CDPD is nearing extinction...know your wireless WAN options

Since 1995, CDPD has been one of the preeminent wireless wide area network (WWAN) technologies used for transmitting data to and from wireless devices in "the field."

Today, CDPD has two main attractions--coverage and cost. CDPD is available in most of the places where existing cell phones work, and is one of the few networks that satisfies my three most important criteria for choosing a WWAN: coverage, coverage, and coverage (C3). I'm often asked what the best WWAN is and why. People tend to focus on features such as speed (CDMA 1xRTT's forte) or the graceful co-existence of voice and data (the main strength of GSM and GPRS). While both criteria are important, absence of a signal when you need it most renders them moot.

The other advantage to CDPD is cost. Because it's a standard with which various carriers could comply, the competition between those carriers helped to make the monthly costs manageable. Depending on your service provider, the monthly cost for an all-you-can-eat data CDPD plan ranges from $30 to $50. I've been paying $40 per month for CDPD service from Earthlink.

On the down side, CDPD is painfully slow when compared to the speeds of 2.5G and 3G networks that are coming on line now, such as 1xRTT, GPRS and their successors. Performance hasn't been an issue for most mobile applications. (Whether the majority of wireless application types in use have been designed to work with the available bandwidth of performance or, it just happens that performance isn't a major requirement for most applications, is a good subject for debate.) As someone with fairly horizontal needs (e-mail and the Web), I tend to absorb whatever performance I can get. However, for plenty of other wireless applications --especially verticals -- performance is a non-issue.

The other down side for CDPD is that it's on death row. At RIM's Wireless Enterprise Symposium last week, AT&T representatives explained to me that its nationwide CDPD network will be turned off on June 30 or July 31 of 2004. Verizon, another CDPD provider, will shut down its network in December of 2005. Rumor has it that Verizon would have preferred to decommission its CDPD network sooner, but a contract with the Illinois State Police that guarantees availability until the end of 2005 is preventing that earlier retirement.

Given CDPD's limited future, it doesn't matter whether you're an existing CDPD customer or someone planning to deploy wireless applications for the first time. Now is the time to be considering your options. Provided you're willing to go the data only route, and performance is a non-issue for your users, the two best options at this point are Motient or one of the Mobitex network providers.

Given my C3 requirement, which of the two networks has a better footprint is hard to tell. Each claims to have the best national coverage of any packet radio network. (Neither has anywhere near the coverage of CDPD.) On its site, Mobitex provider Cingular says it is the "only nationwide Packet-Data Network in the country," a claim that Motient's site doesn't seem to corroborate. For example, whereas Motient's coverage map indicates some hot spots around Alaskan cities Fairbanks and Anchorage, Cingular's site reveals no coverage in Alaska.

Ultimately, both networks compete on Metropolitan Statistical Area coverage (MSA)--a statistic that's worth familiarizing yourself with. Motient's site claims the "network provides in-building and on-street coverage to all of the nation's Metropolitan Statistical Areas (MSAs) and extends service to the 520 most populated U.S. cities, touching more than 220 million people." Meanwhile, Cingular's site says the company covers "more than 93 percent of the business population located in 492 Metropolitan Statistical Areas (MSAs) and non-MSAs including all 50 of the top markets, with a total population of 200 million people and over 130 of the top airports in the nation."

I won't even attempt to make heads or tails of the claims because they don't matter. What matters is whether the area your mobile workforce operates in has coverage.

Increasingly however, data-only WWANs are being overlooked in favor of networks and plans that support both voice and data. One reason is that handheld devices like BlackBerries and cell phones, which were previously confined to voice or data, are increasingly supporting both. At its inception, the BlackBerry was primarily a wireless e-mail device that worked over Mobitex and Motient. More recently, certain models work as both a wireless phone and data terminal, supporting a range of applications beyond e-mail such as Siebel- and SAP-based customer relationship management systems (CRM).

The main benefits are obvious. Our lives are simplified by carrying a single integrated device instead of multiple devices (and their chargers, etc.) and having a billing relationship with a single wireless carrier. The downside? Compared to Mobitex and Motient networks, the national footprint for these evolving networks is substantially smaller, which means that for both voice and data, your device will not work in as many places.

