3G at CTIA: Market Momentum -- But an Uncertain Future

Verizon continues to lead the way with its CDMA EVDO rollout, which ismore broadly available and offers better performance than Cingular's GSMUMTS system. While recovering from my strenuous daytime activities inTampa, I used Verizon's network to fool a few people into thinking I wasworking rather than vacationing, and it consistently delivered in excessof 300 Kbps of downstream data performance. That's not exactly broadbandby my definition, but it got the job done. Oddly enough, while I was inTampa, Verizon was running TV ads in the area making fun of the spottycoverage of Wi-Fi hotspots and hawking its BroadbandAccess service.Although the company may have the most compelling mobile high-speed datastory to tell in today's market, the $79.99 monthly charge for unlimitedusage is clearly a big obstacle to adoption.

At CTIA in New Orleans, Verizon was singing the praise of Nortel's newEVDO Rev A base stations, which promise downstream peak data performanceof 3.1 Mbps in a single 1.25-MHz channel. The two companies arepromising market trials in 2006. Not to be outdone, Lucent was showingoff new HSDPA gear, the logical successor to UMTS and the biggestcompetitive threat to EVDO, showing actual session throughput in excessof 1 Mbps and latency of less than 100 ms. That's an acceptabledefinition of broadband wireless in my book.

It's one thing to offer base stations, but what about the clients? Myassociate, Frank Bulk, has been tracking recent smartphone productintroductions, and his list now includes 16 such devices that supportEVDO and 14 that support UMTS. Samsung even had a prototype HSDPA phoneat CTIA. Unfortunately, it isn't getting much easier or cheaper toconnect your notebook computer to these networks. Falcom USA wasdemonstrating a USB EVDO. Nortel was demonstrating the Sierra WirelessAirCard 850 for HSDPA networks. Sony Ericsson announced a low-cost CDMAtelematics module, coming to a sports sedan near you.

Sprint made a slew of announcements, many focusing on mobile multimedia,but the one that caught my eye was its new wireless SLA (service levelagreement). I've praised Sprint in the past for offering the industry'sbest wireless data services bargain, a $15-per-month all-you-can-eatplan. But there are caveats. First, that's the price for unlimited useon a PDA or smartphone, though with the right combination of phone andBluetooth, it's possible to use the system as a modem for your notebookcomputer. Second, Sprint's network is based on 1XRTT, a data servicethat delivers, in my experience, a maximum of about 110 Kbps downstream,quite acceptable for downloading e-mail. However, I've found thatSprint's 1XRTT data performance varies dramatically from cell to cell,such that some deliver a maximum of only 55 Kbps on my Treo regardlessof the time of day. And for me, that's the service I have in cells thatserve both my home and our Syracuse University lab. When I raised thisissue with Sprint and asked for a technical explanation, the companyrefused to provide one. Instead, Sprint responded that my performancefell within its advertised data rate of 50 to 70 Kbps. But that's onlytrue if you let Sprint do the math, perhaps across its entire customerbase. Using Sprint's service in the two locations I spend most of mytime yields average performance well below 50 Kbps. Will you do better?Maybe, maybe not.Ironically, Sprint's new SLA is billed as "the industry's first wirelessdata service agreements for business customers." Like most SLAs,Sprint's is filled with holes. While the agreement details addresseddata blocks, data drops and network availability, Sprint included nopromises about throughput or latency. One might assume such a commitmentwas implicit in the company's pledge to provide 99.5 percent system"availability," but there's weasel language in the press release thatleads me to conclude Sprint really means 99.5 percent reliability ratherthan availability. In other words, the company can meet its SLAcommitment by delivering reliably poor throughput and latency.

While there's reason to be somewhat enthusiastic about the present, theindustry faces some critical long-term challenges beyond price andreliability. One of my CTIA sources was Network Computing ContributingEditor Peter Rysavy, who has written numerous feature stories for usabout cellular data networks. Through Datacomm Research, Rysavy hasrecently issued a report that should make carriers take notice. Entitled"Hard Numbers and Experts' Insights on Migration to 4G WirelessTechnology," it includes Rysavy's technical analysis of existing 3G andemerging 4G mobile data technologies along with insights gained througha panel of leading corporate wireless experts. According to Rysavy,emerging 4G technology--especially MIMO-OFDM--will offer compellinglong-term advantages as the industry moves toward higher speeds. In theshorter term, CDMA-based systems deliver excellent efficiency, butcarriers will need to develop strategies for adding MIMO-OFDM to serviceofferings, particularly if Intel makes good on its promise to tightlyintegrate OFDM wireless capabilities in future mobile computingplatforms.

Dave Molta is Network Computing's senior technology editor. Write to him at [email protected]

Read more analysis and opinion from Dave Molta.