What became apparent at this year’s Avren Basestation Conference was that there really is a disruptive change in the way mobile networks are evolving. Little attention was paid to residential femtocells at the event, which has always concentrated on the larger basestations that make up the majority of today’s network capacity. Lots of attention was paid to “small cells” which incorporate femtocell technology but are installed in public areas.
The themes were HetNets (Heterogeneous Networks) and small cells – just don’t call them femtocells – with even Wi-Fi being included in the mix. The large mainstream vendors, Ericsson and Huawei, both presented a comprehensive story with these two new terms. Huawei’s SingleSON (Self Optimising Network) would optimise and tune the network for best capacity and coverage, co-ordinating between the different layers of small to large cellsites.
LTE and the switch to small cells
Vodafone group, represented by Andy Dunkin, their head of new technologies and innovation for radio, spelt out how they had successfully trialled and tested a strategy for rapid LTE rollout. The view that LTE will require many more small cells was reinforced by many other presenters, with Vodafone being clear that a simple, quick and cost-effective wireless backhaul must be used.
Doug Pulley, CTO of Picochip, presented a study suggesting that as many as 70,000 small cells would be required to provide complete blanket coverage and capacity for a large city such as London. Perhaps this was somewhat more enthusiastic (i.e. larger) than the figures others were quoting, but perhaps no longer completely out of the question. I wonder how many Wi-Fi hotspots (including private ones at home) could be found in a city with more than 10 million citizens?
Carolyn Gabriel from Rethink Research has been talking with many operators about their future plans, and her research results show that the total spending on small cells will exceed that of macro basestations by 2015 – I clarified that this figure excludes the non-equipment costs of cellsites (site rental, civil works, tower/masts etc.) – making the transition of the investment even more significant.
Giovanna Romano, from Telecom Italia, explained how the much tighter RF emission regulations in his country, combined with growing vocal opposition to adding antenna to existing masts, mean that less obtrusive and lower power small cells are an obvious route to take.
Simon Saunders, representing the Femto Forum, explained that their mission was to encourage the use femtocell technology to benefit 1 billion users. This wouldn’t be limited to those with their own femtocell at home, instead incorporating some of the low cost, self-optimising intelligence from the industry into a wider range of solutions. It seemed to me that other vendors, previously sceptical of femtocells (especially residential ones), seem more ready to accept this view as long as the "F" word (Femtocell) isn’t used – call them small cells (even though they are functionally almost identical) and everyone can buy into that.
Targetting where to put them
Adrian Pike, from the Radio Architecture and Tools team at Orange, has been experimenting with analysis tools to determine more accurately where the capacity hotspots and coverage holes are. Rather than using expensive drive testing with specialist equipment and staff, the solution effectively uses every network customer as a “drive tester”. No special handsets or cellsite equipment is required; the existing measurement reports fed to the network are passed through via the RNC to a high capacity/high speed analysis engine which crunches the numbers. This can determine the location of users to within 50 metres (it works down to 20 metres if you make multiple calls from exactly the same spot), allowing planners to pinpoint exactly where a small cell would address the issue.
By using this method, they have doubled or trebled the total capacity of an existing macro cellsite at relatively low cost by installing one or two new small cells at exactly the right location. It seems they plan to repeat this technique in future, and I could see many other network planners copying this approach.
The fastest 3G network
Mustafa Karakoc, from the Radio Planning team at Turkcell, explained that because they have much more spectrum than most 3G network operators – a total of 20+20 MHz rather than the more common 10+10 – they could in theory achieve peak rates of up to 160Mbit/s by combining it all together (4 carriers x 5MHz each). This would be more relevant to PC dongle users than smartphones – I suspect few smartphones would be likely ever to have the battery power or configuration to support this mode – but may allow them to match or even exceed claimed LTE rates for a while.
The complex backhaul issue
Backhaul is the transmission between the cellsite and the network operator’s main switching centres. These faster 3G and LTE networks need lots of it – a minimum of 100Mbit/s for new LTE sites – so the issue of backhauling LTE is non-trivial – it is a growing expense and can delay network rollout and upgrade. Andy Sutton, Principal Architect for RAN and Backhaul at Everything Everywhere (the new name for the merged UK mobile operators Orange and T-Mobile) ran through the transition of TDM, ATM to IP and Ethernet. He explained how they plan to overlay LTE on their existing 2G/3G backhaul with the new challenges for speed, capacity and performance.
Quick to install, scalable
Airspan demonstrated their LTE basestation product, which combines a single LTE small cell (capable of up to 100Mbit/s) with a wireless backhaul (also capable of 100Mbit/s) that uses out of band spectrum. Because the backhaul doesn’t need as much overhead as the LTE radio to end user devices (it doesn't have to deal with moving devices or quite so much interference etc.), it can be squeezed into half the spectrum. So the cellsite would use a 10+10MHz FDD spectrum for end users, and a separate 10MHz TDD spectrum block for backhaul. This backhaul is shared between a cluster of these small cells in an area, typically terminating at an existing (macro)cellsite which acts as a hub. As the traffic grows, fibre backhaul can be connected to one or more of the small cells which then becomes a hub in its own right. This approach should allow very rapid rollout to begin with, followed by the option to backfill extra capacity using fibre backhaul as required.
The time to install such small cells, which can be bolted onto lampposts and other street furniture, was suggested as about 1 hour – they only need power. Rather than highly qualified RF experts, anyone capable of installing a TV satellite dish should be able to do it.
This wasn’t about the latest fashionable colour – instead operators really are interested in saving power and becoming more efficient. This is both friendly to the environment and to their bottom line. I was surprised to hear just how important this is in developing countries – in Africa for example, energy costs can be as high as 70% of network OPEX . So consideration was given to comparing the costs of larger numbers of smaller cells versus larger ones, the ability to turn off cells (or carriers) at low traffic times.
The most demanding requirement I heard was from one operator asking for small cells which would wake up from a low power sleep when it got busy in as little as 3 seconds and start accepting calls. It was said two minutes would be far too long to respond.
Up in the air
The last presentation of the conference detailed how Aeromobile, a subsidiary of Telenor, has worked for years to gain regulatory approval to put small basestations (femtocells) onboard long haul aircraft. Their CTO, John Little, explained in detail how they had achieved this. 2G GSM service is already commercially deployed, providing competitive charges (it can be cheaper to make a call onboard than after you land) and convenience (just use your mobile as normal when roaming, all your address book numbers are available, caller ID alerts the recipient to who’s calling, billed as for any other roaming call). 3G including data will be coming soon, with up to 5 femtocells onboard linked by a higher speed satellite connection. A comb jammer is used to block frequencies not used onboard and avoid interference with any ground stations. Regulatory approval has been obtained in many countries with more to follow.
Overall a very interesting conference with good range of presentations, attendees and networking. Recommended for anyone in the mobile basestation, RAN or backhaul industry.
Have you already read my initial report from the event, summarising the first day's presentations?