Ericsson recently released a forecast that 5G will have 150 million subscriptions globally at the end of 2021, just a year after services are set to launch. For the forecast, however, Ericsson defines a 5G subscription as follows: “A 5G subscription requires a device capable of supporting LTE Evolved or NX, connected to a 5G-enabled network, supporting new use cases.” NX is defined as the new radio access technology for 5G. This definition means that the vast majority of early “5G subscriptions” will actually be via 4G “LTE Evolved” devices, given that one of the early barriers to adoption of 5G will be limited availability of 5G devices.
This method of defining a 5G subscription makes sense - as 5G will rely heavily on 4G and its evolutions - but inflates the early potential for 5G. A more traditional definition of a 5G subscription would be something like: “a 5G commercial device connecting to a 5G commercial network.” Similarly, a 4G subscription is commonly accepted as a 4G LTE commercial device connecting to a 4G LTE commercial network. HSPA+, which is an evolution of 3G, has been marketed as 4G in some regions, but we should not let marketing claims change the reality that HSPA+ is not actually 4G.
See Also
5G Insights February 2016
...and IoT hype
5G will be the first system built from the ground up with the IoT in mind, but beware the massive hype of the IoT market. In 2009, Ericsson issued its famous prediction that there will be 50 billion connected devices worldwide by 2020, a forecast echoed by other vendors such as Cisco. However, in June of this year Ericsson redefined its forecast and revised it down to 26 billion connected devices by 2020. Then in November the group refined the forecast to 28 billion connected devices in 2021.
Different groups have very different approaches and methodologies to address the challenge of forecasting connected-device numbers. But cutting a forecast roughly in half suggests an element of hype in the initial forecast. And if companies used the initial forecast to build their business cases, what happens to those business cases (and overall businesses) when half the expected market disappears?
Ovum does not have an overall forecast comparable to Ericsson’s, but our forecast for cellular M2M connections can be compared to one segment of the Ericsson connected devices forecast.
Ovum projects that cellular M2M connections globally will increase from 297 million in 2015 to 660 million in 2020, while Ericsson expects devices with cellular M2M connections to increase from 400 million in 2015 to 1.5 billion in 2021.
Mike Roberts is practice leader for the Americas at Ovum
This article was first published in Telecom Asia 5G Insights February 2016 edition