New approaches to an old problem
Wireless users, especially in the enterprise and public-safety sectors, have been waiting for affordable, reliable in-building coverage since the first police officer hit the “xmit” key on his two-way radio. Recent technological developments — as well as deployments of nascent fixed/mobile convergence, or FMC, solutions in some areas — now have operators wondering if the wait is finally over.
In September, Sprint Nextel rolled out a consumer-friendly femtocell base station designed to work with the company’s phones and improve subscribers’ in-home performance. The Airave, manufactured by Samsung and about the size of a conventional 802.11 base station, is available in parts of Denver and Indianapolis, and the company says wider availability is on the way. Semiconductor companies, including Analog Devices of Norwood, Mass., also have announced chip-level products for such applications.
Femtocells are small, stand-alone units designed for deployment in buildings with an eye toward increasing network coverage. Such devices initially used Bluetooth signals to communicate with handsets, but most vendors now are moving toward Wi-Fi as the protocol of choice.
The Sprint Nextel system allows customers to use their mobile phone in their homes, routing the conversation through their broadband Internet connection, says Ajit Bhatia, director of product management for the carrier. The company sells the Airave for $50, with monthly service costing $15 for a single handset or $30 for a family plan.
Whether such products offered by Sprint Nextel and other carriers — T-Mobile rolled out a similar, Wi-Fi-based offering earlier this year — will be successful could be determined by the answer to this question: Will individual consumers, who already pay for service that promises to follow them almost anywhere, pay extra for supposedly better coverage only inside their own homes? From a technical standpoint, other important questions loom: What are the implications of large numbers of femtocells operating in a small area, e.g., a high-rise apartment building or condominium? And specifically, what is the potential for interference with other cellular or Wi-Fi traffic?
This last question is particularly vexing for enterprise users anxious to harness the potential benefits of femtocell deployments, including ease of use (one number to reach a salesperson, for instance, rather than separate office and mobile numbers) and cost reduction. Even in many traditional offices, cellular devices are becoming the primary means of communication as employees move around to meetings in different rooms or even buildings.
Most buildings rely on outside cell towers for coverage, but concrete, steel and other building materials play havoc with cellular signals. Even when such signals can be received inside, they may be weak or intermittent. Also, conflicting signals among multiple outside cellular towers can cause devices to “hunt,” constantly switching from one source to another. And users who rely on outside signals still face capacity limitations due to other users on the network.
Femtocells (and their wider coverage-area cousins, picocells) address these problems by providing strong signals throughout a building. In this respect, they have the same aim as distributed antenna systems (DAS), the traditional solution to in-building coverage.
Femtocells and DAS both typically require an 8-10 dB advantage over signals from outside cells for mobile devices to reliably find and stay on their frequencies for voice communications, said Stefan Scheinert, chief technology officer for LGC Wireless of San Jose, Calif. That requirement may increase to 20 dB for data. Such strong signals often require multiple femtocells, but because each cell uses the same frequency, multiple cells often must overlap. This causes devices to hunt for the strongest signal, which in turn degrades data-rate performance. DAS users address this dilemma through careful antenna-location planning and signal-meter work — precisely the labor- and expertise-intensive efforts that femtocells are supposed to eliminate.
Standards limitations also must be taken into account. When initiating the handoff of a device from an outside cell to a femtocell, for example, only a limited number of cell sites — typically 16 — are scanned and measured; in a crowded environment, there may be well more than 16 sites in range. Also, in a system where the outside and femtocell networks use the same frequency band, the cells may interfere with each other as mobile units increase their transmit power to the femtocell. There also is the issue of coverage area — femtocells on different floors can interfere with other users. In an open system, where the femtocells are not managed, such a scenario can quickly spiral out of control, with users on the wrong cells and cells interfering with each other. In a closed system, the units are spectrum-managed, with a remote controller constantly adjusting the power of different cells to maintain network performance.
There are several partial solutions to this problem, but the best way to prevent interference is to use a different frequency for the femtocell coverage, particularly in CDMA deployments. Less-foolproof solutions include using fixed-power options for handsets, which prevent the mobile unit from increasing its power and causing interference.
Backhaul is another problem. Each femtocell requires backhaul connectivity, but the best location of a cell for backhaul may not be the best location for coverage, Scheinert said. And because femtocells may cause interference to outside signals, they often operate on very low power — typically in the 1-10 mW range.
If a building is small enough to be covered by the weak femtocell signal, there’s no problem. But buildings of several thousand square feet — or campuses with multiple buildings — may require high-output power that interferes with outside cells. And an in-building dominant signal can be difficult to achieve, with some construction materials significantly reducing signal strengths. That means high data rates may be available in only some parts of the building.
Combining base stations with a DAS, Scheinert said, can provide the best performance with the least interference. DAS can enable higher loads of femtocells, with improvements in base stations perhaps eventually allowing more capacity than outside cellular services. And because DAS separates the base station and the antennas, coverage can be tailored to individual buildings.
Another path to femtocell deployment was chosen by Meru Networks of Sunnyvale, Calif., in its efforts to bring FMC to Osaka Gas Co., Japan’s second-largest utility. In what the company calls the largest FMC deployment in the world, Osaka Gas first deployed a pervasive WLAN infrastructure capable of supporting multiple data and voice applications.
The quality of wireless voice-over-IP (VoIP) service was of particular concern for the company, said Racha Ahlawat, vice president of strategic marketing for Meru. As a utility, the company cannot allow delays in times of emergency. Ahlawat said Meru’s system allows wireless VoIP in large environments with guaranteed service quality without any proprietary extensions to the client’s network. Meru’s system also treats all physical access points as a single, virtual access point, making the question of handoffs irrelevant, Ahlawat said. Because the devices recognize only one access point, voice clients have seamless roaming with no loss in quality or dropped calls, she said.
Another improvement in the Osaka Gas system is simplified, single-channel operation of access points, which according to Meru eliminates the need for complicated RF site surveys and allows operators to deploy networks as easily as they might 802.11 or other wireless data networks. In addition, contention management and load balancing inherent in the network eliminate the bandwidth-allocation demands common to high-density deployments, as well as many interference issues.
The Osaka Gas installation began in May 2005 and was fully deployed by the end of March 2006. The network uses 800 access points with 72 controllers and serves more than 5000 handsets, Ahlawat said. Besides the wireless phones, the company still has wired phones at 17 of its 49 offices for emergency communications. Ahlawat said Osaka Gas expects total annual savings of about $4 million from its wireless VoIP solution.
In addition to the cost savings, the Osaka Gas network seems to deliver many of the other efficiencies touted by femtocell boosters: Employees can answer extension calls anywhere, anytime, as if they were at their own desks. And critical documents stored on the company’s central server are now accessible anywhere, allowing employees to deliver information to their customers at a moment’s notice.
Whether deployed alone, Wi-Fi style, or as part of a hybrid femtocell/DAS system, femto networking seems to be ready to serve the demands of even exacting enterprise customers. The long-ballyhooed FMC era may finally be emerging.