As mobile data services and applications continue to increase and mobile Internet use grows with the rise in smartphone sales, operators are struggling to provide enough bandwidth to meet the demand for both voice and data services. The issue of offloading data traffic away from congested 3G cellular networks has become one of the most discussed and debated topics in the wireless industry.
The central challenge is 3G cellular technologies were designed as the next generation of the older cellular voice networks, long before any sizeable mobile Internet really existed, and it was very difficult to predict the sheer volume and extremely fast growth in mobile data consumption introduced by smartphones and other popular Internet-enabled mobile devices.
While the needs of voice services are mostly bound by the number of subscribers served, data services traffic is characterized by a large variety of applications and use scenarios; consequently the data traffic average volume and its variability are vastly larger than those of voice services. Although some look to more advanced cellular generations for salvation (4G, LTE), it is commonly agreed that these technologies cannot scale up enough to support the exponential growth in mobile data bandwidth demand due to the limitation of physics, and they will take years to deploy in any case.
To support adequate data service, many more cellular base stations/access points are needed. The challenge is obviously even greater indoors, since radio signals must penetrate buildings to provide indoor data coverage. Since the cost to operators' business is linearly related to base station deployment density and since ARPU cannot increase proportionally (willingness of subscribers to pay, flat-rate plans), data services pose a great financial challenge to operators, in addition to the logistical hurdles.
To break this linear dependency between deployment density and cost of service, operators must resort to out–of-the-box approaches that meet a number of criteria: They have to be independent of real estate cost, reduce service costs by eliminating "truck-roll" support, minimize need for a new backhaul, minimize changes to network protocols and interface, preferably do not need new spectrum, minimize cost and headaches to users, enable operators to assure service quality and provide a simple way for billing.
The wireless industry is currently considering two options: femtocells and Wi-Fi. Both solutions have merits that will be discussed below.
Advantages of Wi-Fi include the capability to hold a much larger amount of spectrum, while in femtocells there may be spectrum conflicts between the macro network and femtocells, and between neighboring femtocells. Generally, spectrum is a very precious commodity and should be used sparingly. While the maturity of femtocell technology still remains to be seen, the advantages and disadvantages of Wi-Fi are pretty well known by now.
On the other hand, on the surface there are a number of advantages of femtocells over Wi-Fi. The wireless interface in femtocells is identical to the cellular network, and control functions are identical, thus allowing handsets no need to have an additional Wi-Fi radio unit enabled. Femtocells use only licensed operators' spectrum, which can therefore avoid "foreign" radios.
Wi-Fi conversely uses uncontrolled spectrum, which may theoretically make service susceptible to neighboring radio transmissions. In addition, femtocells allow for easier seamless roaming, though some modification is needed to accommodate the different backhaul. Finally, and perhaps the biggest advantage of femtocells, is they are a natural extension of the main cellular network, thus allowing them to support most of the current services provided by the mobile operator. Conversely, Wi-Fi protocol is relatively lenient on quality of service, and therefore enabling multiple services with variable qualities of services may be challenging.
Business considerations of choosing one or the others are still up for debate. Who pays for the backhaul of femtocells? What if family or businesses use multiple carriers? What about the issue of net neutrality? What if that backhaul (DSL, cable) gets clogged?
So Why Wi-Fi?
Fortunately, most Internet-capable user devices now come Wi-Fi capable. Since the main challenge of indoor operation is related to Internet access, the terminal's Wi-Fi can provide for a good ecosystem.
Most existing cellular networks are well provisioned for indoor voice service; the cases of non-sufficient coverage for voice service are rather rare, hence the issue of seamless roaming between indoor to outdoor is questionable at best. If voice roaming is not an issue, we can avoid the complexities associated with "voice continuity" between Wi-Fi and cellular, hence the Wi-Fi solution can be further simplified.
Dense unplanned radio deployment produces spectral conditions that are very similar to unlicensed spectrum conditions. In the case of femtocells, this could nullify the main benefit of licensed spectrum utilization. Consequently, one may be better off using unlicensed spectrum to begin with. Since Wi-Fi is inherently designed to operate over unlicensed spectrum with high co-channel interference, using Wi-Fi will be substantially less risky.
Since Wi-Fi uses unlicensed spectrum that does not exploit the precious operator licensed spectrum, and since the amount of unlicensed spectrum is far greater than the typical operator's licensed spectrum, much more traffic can be served using Wi-Fi.
Wi-Fi networks are already deployed all over; almost every residence, office, coffee shop and other public places where Internet access exists have Wi-Fi. Finally, instead of asking subscribers to buy, install and learn to operate a new device (femtocell), indoor service challenges can be solved by a "software only" approach that requires very little (if any) of the subscriber's attention.
Shimon Scherzer is co-founder and chief technology officer at WeFi.