The US has quite a different topology from other regions
Driving into Dallas from the airport, you are reminded of just how different the environment is here compared to many parts of Europe or Asia. The city grid viewed from the airplane on arrival makes it look like you're landing on a printed circuit board. Eight lane freeways interlink vast areas of concrete construction. Tower blocks in the downtown area are dwarfed by the sheer acreage of the metropolis. Cell towers are highly visible on the highways, and nearby buildings. Travel by car, rather than public transport, is the norm.
But in this somewhat different environment, we heard of the same problems and concerns facing mobile network operators. Data traffic growth will require substantial new capacity, and that will be delivered using small cells and (much more controlled) service provider Wi-Fi. Network operators have gone public with substantial investment plans.
Perhaps that was why numbers at this year's Small Cells Americas conference in Dallas this week have doubled to around 350 delegates. The show was abuzz with strong representation from AT&T and many smaller carriers, several vendors from the area (Plano is a hotbed of telecom), and many foreign vendors from both Europe and Asia. About 90% of delegates were from North America, with some 35% of exhibitors/sponsors from abroad. We had representatives from small cell vendors, backhaul vendors, software, silicon and even property development companies.
While residential femtocells were mentioned during the show, it was clear that the focus was around urban metrocells and indoor enterprise femtocells. A range of LTE products were on display, but we also heard that Wi-Fi and 3G have their role.
Setting the scene from the Small Cell Forum
Gordon Mansfield, the new Chairman of the Small Cell Forum and also responsible for deploying them throughout AT&T, commenced proceedings. He recapped the success and scale of (mainly residential) small cells so far – 46 deployments in 23 countries - as detailed in the Forum's latest quarterly market status report. They've just completed a survey of operators where 98% said that Small Cells are important to their future. It seems everyone either has deployed small cells, or has plans to do so at some point. The Forum now has operators in all the key roles (Chairman and 3 regional Vice Chairs) to ensure that operators are controlling the action and direction.
The Forum has formed a partnership with the WBA (Wireless Broadband Alliance) which will concentrate on developing Smart Cells and Wi-Fi, enabling smart, fine-grained traffic management. Today's static nature of traffic management, where Wi-Fi is used wherever available using "blind handoff" will make way for a more dynamic and real-time solution with intelligence required in both network and devices.
Asked if Wi-Fi might be used extensively for small cell backhaul, Gordon thought it had a place in rapid deployment scenarios, but wasn't their first choice due to the uncertainty of managing interference in uncontrolled spectrum.
In-building or out
Jim Parker, also from AT&T, talked through some of the issues of in-building coverage. Over 80% of data connections are made from inside buildings, and the DAS (Distributed Antenna System) technology used in large buildings and stadiums simply isn't cost effective for smaller enterprises. While DAS still has a role to play, enterprise small cells will be adopted much more widely.
He explained that new construction materials used in buildings to make them more eco-friendly was exacerbating the issue. Perhaps the most amusing anecdote was that when the new FCC building was opened, it didn't have good cellphone coverage from any mobile operator inside – people had to stand outside to make calls!
NEC gave a thorough explanation of why DAS isn't suitable for most in-building environments any more. High costs, deployment time and specialist skills are all prohibitive. This was reinforced by one property developer delegate who said he'd been asked to pay the cost of $350K to install DAS plus $150K for basestation equipment per operator. Instead, his residents were using their own femtocells which he also felt wasn't ideal – instead wanting an operator agnostic solution at a reasonable price.
The theme of whether small cells should be deployed in the streets outside versus inside the buildings was picked up by several speakers. Ubiquisys firmly believe that the vast majority of traffic is indoors. Their CEO, Chris Gilbert, said that five nines (99.999% availability) was vital and they now have it (although it nearly killed them to achieve it).
Mike Schabel, Alcatel-Lucent, re-iterated his theme this year that it isn't about the box – practical deployment issues from zoning, backhaul, logistics are far more important. This message seemed to ring true with many operator delegates.
Several speakers see this as the long term requirement. Ubiquisys believe multimode is essential – if handsets have it, why shouldn't small cells.
Qualcomm are committed to multimode small cells. They have assembled an end-to-end ecosystem covering both 3G and LTE, including their acquisition of DesignArt. They believe large numbers of small cells can meet the demands of "hyperdense" locations, with 144 small cells per macrocell achieving 1000x traffic capacity.
There are different routes to this nirvana. I've seen the early pioneering 3G femtocell vendors develop their solutions into 3G/LTE boxes, initially using two boards. I've also seen LTE only small cells (especially Korean) with roadmap plans to add 3G subsequently. Most if not all have Wi-Fi as an option. Arguably most of the mature commercially deployable products today are single technology (i.e. either 3G or LTE), but that could change as soon as by next year. Many people I spoke to thought the timeframe for volume takeup of LTE metrocells was more likely to be 2014, but next year will see growing numbers of real world field deployment as operators get to grips with how to deploy the technology.