Once you go the integrated path, however, your choices narrow to two types of carriers--those that support GSM/GPRS and those that support CDMA 1xRTT. While I've already written one article on the differences, it was interesting to hear the pitch from the carriers at RIM's Wireless Enterprise Symposium.

One issue I didn't mention in my previous article was that devices supporting the GSM/GPRS option (available from carriers like T-Mobile, AT&T, and Cingular) generally work in other parts of the world, making them ideal for international travelers. Another benefit of the GSM/GPRS option is that it provides an important voice/data co-existence feature that's not available on CDMA 1xRTT (available from carriers like Sprint and Verizon). In the GSM/GPRS world, if your device is in the middle of passing data back and forth across the network, and an inbound call comes in, the data exchange is interrupted to accept the call. When the call is complete, the data application resumes where it left off. Verizon representatives admit that if a CDMA 1xRTT device is in the middle of exchanging data with the network when an inbound call comes in, that call will be forced to roll into voice mail. The problem is not solved by the targeted successor to CDMA 1xRTT ,CDMA 1xEVDO.

I say "targeted" because it's not absolutely clear whether 1xEVDO is a step that carriers like Verizon are willing to take. Upgrading the infrastructure apparently isn't cheap. 1xEVDO's chief selling proposition is that it will be significantly faster than 1xRTT. The theoretical limits for both are 2.4mpbs and 144 kpbs, respectively, with the real world experience throughput close to half the advertised speed. Already, 1xRTT, which averages 50- to 70 kpbs, has proven to be almost twice as fast than GPRS. Currently, Verizon is testing 1xEVDO in the San Diego and Washington D.C. markets. The company is trying out Nortel's 1xEVDO equipment in the San Diego market, and Lucent's in D.C.

Verizon representatives said the decision to deploy nationwide will depend on how well the technology is received in those markets. Though no information is available, my guess is that the service must cost substantially more than 1xRTT, or acceptance wouldn't be much of an issue. After all, for those on or considering 1xRTT, 1xEVDO seems like a no-brainer because it's backwards compatible with 1xRTT. That means if you buy 1xEVDO-based client devices, they'll work on 1xRTT networks (though not the other way around). This is important because, if and when 1xEVDO gets rolled out on a wider scale, you'll want your mobile users to have continuous connectivity, even when they roam out of 1xEVDO coverage areas into ones that have 1xRTT. This should be another critical decision criteria when picking between GSM/GPRS and 1x-anything.

According to the Verizon reps, the next step for GSM/GPRS, known as EDGE (Enhanced Data for GSM Evolution), has a theoretical maximum of 384 kpbs but will deliver only 100- to 120 kbps in practice. More importantly, they said, achieving backwards compatibility along the lines of what should be available with 1xEVDO and 1xRTT has yet to be demonstrated. If backwards compatibility is achieved, it will invariably require more expensive client devices that have two radios because GPRS operates on a different radio frequency than does EDGE.

For those considering their WWAN options for new applications or because they have to migrate off of CDPD, cost will also play a role. Whereas all-you-can-eat-data plans are available and reasonably priced for CDPD, the price will go up if higher performance or merged voice and data are on your checklist. For example, with AT&T GPRS, data is billed separately from voice and the highest monthly volume plan, which costs $100, allows for 100 megabytes. Additional data transfer is charged at $ .001 for each additional kilobyte beyond the 100 megabytes monthly.

GPRS carriers generally allow users to use phones like Ericsson's Bluetooth-enabled T68i as a GPRS-modem for data devices on computers and PDAs. CDMA 1xRTT-based Sprint, as another example, provides all-you-can-eat-data with all of its voice plans, but doesn't endorse using one the devices it provides as a 1xRTT modem for PDAs (unless the PDA functionality is integrated into the device) or for PCs. There are workarounds, but they're not endorsed by Sprint. Sprint has a data-only, all-you-can-eat plan for $120 per month, and it requires the purchase of a PC Card-based adapter.

I could go on with more pricing examples, but you get the picture. Expect to pay more if your CDPD migration/WWAN plan includes one of the voice/data networks versus a lower bandwidth data-only network. Cost and features will factor into your decision, but don't forget that at the end of the day, C3 is the golden rule.

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.

Is wide area wireless for you and if so, what's your strategy? Share your thoughts and ideas with your fellow ZDNet readers using TalkBack below, or write to me at david.berlind@cnet.com. If you're looking for my commentaries on other IT topics, check the archives.