Fixed Wireless Business

But How Much Does It Cost?

Setting up a broadband fixed wireless access network sounds fairly simple, right? Cheap too. I mean, you can buy 802.11b access points at CompUSA for a few hundred dollars. And what else is there?

by Gerry Blackwell [April 2, 2002]

Hmmm. Where should we begin?

Yes, offering fixed wireless services using 2.4 GHz license-free radio equipment can be a good business opportunity for ISPs. And it's not rocket science. But to do it successfully does require an investment of brain cells, time and cash.

How much of the latter? That's what we start to find out in the first of this multi-part series. To help us sketch out a rough WISP start-up budget, we asked two guys who between them have built more than a few fixed wireless networks.

James Portaro is chief technical officer (CTO) and co-founder of NeTeam Corp., a broadband fixed wireless systems integrator based in Akron, Ohio but active across the eastern half of the country. NeTeam built one of the largest campus wireless networks anywhere—if not the largest—at the University of Akron.

Fixed wireless pioneer Jack Unger, president of Los Angeles-based Wireless InfoNet Inc., has built "a couple of hundred" wireless nets since 1993. His company now provides mostly consulting and education services to WISPs. Unger is working on a book for the Cisco Press about how to build successful fixed wireless networks.

Men don't ask
We asked Portaro and Unger to help cost two types of wireless networks: access networks for ISPs who want to offer broadband services using wireless—whether existing dial-up operators or pure-play WISP start-ups—and campus or industrial networks for enterprises or institutions.

The first order of business was identifying the major cost components for ISP networks. Pure-play WISPs, of course, will need to purchase and configure routers, switches and servers and establish a conventional ISP network operations center (NOC). We focused on just the wireless components.

There are three broad areas of cost: upfront network design and testing, establishing antenna sites, and purchasing and installing network equipment.

Costs in the first two categories, Portaro points out, will depend to some extent on location—whether the setting is urban, suburban, or rural/small town.

The first broad category, both agree, is crucial. It includes site surveying, network design, equipment selection and reliability testing. Unger characterizes them as "the non-obvious costs that are time and labor based. We find many people going into the WISP business are not fully aware of them."

Site surveying involves both physical surveys to determine where to place wireless POPs for maximum line of site access to target customers, as well as surveys of existing wireless activity that could cause interference. There may be other WISPs in the area, more and more enterprises are deploying 802.11b wireless bridges outside, plus there are still some ham radio operators at 2.4 GHz.

"If you interfere with a ham operator, your stuff has to come out, no questions asked," Portaro points out. "If others are communicating on a bridge in the area, your network may be slowed down because of RF congestion—it becomes a quality of service issue. So the research has to be done on the front end."

How much does it cost? We tried to nail down costs per stand-alone wireless POP—i.e. for each antenna site. Up-front costs depend in part on how wide and deep the coverage and how intense the RF activity. "It takes a lot of legwork to do the testing," Portaro notes.

His estimate, based on what his company would charge: $5,000 to $15,000 per POP. That includes network design and equipment selection.

Unger's estimate, based on establishing a wireless access network in a primarily suburban setting—a small city near a large population center —includes only the site survey and is much lower: $2,000 to $3,000. In Unger's experience, new WISPs want to do their own equipment selection. He recommends allowing at least two weeks—that is, 10 business days—to complete this step.

This raises an important point about costing wireless networks. If you hire a NeTeam to build you a turn-key wireless network, you could theoretically assign a dollar value to every part of the process—or one big number for the whole project. But most ISPs will do at least some of the work themselves, and probably should.

So you have to calculate either the cost of hiring someone full-time with wireless expertise, or the opportunity cost of freeing up existing management and staff to learn about wireless. If it's the latter, you can add the cost of books, seminars, travel, etc. Hiring wireless expertise may be difficult and expensive because the skills are still rare—though growing less so. Think $100,000-plus per annum.

It's worth pointing out one other step here, though it comes a little later in the process, because it too involves often hidden time and labor costs. Once the network is installed, you need to invest time in reliability testing to make sure everything works as it should—that customers in all parts of the coverage area will actually get the bandwidth you're promising.

"I've found that not allocating enough test time is the most common mistake new WISPs make," Unger says. "They figure since the equipment comes in on Monday and they install it on Tuesday and Wednesday that by Thursday or Friday they'll be hooking up paying customers who are going to be satisfied with the service."

"In the best case scenario—in a tiny little town where there is no other RF around and you've done a faultless wireless and site survey—it could happen that way. But 90 percent of the time it doesn't."

Unger's recommendation: allow at least two weeks of elapsed time to do the testing and work all the bugs out of the system. "It's always time very well spent," he says. "If you try to cut corners, it can easily come back and bite you later."

Next week: establishing antenna sites.

—End