Other operator viewpoints
Mexican operator Iusacell have plans for a so-called "maxisenal" femtocell offering using Alcatel-Lucent. They plan to start with 500 metrocells for locations such as subway stations, tall buildings, shopping malls and their own retail points. "Next year small cells will be very big"
The smaller US CDMA networks could also join the party. Wisconsin based Cellcom, having deployed their own CDMA femtocell solution, understands the high initial costs and technical learning curve which would deter other smaller operators. So they now offer this as a managed service to others. They'll setup the E911, lawful intercept, switch routing and billing etc – running the service on an OPEX basis. They reckon it should take about 3 months to bring a new operator into commercial service.
Rick Vergin of Mosaic configures all their femtocells for open access, and "gives them away like candy" because each one installed noticeably improves customer satisfaction.
An Asian Perspective
Rapid LTE rollout was a feature of Korean and Japanese operator presentations.
NTT DoCoMo launched LTE two years ago and now have 23,000 macrocells for 11 Million LTE users. (I estimate that's around twice as many cellsites per user as found in the US for 3G). They forecast a further 12x traffic increase by end 2015, aided by imminent launch of faster tablets and smartphones capable of 100Mbps. They like multimode small cells, especially for use with CSFB (Circuit Switched Fall Back, which handles 4G voice calls through the 3G network).
Softbank, also in Japan, had invested heavily in TD-LTE network capacity, and more recently also in FDD LTE. They've seen "crazy" levels of traffic due to the iPhone 5, which will drive a higher percentage of traffic onto FDD than their TDD network. Having originally planned for 27,000 LTE macrocells by March 2014, they now need to accelerate their rollout to handle the traffic. They also have over 100K 3G small cells installed.
The two main Korean network operators, KT and SKT, continue to battle it out with extensive LTE deployments. This is giving a headstart to their local small cell vendor community. For example, Contela, already having extensive 3G/Wi-Fi public access femtocells in service, has extended their system to handle LTE as well, sharing the same gateway and management systems. Competitor Qcell (visible for the first time at this conference) demonstrated the high throughput of their LTE small cell by simultaneously driving 20 smartphones with heavy traffic in real time on their stand.
China Unicom explained their views of the various LTE and LTE-Advanced features and options, but I found it hard to discern the size or scale of their deployment plans.
The conference focussed on public access small cells, rather than residential or rural, and reflected that (in the US at least), we are still at an early stage.
The event was full of contrasts.
At the headline level, the burgeoning demand for wireless data traffic capacity is clear. The technical solution of small cells is widely understood and accepted. High level investment plans from the major operators have been announced, involving billions of dollars.
In parts of Asia, that's already translated into enormous rollout of LTE macrocells and 3G small cells. LTE small cell products are actively demonstrated and already in trials.
In the US, vendors have a wide range of products and services to choose from, small cells in every conceivable package/format, radio technologies (3G/LTE/Wi-Fi) etc. serving indoor and out.
Service Provider Wi-Fi, which is popular here, looks likely to be extended to form a more integrated and controlled part of the wireless network.
Low cost alternatives for backhaul and deployment which enable the scale are now available to make the business case attractive.
But vendors don't appear to be getting clear sight of decisions (or actual orders for product) being made. Many are being asked to jump through ever more hoops to prove or adapt their product, explain or train their customers, and justify the business case.
It takes time to ramp up deployment of a disruptive new technology such as small cells. This is clearly happening in Korea and Japan today. Progress is less visible here in the US. The industry needs to make some clear technology choices, and get to grips with the dramatically different logistics required to deploy and operate a network composed of mainly small cells.
Some other snippets:
- Askey, a Taiwanese ODM, has licensed Alcatel-Lucent's femtocell software and was showing off it's residential femtocells in quite a funky shape and style
- Small Cell Forum quarterly market report includes a case study from Greece, where small cells can provide free cellular service in designated zones (similar to free Wi-Fi in cafes)
- Taqua thinks that the $300M US government funds to enable internet in rural/unconnected communities could drive small cell/backhaul rollouts.
- Putting cellular service in your business jet would typically cost you around $500K
- iSuppli forecast 1.4Billion smartphones shipped in 2012. Over $42M spent on new infrastructure annually worldwide (more spent maintaining the installed base).
- During June 2012, SK Telecom handled 16,657 Terrabytes of traffic, up from 5,880 in June 2011.
- Uplink traffic is growing as a percentage of overall data. Many more people want to share videos, photos etc. especially at public events.
- ... oh, and the iPhone 5 LTE traffic figures are "really crazy". But you knew that. Did you know that US operators will be subdisidising iPhone 5 sales to the tune of $10 billion during 2012 alone?
Apologies for omitting details of many other presentations and conversations which I didn't have space to include here.
The next Avren small cells conference will be in Korea in March.