By Ian Sugarbroad, President and CEO, LGC Wireless*
Twenty years ago, few people would have believed that cellular phones would be more common than wristwatches, but they are. Now wireless devices of all kinds are bringing big changes to the enterprise. Wireless LAN deployments are growing, smartphone use has skyrocketed and at least one-third of business users are relying on their cell phones as their primary communications devices. With things the way they are today, we can begin to see a future in which everything is wireless. The question is how do we get there? In this article, we’ll look at some current and future trends in wireless voice and data networks for the enterprise, and then explore a vision of tomorrow’s converged wireless office infrastructure.
Wireless solutions today have grown up as responses to specific applications such as data service or voice. Cellular has been the primary choice for voice service, and WiFi has been the popular solution for data service. But as with any network operator, enterprise I.T. managers want to integrate all services onto one cost-efficient network to reduce maintenance costs and streamline management.
To fulfill this role, the wireless system must support all voice and data applications, giving appropriate priority and quality of service to each user. It must support any number of users anywhere in the enterprise, and it must integrate internal and external voice networks to form a seamless bridge between the enterprise PBX and carrier networks. Based on where we are today, there’s no one wireless technology that fulfills all the requirements.
WiFi: Fast and Cheap, but Not Carrier-Friendly
WiFi technologies such as 802.11a/b/g dominate the enterprise when it comes to data service, and companies will eventually upgrade to faster new standards such as 802.11n, which delivers up to 200Mbps. WiFi technology is popular because these Ethernet-based systems are relatively inexpensive and I.T. managers understand how to integrate wireless networks into their wired Ethernet LANs.
However, most WiFi deployments to date have been for hotspot coverage only, and most only support data services. There are several technical issues inherent in WiFi that pose barriers to broader penetration and application support, including capacity, scalability and ease of roaming.
Capacity: WiFi networks were initially designed with the idea that one access point (AP) would serve a small group of users. Ethernet’s CDMA/CD collision avoidance protocol is designed to prevent two users from accessing the same AP at the same time. As the number of users increases, WiFi’s collision avoidance mechanism impacts access, particularly when there is a mixture of slower 802.11b users and faster 802.11g users.
Interference: WiFi APs must be deployed using alternating channels to avoid situations where a user device is “seeing” identical channel signals from two APs at the same time. It is frequently difficult and expensive to plan for and minimize such interference in large deployments. WiFi equipment vendors offer sophisticated coverage planning tools as well as the ability to regulate AP output power to minimize such interference. Some newer WiFi technologies use a licensed frequency in the 5GHz band, but mainstream WiFi remains an unlicensed frequency that will always be subject to interference—not just from AP channels but from other devices using the same frequency band, such as cordless phones and microwave ovens.
Roaming: Since WiFi networks were initially designed to permit access with a single AP, users must re-associate with a new AP as they move from one AP’s coverage area to another’s. This need for re-association disrupts connections for voice calls. Many WiFi vendors are now using pre-authentication technology to eliminate the need for re-association, but between this issue and the capacity problem, voice is still a very difficult service to support on large WiFi networks. Although there are many dual mode (WiFi/cellular) phones on the drafting tables now (and some are even available), having access to a WiFi network doesn’t eliminate these fundamental problems.
While WiFi blossoms, other developers are working on mobile WiMax. The 802.16e, or 4G mobile standard approval process is well underway, and Sprint has already committed to rolling out a mobile WiMax network that promises data transmission speeds of 6-10Mbps.
Cellular: Carrier-Grade, but Not Cheap
Cellular networks use licensed frequencies and were designed to deliver service to tens of thousands of users in a given area, so capacity and frequency interference are not the major issues they are with WiFi. While initially designed to support voice only, all of the major cellular providers now offer modem cards and data plans that provide connectivity for any smart phone or laptop user in the carrier’s coverage area. In addition, major laptop manufacturers such as HP and Dell are now offering cellular modems as a “built-in” option for some of their models.
But 3G cellular data is still in its infancy (at least in the United States), 4G is just beginning trial rollouts, and many say neither will ever overtake WiFi as the data carrier of choice within the enterprise. There are several roadblocks to adoption, including bandwidth, coverage, ease of deployment, carrier-based management, and cost.
Bandwidth: Today’s 3G network technology supports real-world data rates of 800Kbps—1Mbps, far less than 802.11b’s real-world 5-6Mbps. WiMax will raise the bar to more than 6Mbps, giving it the same throughput as 802.11g, but of course this will be dwarfed by 802.11n’s throughput, which will be available more quickly.
