My comments about General Packet Radio Service (GPRS) problems in the January 30 edition of Mobile & Wireless UPDATE generated quite a few responses. (To read that commentary, "GPRS Moving Violation?," go the URL at the end of this paragraph.) I want to present a sampling of the messages I received.
Tim Boden reported excellent GPRS coverage in London--both when he's stationary and when he's on a train. (He said the only time a connection is lost is in tunnels.) He added that some European users get good results by using USB cables to connect a notebook PC running Windows 2000 to Global System for Mobile Communication (GSM)/GPRS phones. According to Boden, this solution yields "a steady wireless connection of about 40Kbps." He believes two problems are holding GPRS back: high cost and a lack of roaming agreements. "European businessmen are used to roaming throughout Europe (and beyond) over the GSM network (i.e., phone calls). At present, and mainly through agreements made by Vodaphone, GPRS roaming is limited to only certain carriers, and then only between the United Kingdom, France, Germany, Holland, and Italy," he said
Szymon Slupik wrote from Poland to report that he's experienced good results with a Bluetooth-connected GSM/GPRS phone-and-notebook-PC combination when he travels by train. He reported continuous maintenance of the logical connection, even in areas without GPRS coverage. "When I enter such a spot, the transmission obviously seizes, but then it resumes automatically when the train enters an area covered by another base station. The entire session is preserved, including connection state, IP address, and so on."
On the other side of the coin, Timothy Huber reported trouble with GPRS in Pennsylvania. "This morning, I stopped at Starbucks to pick up some coffee. When I returned to my car, my T-Mobile Pocket PC Phone Edition device received a Push SMS from my company's mobile server and started to synchronize my email, calendar, and so on. Everything was going fine until I started to drive. Then, I quickly received a network-connection error message. I stopped at a signal and restarted the synchronization, which worked until I drove off again and received another error message."
Matt Ward provided some interesting background information about how GPRS works--at least, in the UK. In particular, he noted that "GPRS is a high-latency technology, which means that you'll always suffer from intermittent delays during data communications. GPRS is also a 'lossy' connection, which means that when latency is particularly bad, it's possible to lose data packets, resulting in automatic retransmission of the lost packets. Hence, the user can experience random delays in transmission (i.e., 'timeouts')." Ward said that these problems, as well as others that I don't have the room to delve into here, can "cause havoc when the GPRS transceiver moves from cell to cell (such as when it's used in a car)."
These thoughts leave us with a conundrum: Why do Boden and Slupik (and others, no doubt) have good results on European trains while Huber, Bob Anderson (whom I mentioned in the January 30 UPDATE), and others have bad results in US cars? I have two thoughts about these differing experiences: First, GPRS is a new technology in the United States and offers poor coverage outside major metropolitan areas. The technology might work better as coverage improves. Second, and more problematic, the population density in Europe is much higher than it is in the United States; therefore, in Europe, telecommunications companies (telcos) can more easily provide widespread GPRS coverage. Telcos simply don't need to cover as much ground--particularly because many commuters travel by train. By covering the main stations, a telco has already done a pretty good job. By contrast, to provide adequate coverage for US commuters, a telco must provide coverage on all major highways--a much bigger job.
Software that's tolerant of intermittent connectivity obviously helps. I'm particularly interested in Slupik's experience with GPRS, which gives him a continuous logical connection even when he moves outside physical GPRS coverage. We're not experiencing that functionality in the United States--at least, with the Pocket PC Phone Edition software that Huber's using. I urge Microsoft and its partners to take a close look at these concerns. Presenting the user with a network-connection error message isn't terribly helpful. A more helpful solution would be to display a poor-coverage message, then offer to try again when the user moves into a good-coverage area (or perhaps monitor the signal and perform this functionality automatically).
* 802.11 IN HOTELS
While we're on the subject of wireless connectivity while traveling, I want to pass along a recent experience. Last weekend, I attended Windows & .NET Magazine's annual editorial conference at Denver's Magnolia Hotel, which offers in-room 802.11b wireless Ethernet. I had anticipated that functionality, so I brought along Symbol Technologies' Wireless Networker CompactFlash (CF) card for use with my NEC Computer's MobilePro 200E handheld PC (H/PC). The card detected the network but wouldn't give me a connection. I called the hotel's technical support line, and a representative told me that the only way to use the wireless network was to install software (available only for PCs) that would provide its own protocol stack for access to the network.
That's ridiculous! The whole point of a standard protocol such as 802.11b is to assure that any device can participate. Undoubtedly, the hotel and its wireless provider implemented its 802.11b service to provide a revenue stream (the charge was a high $11 per day, but I would have paid), but couldn't they accomplish it without the requirement to install proprietary drivers? This necessity limits network access to those who have a full-blown notebook or (presumably) a Tablet PC and are willing to install the required software. I'd like to hear from readers who've had similar experiences.
* MOBILE DEVICES IN EMERGENCIES
With the country again on alert against possible terror attacks, now is a good time to review the capabilities of mobile devices in an emergency situation. Cell phones played a huge role in rescue operations during and immediately after the September 11 attacks, and sometimes text messages got through when voice calls didn't. If you support mobile users, you might want to remind them of this fact.
One problem with using cell phones in an emergency is that if everyone tries to make a call, cells can become congested, and no calls will go through--even from "first responders" and other key personnel. To help with this problem, the White House's National Communications System (NCS) has established a Wireless Priority Service (WPS) program. WPS is available on the T-Mobile network in a variety of major metropolitan areas. (And if the T-Mobile system happens to go down completely, satellite phones can be used as backup.) WPS users can place priority calls, which will "queue for up to 30 seconds and grab the next available path," according to an NCS briefing document. A similar program called Government Emergency Telephone Service (GETS) is available for prioritizing wireline calls. If you're a "first responder" with a legitimate need for WPS or GETS, contact the NCS ( http://www.ncs.gov ) for more information.