Why bother with Wi-Fi when CDMA can do?

I have a question about Wi-Fi (802.11) wireless LAN technologies: Why bother?

If you followed any of my recent trials or tribulations with Handspring's Treo 300, you know that I stumbled upon an application from June Fabrics called PDANet that allowed my notebook to get access to SprintPCS' 2.5G CDMA 1xRTT network at a throughput rate of about 70 to 80 kbps. From a hotel room in Washington, D.C, accessing my email actually worked faster via CDMA 1xRTT than it did via dial up. Can you imagine that? While 70 to 80 kbps doesn't sound like a lot, you'd be surprised at just how sufficient this is for 99 percent of what you do (unless all you do is watch streaming video).

If I can get 70-80 kbps and my hotspot is all of SprintPCS' and Verizon Wireless' coverage area and the "line" is secure (which it is, compared to Wi-Fi) and streaming media isn't a priority for me, why bother with Wi-Fi?

Wi-Fi coverage is spotty, and there are virtually no roaming agreements. If I subscribe to T-Mobile to get coverage at Starbucks, chances are I won't get coverage at the local airport where the provider is Wayport without paying for another subscription. Meanwhile, there's a chance that a single CDMA 1xRTT network covers both and maybe even my office and my home. The more I think about it, the more I wonder whether I can do without a Wi-Fi radio altogether. That way, I don't have to bring two cards with me everywhere I go, and install or fiddle with any software that helps me to manage which wireless network to connect to. Keep it simple stupid: Isn't that what they say?

Want another reason? Recently, there's been talk about voice over IP (VoIP) for Wi-Fi-based handhelds. I've played around with a Microsoft VoIP technology from the company's labs called Portrait that works on PCs and handhelds. Using a PocketPC-based device such as an HP iPaq with a Wi-Fi card, Portrait makes it possible to have a VoIP conversation with someone else using Portrait on another, similarly equipped handheld connected to the Internet. Considering the cost of VoIP (free), the come-uppance of Wi-Fi along with its availability in public settings raises some interesting possibilities. Take the same Wi-Fi/VoIP-equipped handheld into a public hotspot and, from a functional point of view, what's the difference between that and a cell phone? In that scenario, could Wi-Fi present a threat to the existing wireless carriers?

But given the uphill climb that's necessary to turn VoIP into a killer application for both the Internet as well as wireless handhelds (thus turning the handhelds into data- and voice-capable terminals), there is little choice but to reconsider the CDMA-connected handheld. The wireless voice provided by CDMA-based, data-capable phones is as good as wireless voice gets and the calls are practically free when you look at the number of minutes you get with just about any package. Combine that with a better hotspot and built-in security, and CDMA continues to look better than Wi-Fi.

What about the GSM(voice)/GPRS(data)-based 2.5G alternative to CDMA? Currently, the real world data throughput for GPRS is running at about half (30-40 kbps) of what people are getting out of CDMA. At that speed, most networked applications are noticeably degraded. Until that situation improves, I can't recommend the GSM/GPRS option--except for world travelers (GSM/GPRS is the standard most everywhere else in the world) and for people who want their data sessions to be interrupted for inbound calls. That feature is not yet available for CDMA.

Another feather in CDMA's cap is its future. Currently, most CDMA deployments are of the 1xRTT flavor. However, Verizon Wireless is prepping two cities --- San Diego and Washington, D.C. --- for a CDMA 1xEV-DO rollout. What's the difference? Oh, just a few hundred kilobits per second. Compared to 1xRTT's theoretical peak throughput of 144 kbps, 1xEV-DO's peak is 2.4 mbps. But like 1xRTT, which delivers far less than its advertised peak, 1xEV-DO is predicted to deliver somewhere between 400 and 800 kbps. That's faster than most cable or DSL connections.

I asked Verizon Wireless' corporate communications executive director Andrea Linskey whether her company was having any thoughts about how CDMA (1xRTT or EV-DO) could be a better strategic wireless target than Wi-Fi, even for people working in their offices and homes. In addition to CDMA 1xRTT-based connectivity, Verizon is rolling out public Wi-Fi hotspots in places like airports through a partnership with Wayport. Although Linskey said she believes that Verizon "could go pure CDMA," she made it clear that the carrier was not trying to intercept businesses leaning in the direction of Wi-Fi in hopes of getting them to reconsider that decision.

"For now," said Linskey, "we don't view Wi-Fi as a threat to the CDMA business. We see it as complementary. In campus settings, businesses may be likely to launch with Wi-Fi as the in-building, wireless alternative. Out of the buildings, they can choose to go with Express Network." Express Network is Verizon's brand name for its wireless wide area network (WWAN). "Carriers like AT&T and Verizon are hedging. While we are waiting to see what becomes of the de facto wireless standard, we're preparing to roll out in 2004 a product that makes switching between a company's native Wi-Fi network and Express Network automatic for someone [with radios for both]. It will detect and automatically connect to the right network. Before then, however, customers will have to manually select which network to use."

Today, Verizon Wireless offers an all-you-can-eat data plan for approximately $80 per month. It requires a CDMA 1xRTT-based PC Card or one of Verizon's phones with Verizon's Mobile Office Kit. Ideally, the way to go is to share one radio --- presumably a phone --- for voice and data communications. One radio means one account and keeps things relatively simple. The best way to connect a device to a phone is through Bluetooth. But, currently, Verizon Wireless isn't reselling a Bluetooth-based CDMA-enabled handset, which means that any time a connection to the CDMA network is desired by a PDA or notebook computer, a hardwired connection is required between the device and the phone. Is a Bluetooth-enabled phone in the offing? "Verizon Wireless has extremely stringent quality requirement for its handsets," said Linksey. "We're testing some units now but have nothing to announce."