Coverage: Many commercial buildings do not receive uniformly high-quality indoor cellular service. While service may be good enough for occasional voice calls, it is not good enough to support data services. In-building distributed antenna systems (or DAS systems) remedy this problem, but enterprises will either have to purchase these systems on their own or negotiate with their carriers to install them.
Ease of deployment: Since in-building cellular systems tie directly into a carrier network, the carrier is involved in the deployment. In carrier-funded DAS systems, carriers install and manage the whole network and even in enterprise-funded DAS deployments, the carrier must provide a signal repeater or base station that connects the DAS to its network, along with a link to a wireline backhaul connection from the building to the carrier network. While the physical details of such deployments are no more complicated than deploying an enterprise-scale WiFi network, the carrier’s involvement makes them more complex and time-consuming.
Carrier-based management: I.T. managers like to own and exclusively control their communications equipment, whether it’s a network switch, a PBX, or a wireless network system. With their link to a carrier network, cellular systems deny I.T. managers their full ownership of and control over the system.
Cost: Cellular voice and data services require carrier service charges, and to date cellular data service in particular has been deemed too expensive for broad deployment in the enterprise. Mobile WiMax is expected to change this, but it won’t really get started until 2008. Today, most companies are happy to rely on external WiFi hotspots for employee connectivity outside of the office.
Both WiFi and cellular have drawbacks in terms of supporting a do-it-all wireless network, but cellular technology’s shortcomings will ultimately be a lot easier to overcome. No matter how much work WiFi developers put into working around the capacity and roaming problems and no matter how many dual-mode WiFi phones come on the market, there’s no getting around WiFi’s use of unlicensed spectrum that will always invite interference—interference whose sources could well be beyond the enterprise’s control. With their licensed spectrums and an architecture designed for widespread scalability, mobile carrier technologies are much more likely to carry the day.
But if cellular is to assume the role of one-stop wireless technology, it will need some changes. Fortunately, those changes are now on the horizon.
One Big, Happy Wireless Network
The do-it-all wireless network must combine the best of both worlds. It should deliver the plug-and-play installation, ubiquity, and cost-effectiveness of Ethernet-based WiFi. It should also have cellular networks’ massive scalability, superior voice quality, ease of roaming with transparent handoffs, and carrier network connectivity with integration between the IP PBX and the carrier network.Converged IP technology: IP is the future of networking for all applications, and the converged network will be IP-based. Cellular carriers are already adopting IP-based softswitches to support higher scalability with new services. We can expect that this new generation of softswitches will soon appear on single blade servers with capacities sized for an enterprise rather than the public network. These call servers (together with standard media gateways) will create the IP connectivity between the enterprise PBX and the carrier network while controlling the cellular base stations in the enterprise. Similarly, these base stations will not be the old style base stations designed for the macro network, but pico base stations—compact products that support higher packet data rates and talk to the call server over IP. It may be a long time before cellular networks match the raw data throughput of WiFi networks, but it’s not a question of which has more bandwidth – it’s a question of whether or not cellular networks can deliver enough bandwidth to get the job done.
User-installable DAS: Cost and time of deployment for cellular networks will be reduced with new DAS systems that are easier to deploy than today’s WiFi systems. Remote antennas already look and install just like APs (using Ethernet connections and similar form factors), and soon we will have pico cell base stations that fit in an I.T. data center rack. Because frequency planning and site surveys are minimal with licensed cellular frequencies, enterprises can build networks that reach every part of their facilities and support any number of users at a lower cost of deployment.
Low-cost backhaul: Tomorrow’s pico cell base stations and call servers will use standard DSL lines for backhaul rather than T1 lines, reducing the cost of this piece of infrastructure by roughly 75 percent. Pico cells won’t be appropriate for very large facilities, but they will bring high-performance mobile coverage and capacity to small and medium-sized buildings.
Lower service costs: Cellular providers already offer voice service plans that include no-cost in-network calls (such as between employees in companies with a corporate service plan). With the greater cost efficiencies of running all-IP networks, carriers will also bring down the cost of data service plans to the point where the benefit of ubiquitous on-site and off-site access combined with seamless roaming will easily outweigh WiFi’s alleged “free” calls (which in any case aren’t free, given the cost of deployment and management).
One-call carrier integration: In the future, enterprise-friendly softswitching, base stations and pico cells will offer fast and easy integration with the carrier network. Once the enterprise IT department installs a DAS and pico cell base station and connects the base station to a DSL line, one phone call to the carrier will activate the service.
Converged, IP-based wireless networks will ultimately support any application for any number and density of users inside the enterprise. The vision may take a few years to become reality, but moving to such convergence is in everyone’s interest. Vendors and service providers already know how to make convergence happen – now it’s just a matter of watching it unfold. //
*This is a Contributed Article. The opinions expressed above are those of the author.