Microcells, oDAS and picocells: Small-cell architecture to stem wireless data deluge

AT&T distributed antenna system das palo alto californiaWhen it comes to the future of mobile networks, small is in. Wireless vendors since the beginning of 2011 have been tripping all over themselves introducing small-cell architecture strategies as a solution to users' increasing demands for mobile data. Soon, many of the technological tricks that operators have already used to boost wide-area network capacity--such as increasing the spectral efficiency of the air interface--will be exhausted, network experts agree. Small cells then become the answer.

A recent survey from Infonetics Research found that 58 percent of the global operators the firm interviewed plan to deploy small cells by the end of 2011, while 68 percent plan to do so in the future.

And while many operators acknowledge that shrinking cell coverage is an answer to the network capacity crunch going forward, many vendors still have work to do when it comes to constructing small cells, backhauling their traffic and mitigating their interference with the macro network.

Alcatel-Lucent's lightradio cube

Alcatel-Lucent's lightRadio cube

Big vendors have jumped in the fray with new product offerings, some of which still need to be developed further. Alcatel-Lucent (NYSE:ALU) introduced a miniature base station called lightRadio. Huawei is working on microcells while Ericsson (NASDAQ:ERIC) is building integrated antennas and radios to reduce the footprint of equipment. Nokia Siemens Networks recently announced Liquid Radio, which uses distributed antennae and virtualized baseband processing to provide a highly distributed architecture built around small cells and miniature base station designs.

Het-Nets--heterogeneous networks using small-cell architectures that complement the overall macrocell network--are a major focus for the future of AT&T Mobility's (NYSE:T) network, Kris Rinne, senior vice president of architecture and planning with AT&T, said during last month's "Path to 4G" event hosted by FierceWireless.

"LTE spectral efficiency will only get you so far. You can only build so many traditional cell sites. There needs to be a new way of thinking and designing networks," she said. "Since a significant percentage of growth is in high dense areas, we need smaller and dense solutions to achieve the full promise (of higher data speeds) ... A major focus for us is developing a small-cell strategy and expanding that as options into different areas."

AT&T is embracing small-cell network architecture in the form of microcells, outdoor distributed antenna systems (oDAS), and picocells at the street level. For instance, the operator is proposing to build some 80 new small-antenna tower sites on top of utility poles across downtown Palo Alto, Calif., in a bid to bolster voice and data capacity in areas that experience heavy data traffic. AT&T plans to use oDAS, but the city has yet to approve the plan.

The notion of small-cell architecture has existed for some time. DAS has always been used to bolster coverage indoors, but oDAS increases capacity outdoors. Femtocells and picocells have existed too, but have yet to be put to use on a wide scale inside a macrocell network as a strategic way to bolster coverage and capacity.

"These are not new things. What always broke the back was the backhaul infrastructure, enclosure aesthetics and the heavy battery," said Alan Solheim, vice president of corporate development with backhaul specialist DragonWave. "The key is to get all of those elements in one package that doesn't break the bank."

In other words, it's not economically sound for an operator to increase the number of small-cell nodes 25 times and see a corresponding 25 times increase in the cost of ownership.

The problem with small cells

Three primary problem areas associated with small cells are mounting, interference and backhaul.

Mounting If operators are looking to deploy thousands of small-cell nodes, they can't incur heavy costs in mounting the devices throughout an area. Operators may have to reach agreements with cities to deploy nodes on lamp posts and other structures, and the nodes need to be as plug-and-play as possible so virtually anyone can mount them and plug them in. AT&T has been waiting on Palo Alto to approve its oDAS plan for several months.
Interference and management

In addition, operators will have to grapple with interference concerns and network traffic policies. Small cells underlying bigger cells create more interference and hand-off problems because devices may switch back and forth based on where the stronger signal is. The interference issue is already being addressed by the femtocell community by incorporating elements of self-organizing networks, which would allow the nodes to automatically discover each other to avoid interference.

Het-Nets are a primary goal of 3GPP Release 10 of the LTE standard, but network equipment remains several years off, noted Yankee Group analyst Ken Rehbehn in a recent report. "With capacity-focused spectrum for LTE deployments found at 2.6 GHz, small-cell solutions are an essential component of many MNOs' LTE deployment in congested areas," he said. "However, Release 10 network equipment remains several years off. Absent Release 10 features, MNOs may fail to get the desired level of integration into the larger radio plan. Compromised interference management and hand-off support will result."


Another major tripping point is backhaul. Operators need to have solutions in place capable of aggregating, networking and switching backhaul from the thousands of nodes on the street level. To put a single connection to each node would be an expensive proposition. Moreover, operators are increasingly using fiber for backhaul, but fiber is not available to light poles and other high places beyond the traditional cell tower, noted Aircom CTO Ricky Watts.

However, a number of vendors are clamoring to solve the problem. Last month Bridgewave introduced a new family of millimeter radios called PicoHaul, which is targeted specifically at small-cell deployments. The idea is to use a small-form radio to connect small cell sites back to cellular towers, where traffic could then ride on the fiber links that already serve the macro network. Bridgewave said it has completed a successful trial of the setup with an unnamed North American operator.

Other vendors such as Aviat and Exalt Communications are touting point-to-point wireless links for small-cell deployments.

Finding small-cell solutions

BelAir's LTE picocell

BelAir's LTE picocell

BelAir Networks may be the furthest along in the small-cell game. The Wi-Fi vendor, which sells equipment to AT&T and cable operators, has already deployed a Wi-Fi/3G picocell that is being used by a cable provider that is wholesaling the connections to an unnamed wireless operator. The service is carrying live commercial traffic over 25 nodes using the cable operator's outdoor hybrid fiber-coax lines as backhaul.

BelAir also recently introduced a Wi-Fi/LTE picocell that can be used by cable providers and wireless carriers.

Greg Williams, senior vice president of business and corporate development with BelAir, said the vendor has solved many of the issues that have made small-cell deployments problematic in the past.

"We've been essentially working on metro picocells for the last eight to nine years," Williams said, referring to the fact that outdoor Wi-Fi networks are similar to wireless network small cells. "We've got the cost down to a place that makes sense."

While the picocell can be mounted on the strand of a cable plant, it can also be placed underground to connect to underground fiber or fiber-coax lines for the power and backhaul it needs. Or the picocell can be placed on light poles and other fixtures. Williams said the benefit of using a cable strand is that it eliminates the need to secure additional zoning since the picocell can be cut into the strand.

Where there isn't any fiber feed, BelAir incorporates mesh networking to create a short-hop backhaul network that the company said won't degrade the signal as it passes through hops, Williams said.

When it comes to the deployment of nodes on the cable strand, Williams said the process literally takes minutes. "One node can take about 15 minutes to cut into the cable network, and the person installing it doesn't have to have a lot of knowledge," he said.

Given the lack of zoning requirements and fiber backhaul, cable operators are likely to find a lucrative wholesale business in the small-cell architecture game. Recently CableLabs introduced a technical project that involves picocells and other small-cell architectures that cable operators can use with their own wireless deployments and cellular backhaul offerings.

Microcells, oDAS and picocells: Small-cell architecture to stem wireless data deluge