Avren's Basestation Conference has been running for quite a few years. While it attracts a smaller audience than some events, it tends to be more technical and address some of the wider issues affecting the entire RAN (Radio Access Network) ecosystem. From a small cell perspective, it helps to assess what's happening with the "big iron" including Cloud RAN, DAS and macrocell basestations. The venue moved to London this year from Bath and Bristol, making travel easier for some coming from afar.
The Big Picture
Stuart Carlaw of ABI opened the event, painting a picture of the industry context. Recapping what we've all seen, the "givens" included growing traffic levels, a shift to data from voice, and lower overall CAPEX budget spend.
The cost, effort and risk of running a mobile network have all increased – it's not such an easy business as before. Competition has become multi-faceted and is fiercer than ever. R&D budgets are pressured while expected to cover more areas.
There is a landscape change in the vendor space. The older RAN vendors (Ericsson, NSN, ALU, Huawei, ZTE) are seeing a new breed of competitors selling services and software (Oracle, IBM, Juniper, Cisco). Many new technologies are being introduced, moving the telco providers into the services space.
He believes the "bump" of increased revenues will be lower for LTE than it was for 3G, forecasting a drop of $7 billion per year by 2017 in macrocell revenues. There will be a shift of investment towards small cells, especially when capacity expansion of existing installed macrocells are exhausted (adding spectrum, LTE, sector splitting).
LTE being rolled out
Andy Sutton from UK operator EE still believes that Macrocells remain the most cost effective way of delivering additional capacity over an area at the moment. Perhaps that's because they have the luxury of plenty of new 4G spectrum to deploy. Their LTE network reaches 60% of the UK population today, targeting 98% by end 2014. Their double speed LTE (using a 20MHz carrier rather than 10MHz) is already available in 20 cities, and they have plans for quadruple speed (using carrier aggregation of 2 x 20Mhz carriers at 1800MHz and 2600MHz) during 2014. 3G continues to be expanded too, with dual carrier HSPA (DC-HSPA) at up to 42Mbps covering 80% of population by the end of 2013.
Usage figures and patterns for this LTE service were also quite revealing. 25% of data is streaming on smartphones. Larger screens need more data – the iPad and "Phablet" devices consumed 40% of total capacity. It seemed to make sense to me that smartphones are used more during the day, and tablets more in the evenings.
Around 13% of EE traffic overall is social media related, and the enhanced LTE uplink allows more content to be published directly by users. There's no particular generic data profile – 14% is YouTube, 12% web browsing, 11% secure web browsing.
Small cells will be deployed in due course, but would have to deliver the same or better end user experience as the macro network. I asked if this would mean they had to have the same features, such as the same peak 300Mbps quad speed, and was told that it was the QoS experience that was most important, not matching the precise technical performance.
An Italian delegate asked whether the aggregate RF transmission power from shared sites with multiple frequencies and technologies was a limiting factor. While there are limits in the UK, and these do apply in a few cases, these aren't as severe as Italy.
ThinkSmallCell chaired a panel session later during the conference which explored where and when it should be cheaper to deliver data traffic via small cells vs macrocells, and we'll publish some of the findings and comments as a separate article.
Deutche Telekom continue to evaluate and understand small cells
Ole Klein from DT explained his classification of small cells was based on RF transmit power:
- Residential femto <100mW
- Indoor pico/Enterprise Femto <250mW
- Outdoor Pico (micro, metro) 250mW to 10W
If you use more sectorisation or a smart antenna splitting horizontal and vertical aspects, then these also generate additional (virtual) small cells and greater spatial efficiency. At the end of day there will be toolbox solution set and they won't choose entirely one or the other.
Repeating the message I've heard from other DT presenters before, he foresees three main use cases for small cells:
- Extending coverage of macro network
- Improving data rate at cell edge
- Cost efficient capacity extension for certain traffic scenarios; this could be hotspot, hot zone or a complete small cell layer
Deployment relies heavily on the end-to-end cost, not just the cost of the box, and will need further cost efficiency and also greater diversity of product.
Ericsson present their new Radio Dot
Ericsson have recognised that Macrocells can't do everything. They believe small buildings can be adequately handled from an outside macrocell and very large buildings would justify their own DAS (Distributed Antenna System). In between, picocells can be used to serve some small to medium sized buildings. But there is quite a wide gap for a cost effective and efficient solution for medium/large buildings.
They believe that today's enterprise small cell solutions that address this market segment have some serious drawbacks, including a lack of deployment flexibility and are difficult to upgrade in the future.
Their solution is called the Radio Dot, developed as a working prototype in their research labs. I would view this as being a DAS system with the RF heads connected through dedicated LAN cables. Announced last month, this won't be commercially available for 18 months. It effectively splits their existing remote radio head into two functions, with separate RF heads and distribution boxes comprising overall:
- a central macrocell which handless all the baseband processing and control
- a distribution unit (DU) connected by fibre, typically one box on each floor of a large building
- small radio heads (Radio Dots) connected by dedicated CAT5 LAN cables. These don't use Ethernet and can't share an existing Ethernet LAN.
Each Radio Dot can transmit both 3G and LTE at multiple frequencies, sharing the capacity delivered by the macrocell basestation. This is primarily a coverage, rather than a capacity solution.
From my perspective, this still requires the cost of the full macrocell and associated equipment. Dedicated fibre and LAN cabling needs to be pulled throughout the building. Radio Dots need to be installed at each point of presence.
I find it hard to believe this would be cheaper than an enterprise small cell solution, especially one which could expand or evolve the existing Wi-Fi access points already deployed. Unlike a full DAS solution, it isn't natively multi-operator. Unlike an enterprise small cell solution, it doesn't add capacity or share any of the in-building wiring/networking.
Advantages of this approach are that it runs same software as the macro network, so is robust and fully interoperable with other vendors systems. Future upgrades (e.g LTE Advanced features) could be applied at the macrocell without needing to upgrade or affect the installed Radio Dots.
This is a working lab prototype, and commercial product won't be available for another 18 months. Meanwhile, we have several proven, high capacity indoor enterprise small cell solutions available today which run through existing/shared Ethernet networks. Read Spidercloud's robust assessment of the Radio Dot, and I'm sure Cisco and ip.access amongst others would have something to say.
The macrocell business continues to enjoy a short term surge of investment in both 3G capacity and initial LTE rollouts. Macrocell site sharing and the growing number of frequency bands/radio access technologies are important technical factors, driving the need for specialist hardware (filters, combiners, antenna) and physical formats (remote radio heads).
Small cells remain an option, with quite different considerations being taken for indoor and outdoor use. At this event, residential femtocells weren't being discussed much, but enterprise/indoor small cells and metrocells are both being actively trialled and developed as separate activities.
Ericsson's announcement of their Radio Dot solution flags up that there is a real market requirement for indoor enterprise small cell product – they are (belatedly) reacting to what their customers are asking for. With proven, existing indoor small cell solutions available today perhaps this indicates we might see a growing interest in this segment now that the first LTE launch and rollout engineering programs are well in hand.