re: Wireless Data Transitioning
* DoCoMo ... users who got to try the technology last week were amazed at its speed ... None of the reports mentioned the actual throughput on NTT's new system.
* Compare that with the 6 to 27 Kbps I've been getting with the latest wireless wonder I've been trying out, the Sprint PCS Wireless Web Modem ... The device connects to Sprint PCS' CDMA network, which delivers only 14.4 Kbps, but uses "bandwidth-optimization" (read: data-compression) technology, provided by a San Francisco company called BlueKite, to boost the effective throughput.
* Metricom's Ricochet high-speed wireless technology, which is supposed to deliver 128 Kbps.
* 802.11x ...
>> Tech21; Wireless Waiting Game
The San Francisco Chronicle
June 05 2001
Japan got its first taste of high-speed wireless Internet connectivity last week, but here in the United States users continue to face a range of wireless-data options that can only be described as pathetic.
As you may have seen in the news last week, NTT DoCoMo, the Japanese company behind the fabulously successful I-mode wireless voice-and-data service, began commercial deployment of the world's first 3G wireless network.
For now, availability is limited - only 3,300 Tokyo residents, selected from 147,000 applicants, got to use it last week, and NTT has delayed its rollout to the rest of the country until fall. Prices are high, too - roughly $1 for three minutes online. And the technology is clearly still in its shakedown phase: A software glitch brought it down within 18 hours.
But at least they're moving in the right direction, and users who got to try the technology last week were amazed at its speed, according to news reports. None of the reports I've seen mentioned the actual throughput on NTT's new system, but in principle 3G is supposed to provide 2 Mbps to stationary users and 384 Kbps to users on the move.
Compare that with the 6 to 27 Kbps I've been getting with the latest wireless wonder I've been trying out, the Sprint PCS Wireless Web Modem. The Sprint product, which costs $399 plus airtime charges, is a PC Card - one of those credit-card-size expansion cards for notebook PCs or Compaq's iPAQ Pocket PC handheld (with bulky expansion jacket). The card works with Windows 95, 98, NT and 2000, as well as Windows CE 3.0, but not yet with Windows Me.
Manufactured by Sierra Wireless of Richmond, British Columbia, as the AirCard 510, the card is notable for a clever design feature: A thin, 2-inch-long antenna retracts into the card for safe storage when you're not using it. When you want to go online, you just pull it out, then pivot it to any convenient angle.
But even a giant antenna wouldn't help much with speed. The device connects to Sprint PCS' CDMA network, which delivers only 14.4 Kbps, but uses "bandwidth-optimization" (read: data-compression) technology, provided by a San Francisco company called BlueKite, to boost the effective throughput. Although Sprint's posted specs (www.sprintpcs.com/ccb/business/wirelesseb modem.html) don't mention connection speed, a press release issued jointly by Sprint, Sierra and BlueKite claims that "BlueKite's solution enhances wireless Web browsing speeds by an average of 5x" and that users can expect "speeds comparable to a landline dial-up connection averaging 56 Kbps or higher."
Note, though, that comparable doesn't necessarily mean equal, or anything close to it. When I got online, the software that comes with the card informs you that you're connected at 38.4 Kbps. A BlueKite marketing exec I talked to said he regularly gets 40.8 Kbps.
I would have been delighted to do as well. When I tried to measure my actual throughput, using several online bandwidth meters, I never got above 27 Kbps, unless you count Cnet's report of "infinite Kbps" (I wish!). I've never been able to connect at all from my East Bay home. From The Chronicle office, even though the software reported a strong signal (usually five green bars, the maximum), I can connect only intermittently, and the bandwidth meters typically report speeds below the network's nominal 14.4 Kbps.
Outdoor connections are generally better; the problem there is that the screen on my notebook, like most portables, is barely readable outdoors.
It may be that BlueKite's compression technology works better with actual Web content than with the test data transmitted by bandwidth meters. It's also undoubtedly true, as the wireless companies always say, that "many factors affect connection speed," and some users will surely have better luck than I did.
I'll also admit that I'm a little spoiled, because I have a very fast DSL connection at home and The Chronicle's T1 line at work.
But after all the qualifications, the bottom line is clear: Surfing the Web with this product is painfully slow. For $400, plus Sprint's substantial per-minute fees, I expected better.
On the bright side, the product does work (in principle) anywhere on the Sprint PCS network, which encompasses most metropolitan areas across the country. And the speed may be good enough for users who mainly want to send and receive e-mail, as opposed to browsing the Web.
For real wireless surfing, though, I guess we'll all just have to wait until the carriers finally upgrade their networks. I'm not expecting 3G here for years, but most of the major carriers claim they'll have some so-called 2.5G systems, offering more than 100 Kbps, up and running by the end of this year.
Whether that will actually happen remains to be seen. The severe slump plaguing the whole telecommunications sector obviously doesn't help, and even though 2.5G is supposed to be a modest evolutionary step from current networks, technically speaking, getting it to work is apparently more of a challenge than the carriers realized. Just last week, Cingular postponed until fall the rollout of 2.5G services it had planned to introduce in California and Nevada this month.
