Erol Hepsaydir, Head of Ran Strategy at Three UK, gave an excellent presentation at HetNet world last month, outlining the perspective of how he sees future technology meeting their goals of coverage, capacity and quality. Key points included differentiating RAN features that do/don’t need device support, tricks and tips for small cell integration and the importance of SON.
A snapshot from today
Although the smallest of the four UK mobile networks, Three UK carries more data traffic and still seems to thrive by offering unlimited data packages. There’s no premium for 4G which now carries 29% of data. Streaming media (video and audio) soaks up 60%, web browsing 26% leaving interactive messaging, voice and everything else to fit into the remaining 14%. This validates previous industry forecasts that video will comprise an ever greater share of our wireless data consumption.
Reliability and reputation are important, and like any other network Three want to balance the (sometimes conflicting) requirements of coverage, capacity and speed. Rather than focus on peak speed which is not seen as a critical factor for customers, Erol commented that they target reliability and consistency with adequate speeds – those watching a video aren’t going to be interested in occasional bursts of 300Mbps, they want their continuous feed of (say) 1Mbps (often less).
10 Year capacity growth
We’ve all seen those 1000x demand curve slides in recent years. Erol reminded us of the three main options available:
- more sites (spatial spectrum reuse),
- more technology (spectrum efficiency)
- more spectrum.
He gave what I thought was a somewhat pessimistic but certainly pragmatic view of perhaps 4.5x known and 40x growth potential using known technology, positioning small cells as a key element in that growth.
Differentiating between RAN features that do and don’t need new devices
There are about a dozen new LTE-Advanced features in the 3GPP standards pipeline which promise to deliver additional capacity. Erol reminded us that half would require new smartphone devices before they can achieve that promise. With smartphone replacement rates slowing down, and many older devices remaining in circulation, it adds a few years to full return on investment. The table below shows the split.
| No Device Support Required | Device Support Required |
| ICIC (Inter-Cell Interference Co-ordination) | CoMP (Co-Ordinated Multipath) with joint processing |
| FSS (Frequency Selective Scheduling) | NAICS (Network Assisted Interference Cancelation and Suppression) |
| CoMP (Co-ordinated Multipath) transmission with co-ordinated scheduler | eICIC/FeICIC (Enhanced ICIC/Further Enhanced ICIC) |
| 4x2 MIMO | Dual connectivity |
| 3G refarming | 4x4 MIMO |
| Cell range extension | evolved Multimedia Broadcast Multicast Service (eMBMS) |
| New Spectrum |
While Small Cells weren’t explicitly the list, some of the features that would exploit them more extensively are. They don't require any new device features for initial deployment, but could benefit from some of the more advanced aspects that do.
| Radio Feature | Capacity Gain | Device | Sync/ Latency | |
| Evolution | NACIS | 6-10% | Y | Y |
| MIMO | 4x2 MIMO | 23% | ||
| 4x4 MIMO | 77% | Y | ||
| DL intrasite CoMP | 10-25% | Y | ||
| Small Cell | 3 x ePico | 350% | ||
| eICIC/FeICIC | 320% | Y | Y | |
| Smart FeICIC | 350% | Y | Y | |
| Dual Connectivity | 32% | Y | Y | |
| LTE-U | 200% | Y | ||
| Multi-Sector | 6 Sector | 80% | ||
| Mobile Broadcasting | eMBMS | Depends on service usage | Y | Sync |
Densification with Small Cells
It seemed to me that Three are considering Small Cell densification primarily for urban/public spaces. While they already offer a 3G residential/SOHO femtocell, I’ve not seen much public activity around in-building/enterprise systems. This may be because their customer base is more predominantly consumer oriented. The charts shown indicate more of an “outdoor-in” approach with urban small cells deployed in areas of highest traffic demand and/or coverage notspots. The presentation focussed on LTE rather than 3G, which is where I sense most of the urban small cell investment is headed.
Interference management was clearly a concern, and the use of eICIC (and FeICIC) to synchronise between the small cells and macrocells very important. This will require devices that support those features and tight synchronisation throughout the network.
Cell Range Extension enables the small cells to capture more traffic and don’t require any additional smartphone device support. The stronger signal from nearby macrocells is ignored until a threshold is crossed. The amount of that bias is limited by the performance of control channels, recommended to as much as 6dB if eICI/FeICIC is enabled. Unlike some other feautures, it works with today’s smartphones.
Dual connectivity, where the smartphone communicates with both small cell and parent macrocell effectively allows efficient data offload from the macrocell while remaining under it’s control.
None of the above features are essential to the first deployment and densification of urban small cells but would improve the performance and long term capability of the network.
The growing relevance of Self Organising Networks
SON seems to have been a topic that lurks in the background, visible but not yet mainstream at many conferences. It’s clear that Three, like many networks, recognise the growing importance of automation in their back office configuration and the need for real-time optimisation systems. As the number of cells increases, SON enables faster deployment and increases efficiencies all round.
SuperVoice strategy
Erol explained that to increase coverage specifically for voice calls, Three acquired a sliver of LTE spectrum at 800MHz and has just launched VoLTE. It is aggressively upgrading their macrocells with this capability so that it can reach inside buildings and increase the range of coverage from their existing sites, with 50% population coverage already in place. The feature is marketed as SuperVoice with an emphasis on coverage rather than HD Voice quality or faster call setup.
Unlike some of the other networks which have access to 2G GSM using 900MHz, their higher frequencies at 1800 and 2100MHz restrict their coverage footprint. For a year or two they’ve also offered an iPhone/Android App that uses Wi-Fi for voice/text. Since this isn’t integrated into the dialler, I’ve found it can be quite awkward and sometimes unpredictable in use.
So while SuperVoice only works with the latest VoLTE capable smartphones, it does provide a fairly quick solution for voice coverage. I still have the opinion that some inbuilding systems will be required in the longer term to attain the balance of coverage/capacity/speed which Three seek throughout their network.
Summary
Three UK have accommodated dramatic increases in data traffic in recent years and continue to evolve their network with increased support. This includes a controlled adoption of LTE and future deployment of small cells.
Those features, such as Cell Range Extension, which don’t require device support will bring the quickest returns. Some that do, such as eICIC and dual connectivity, will bring useful benefits in the longer term. It’s clear that RAN strategists such as Erol consider carefully how to make best use of their installed assets first.
But with available LTE spectrum now becoming fully deployed and more advanced features having strong dependencies on the the latest smartphone device capabilities, it seems to me that the inevitability of small cell deployment is noticeably closer.