Meanwhile, Verizon's biggest CMDA 1xRTT competitor, SprintPCS, has announced it will be offering Sony Ericsson's Bluetooth-enabled T608i starting this month. However, that offering could be in jeopardy: Sony Ericsson announced recently that it would be discontinuing CDMA phones in the U.S. In response to an inquiry about what SprintPCS plans to do should Sony Ericsson stop production on the T608i, company spokesperson Suzanne Lammers told me, "We're still looking at our options, and should be able to give you an answer fairly soon. It's possible that we may carry only limited quantities through select channels, but nothing has been confirmed yet."

		As part of our year long Wireless that Works special report, we want to hear your tales from the frontier that highlight the successes and failures of wireless technology deployments. To have your company considered for a "Wireless that Works" case study profile on ZDNet, send us an e-mail. We need the name, phone number, and e-mail address of a business contact at the company, and ideally a technical contact as well.

"1xEV-DO, if it works, could be a confrontation to Wi-Fi," said Bakhshi. "Once 1xEV-DO comes on line, it only has to come at 400 kbps of throughput to satisfy most users. The question is whether the underlying backhaul to the Internet will be able to handle the load. It could be like sucking a watermelon through a cocktail straw." In light of the fact that Verizon Wireless is just now piloting 1xEV-DO in only two cities, Bakhshi doesn't see that confrontation happening any time soon. In addition, Bakhshi told me, "Neither CDMA nor GPRS are very good at penetrating buildings. In fact, outside, they're still not particularly good."

But, in the same way the WWAN technologies exhibit weaknesses indoors, Bakhshi questioned the business model for all of these public Wi-Fi hot spots that are showing up everywhere. "Maybe in specific areas like at airport gates," he said, "where a lot of people need to get on just before the plane takes off. In situations like this, the cellular technologies don't scale very well. The radio resources of the cell network are constrained. For example, if 15 people make calls from the exact same spot, do they all get on? No. So, in certain spaces where there's lot aggregation of people, maybe that's where it makes more sense to improve the local radio resources through Wi-Fi."

Like Verizon Wireless' Linskey, Bakhshi thinks that, for the time being, the best solution for people who need anytime, anywhere access is to run a blend of both technologies. Echoing Verizon Wireless' recognition that switching between two wireless technologies needs to be seamless, Bakhshi said: "A CDMA 1xRTT / Wi-Fi combination card would be a nice thing to have along with the software that makes it simple to switch between the two." Such combination cards do exist. For example, last year, GTRAN Wireless introduced a dual mode CDMA 1xRTT/Wi-Fi card using Intersil's Wi-Fi chipset. On the software side, solution providers appear to be responding to users that want seamless switching between the two radios. In addition to Verizon Wireless' promise to address the issue by next year, companies like NetMotion Wireless [link: www.netmotionwireless.com] are providing software solutions like Netmotion Mobility and partnering with wireless network providers such as GSM/GPRS-based AT&T.

Ultimately, whether CDMA or any other 3G technology will be able to kick Wi-Fi's butt may be determined by cost. Right now, all-you-can-eat data plans from Verizon Wireless and SprintPCS cost about $80 per user, per month , How will that compare to the per user total cost of ownership for Wi-Fi? One cost and administrative headache that's eliminated by a pure-CDMA strategy is the securing of the local air. When you outsource your wireless networking to a wireless carrier, you're also outsourcing at least one element of your security. In either case (Wi-Fi or CDMA), going wireless in a public scenario requires a virtual private network if someone has to get behind a corporate firewall, so that cost is a wash between the two options anyway. But with Wi-Fi, the additional measures needed to secure the air require a measure of expertise, expense, and time that some would rather avoid.

Putting a finger on the per user TCO for Wi-Fi isn't quite as simple because there are certain breakpoints in company size where the infrastructure costs take a sudden leap. Gartner analyst Phillip Redman told me, "I don't think you can say [what the per user TCO] is definitively. It has to be based on usage." Even so, Redman has developed two scenarios: one where the WLAN is a primary network for small offices and a second where the WLAN is a complementary network for large offices. Whereas the per user annual TCO for the former (assuming a small office of 45 users with occasional additional part timers) was $10,000, the same for a mid-size to large office is about $3,052.

Regardless of the comparison, the question I keep asking myself is this: For $960 per year, would I be willing to put up with lower bandwidth if it meant being able to cruise around a bigger hotspot without having to manage two separate radios or worry about roaming between providers at Starbucks and the airport? When I asked Verizon Wireless' Linskey what she thought the much higher performance 1xEV-DO service might cost, I was surprised to hear her say, "We haven't priced the data-only service yet. But my expectation is that it will be similar to the things you're seeing on 1xRTT right now."

If that's the case, sign me up. If everyone else is so gung-ho about Wi-Fi, more power to them. If there aren't more CDMA believers like me, perhaps the congested scenario that IDC's Bakhshi speaks of will never happen. Of course, if you're located in the sticks where none of the 2.5G or 3G networks are floating about, then Wi-Fi connected to a local DSL, cable, or T1 backhaul may be your only choice. So, Wi-Fi isn't dead. But I'd love to see something like 1xEV-DO take off.

CDMA over Wi-Fi? Am I off my rocker or am I onto something? If your warehouse or facility had great CDMA 1xRTT coverage from a provider like Sprint or Verizon Wireless, would you consider going with that for all your mobile users instead of Wi-Fi? Especially since they could roam beyond the confines of your campus? Use TalkBack to let your fellow ZDNet readers know what you think. Or write to me at david.berlind@cnet.com. If you're looking for my commentaries on other IT topics, check the archives.