A road map for 5G as a profits engine for network operators
A huge amount of 5G research energy has gone into pushing the technology boundaries to the state of the art and this has been matched by many 5G visions unbounded by current technical or economic reality. All this must now change if 5G is to begin to be rolled out in 2020. Visions and state of the art show pieces have to give way to a road map for the sort of 5G network technology that is mature enough to be deployed by 2020 and gives mobile operators the right commercial incentives to invest in it. A successful strategy has to weave a 5G story line through four critical elements:
- A plausible business justification
- A set of technologies able to deliver it
- Suitable spectrum to roll out the new networks on
- Emerging global standards that will secure early scale economies
The story-line is important in binding together all the diverse actors along the value chain to work together to a common delivery date.
The global (3G PP) standards RAN Workshop in September 2015 has taken a first step in linking the first and fourth elements by narrowing down the business justification options (they’ve termed “use cases”) to:
- Enhanced Mobile Broadband
- Massive Machine Type Communications
- Ultra-reliable and Low Latency Communication
The Enhanced Mobile Broadband has yet to be refined in terms of what exactly is to be enhanced, by how much and where. The obvious thing to be enhanced is data speed (and capacity) but lower latency could be a part of the mix. Driving the decision of “how much” and “where” will be the choice of radio spectrum. The WRC-15 provides the first real glimmer of the spectrum options for 2020:
- mmWave spectrum – Various bands from 24.25 to 86 GHz will be looked at in WRC-19 with 31.8-33.4 looking a likely possibility for Europe
- 3.4 to 3.6 GHz and possibly 3.6 to 3.8 GHz in Europe
- UHF – 700 MHz in Europe and possibly 600 MHz in the USA
These spectrum choices neatly answer the question of “how much” and “where” as mmWaves can support the very wide RF channel widths to deliver up to 10 Gb/s or more but very limited coverage. Spectrum around 3.6 GHz offers coverage of enhanced capacity over dense urban areas, perhaps delivering 1 Gb/s or so depending on the RF channel width made available. Only 700 MHz can deliver national (or universal) coverage but the channel width will limit data rates to 50 Mb/s or down to 10 Mb/s if the aim is to push the extremes of coverage.
Let’s try to map the above to a 5G road map. The mmWave technology, with it very sophisticated MiMo, beam steering and very low latency blows away any worry of capacity crunches and can deliver ultra low latency. It boosts the untethered short range “nomadic” market. It is an exciting prospect. It is also an incomplete solution as it does not”mobilise” high capacity/data speed access over contiguous wide-areas. Nevertheless it is a headline grabbing icing on the cake.
Mobile operators have always enjoyed a price premium over fixed network connectivity that derives from the “convenience” consumer crave. Wide-area mobility frees consumers from having to think about when they are. In this regard the 3.4-3.8 GHz choice is really interesting as the band offers better radio propagation than WiFi at 5 GHz and potentially much wider radio channels than WiFi at 2.4 GHz and licensed spectrum at 2.6 GHz. As licensed spectrum it also offer higher powers than WiFi. So it begins with a competitive edge all the way round just from the choice of spectrum. What about the technology itself? The two WiFi technologies and LTE technology have been around for long enough to know what needs to be done to deliver a better technology. For example signalling for LTE is too over centralised and for WiFi is too decentralised. A technology developed at the 5G IC called MACLite offer the best of both worlds. The cell edge shortcoming with LTE is well known and the 5G IC have a “pilot” technology that can deliver a 14 fold improvement at the cell edge. These are only examples. The list has to be built-up within the standards body to provide a very powerful small cell technology that will allow mobile operators to provide something indoors with an edge over WiFi (due to the choice of 3.4 GHz frequencies), extend those high data speeds over dense urban areas (securing a mobile premium) and re-using the spectrum/technology for low cost high capacity back haul in semi-rural/semi-urban areas. This will allow the mobile network operators to offer more and more data capacity to consumers every year whilst holding down prices and keeping up profits. Such 5G technology will need to be matched by complementary measures to deliver the full business benefits eg incentives for consumers to host “open” base stations to keep site rent costs down and spectrum sharing (such as licences shared access for low powered indoor use) to boost wider use of the technology to drive down unit costs through scale economies…these lower unit costs then drive through to lower dense deployment infrastructure costs.
For me “the sleeping beauty” of 5G network technology is an adaptation of 4G technology (or a new waveform) for the 700 MHz band where the mobility premium is at its maximum. There is much to be gained by a combination of improved technology and basic re-engineering (eg selective taller masts) that could put a massive 20 dB (factor of 100) in the link budget over current 4G networks to deliver much greater reach and reliability. An interesting idea is to create a moderate data rate universal control plane that can also carry priority traffic. The pilot technology mentioned earlier could be applied to improve cell edge performance. This may not all add up to the zenith of “ultra-reliable” but good enough to attract new profitable business opportunities from those currently served by private high resilient networks and new applications that want national coverage of a predictable Quality of Experience. Where I would place a 700 MHz 5G option on the road map would be around 2022.
So far I have avoided mentioning massive machine type communications (IoT). This is because “massive” anything needs exceedingly cheap devices and it will be some while from the first deployment of a 5G device (which traditionally arrive later than the networks) before 5G devices have fallen down the scale economy curve to become that exceedingly cheap. This leads me to best guessing 2027 before IoT comes anywhere near driving demand over a 5G network. The only thing needed by 2020 is a new architecture to be put in place so that 5G networks can be ready for a massive surge in connectivity from 5G IoT when it finally comes.
Finally all of the above has focussed on the radio access. 5G is likely to pull in its wake significant changes to the network including a much flatter architecture, much smarter demand-attentive resource management, prudent energy management and closer integration between some device and network functions to simplify for consumers being always best connected. Deploying these improvements could be done at different times by different network operators but what has to be brought to a consensus time-scale are any associated changes needed within smartphones and other devices for the mass market. This will require a more concerted approach by network operators to implement 5G end-to-end network up-grading. The final piece of the modernisation jig saw will come with the modernisation of Internet TCP/IP protocols. The sheer size of the implementation challenge will put this reform at the very far end of the road map and I would stick a flag in the ground at around 2030.
2020 roll out of much better broadband small cell technology at 3.6 GHz towards dense urban coverage with mmWave hot spots providing headline grabbing icing on the cake.
2022 roll out of a combination of modestly better broadband and greater coverage/reliable technology at 700 MHz toward universality
2025 roll out of small cells at 3.6 GHz start to become dense enough to provide substantial contiguous dense urban coverage. The emergence of such viable “mobile” coverage will allow mobile Gb/s applications to begin to emerge and drive demand for even more pervasive coverage.
2027 mass deployment of exceedingly cheap 5G IoT devices
2030 modernisation of Internet TCP/IP protocols