Ricochet Alternative
You may be wondering why I haven't mentioned Metricom's Ricochet high-speed wireless technology, which is supposed to deliver 128 Kbps. It apparently does deliver that kind of bandwidth, or something close to it, in some of the 13 metropolitan areas where it's available.
But since Metricom kicked off its high-speed service here last fall, I've done three rounds of testing with it - twice with Metricom's external modem, once with a Novatel Wireless PC Card - in and around San Francisco, and the best I can say is that it's flaky in this area. I never got anything close to the advertised 128 Kbps, and all too often I couldn't get any tolerable speed.
There's also the unfortunate reality that Metricom is teetering on the brink of a financial abyss. It has already burned through more than $500 million - last quarter, it reported a net loss of $186.4 million on revenues of just $3.5 million - and its own executives say cash will run out this summer unless a new investor steps in to bail it out.
I wish the company the best of luck, but at this point, signing on to its service - even if it worked better than it has for me - would be at best a dangerous gamble.
802.11.THIS.THAT: While wide-area wireless struggles, there's continuing progress - despite recent security concerns - as well as ever-increasing confusion in the world of local-area wireless networking.
Last month, the Federal Communications Commission, in response to industry prodding, proposed several changes to the regulations governing the wireless technology behind the 802.11b standard, the basis of the wireless LAN gear showing up in growing numbers of homes, offices, schools and public spaces like cafes and airport lounges.
The changes the FCC is contemplating are complex and technical, but the net effect would be to allow the technology to advance in short order from its current maximum speed of 11 Mbps to as much as 54 Mbps.
That won't make much difference for those who just want to surf the Web without wires from their couch or deck - 11 Mbps is already faster than cable or DSL connections. But it will open new possibilities for the future, like wireless transmission of high- quality video around the house, and it will accommodate heavier traffic on school, office and public-space LANs.
If no major problems arise in the public comments the FCC is now soliciting, the agency is likely to make the proposed changes final this fall. Then the IEEE, the body that sets 802.11 standards, is expected to incorporate them into a new version called 802.11g. If all goes well, new products based on it could begin reaching the market by the middle of next year.
But even if all that happens, the new standard won't have the market all to itself. A competing wireless scheme called HomeRF, designed specifically for consumer use, seems to be fading fast now that Intel, until recently one of its chief champions, has decided to concentrate its wireless-networking efforts on 802.11.
But - we're talking high tech, after all - there are several other standards in the picture. One called 802.11a, which also will support speeds up to 54 Mbps, has been around for a couple of years, and products based on it, from Cisco among others, are expected early next year - just in time to go head-to-head with 802.11g.
While the emerging 802.11g, like the current 802.11b, uses the 2.4- GHz portion of the radio spectrum, 802.11a operates in the 5-GHz band. (Got that?) In principle, that's a big plus for the 802.11a camp, because the 5-GHz band is relatively uncrowded and therefore interference from other devices is less of an issue. Both 802.11b and, in the future, 802.11g products have to contend with interference from a variety of other devices, including cordless phones, microwave ovens and Bluetooth, yet another new wireless data standard that also runs in the 2.4-GHz band.
But 802.11g has other advantages. Products based on it are expected to be a bit less expensive than the 802.11a variety. And while the latter won't work at all with existing 802.11b equipment, 802.11g stuff will, if its proponents are to be believed - b cards should work at their current speeds with new g access points, while new g cards will communicate at 11 Mbps with old b access points.
Just to add to the fun, there's at least one more proposed standard - this one called 802.11e - in the wireless-networking stew . Like the current 802.11b, the e version would use the 2.4-GHz band and deliver data at up to 11 Mbps. What 802.11e would add are several innovations developed by ShareWave, an El Dorado Hills (El Dorado County) company that originally tried to market a proprietary home-networking scheme. In particular, according to ShareWave, its technologies would offer much better support for multimedia, including transmission of CD-quality audio and MPEG2 video.
Panasonic already offers a couple of wireless-networking products based on ShareWave technology: a PC Card called the KX-HGC200 (catchy, eh?) goes for $150, while the base station it requires, known as the Concourse Broadband Networking Gateway, is $250. Netgear, a Nortel subsidiary, is supposed to deliver a line of similar products soon.
For years, ShareWave and its partners have been staging impressive demos of wireless multimedia transmission at trade shows. For now, though, there are no audio or video players that use the technology, and for ordinary computer traffic, it offers no discernible advantages and one big disadvantage: In its current form, it's not at all compatible with 802.11b gear.
If the ShareWave technology really has something to offer that's not provided for in the other 802.11 standards, my advice to ShareWave (not that they asked) is to find a way to integrate it with the versions that already have broad industry support and market momentum. Panasonic and Nortel are big companies, but as Intel's experience with HomeRF ought to prove, even giants have a hard time bucking real marketplace trends. >>
- Eric - |
