Receive new posts as email.
This site operates as an independent editorial operation. Advertising, sponsorships, and other non-editorial materials represent the opinions and messages of their respective origins, and not of the site operator. Part of the FM Tech advertising network.
Entire site and all contents except otherwise noted © Copyright 2001-2010 by Glenn Fleishman. Some images ©2006 Jupiterimages Corporation. All rights reserved. Please contact us for reprint rights. Linking is, of course, free and encouraged.
Slashdot points to a brief article on boosting power output from a Linksys WAP11: Power ain't everything, but it might be a useful hack for those trying to link locations cheaply without using expensive antennas.
Seattle Times on Boeing's Connexion service: the local angle with details on layoffs and questions about the future.
I'm back in Seattle after a week of sucking watermelons through a 56K modem straw. I'd forgotten how spoiled I am by broadband.
Boeing's Connexion service gets go-ahead: Feds approve further testing of the in-flight broadband service. A competing offering from Tenzing Communications was slated to use Wi-Fi in-flight rather than wired special connections as Boeing is apparently planning. No word lately from Tenzing, however.
Convergence of FireWire/IEEE 1394 and 802.11a? This somewhat obscure story appears to draw a connection between the high-speed FireWire standard marketed by Apple (and by other vendors as i.Link or IEEE 1394) and Wi-Fi5 or 802.11a. Adding to my confusion is the mention of 802.11e in the article, which, to my knowledge, is focusing on scheduling issues: how to get different kinds of data assigned to different degrees of priority. Clarification, anyone?
As networks grow up, they lose their borders. Can anyone remember back to 1994, as Microsoft announced its upcoming Microsoft Network which would only work with its Blackbird authoring system? Microsoft said that when they launched MSN in 1995 it wouldn't conform to HTML standards nor would there be more than an email gateway to the Internet. How quickly they recanted.
Even the sleeping giant, when aroused, recognized that networks profit when connected, not when sundered. All of Microsoft's subsequent efforts, rebuffs, recantings, and courtings have centered around the sheer might of a larger world that they can't keep their users from wanting.
A more recent example is cited by Simson Garfinkel in the January 2002 issue of Technology Review. The article is not yet linked online, but in it he discusses how a lack of standards among cell messaging systems has prevented growth of this useful and fun service in the States. In Europe, billions of messages are sent each month at 1 to 10 cents each.
I've spent a lot of time writing about Boingo Wireless in the last week or so because their emergence signals the next step in the maturity of public space wireless. Up until Boingo announced their deals and their software, each network stood mostly alone. There are hundreds of thousands of Wi-Fi networks in the U.S., and only a fraction are public. Of those, only a tiny fraction were interoperable or linked.
Boingo forged individual deals with each network they wanted to partner with to create consistent pricing and access, but otherwise appear to have left technical details and operational details alone. (Forthcoming best practices reports from the Wireless Ethernet Compatibility Alliance for wireless ISPs should pick up the slack on what standards are needed; other groups will certainly help wISPs conform or test to those standards.)
I believe my reports have been mistaken for extolling Boingo qua Boingo: rather, I am excited about the message that Boingo sends. Wi-Fi is open for business. Wireless ISPs are willing to cut the kinds of deals necessary to create more traffic on their networks to reach, ultimately, a repayment of their investment and actual profit.
I don't deny the profit motive was in effect prior to Boingo's appearance, but it's a new world in which all players that can agree on reasonable terms can use the same commons.
Doc Searls, one of my favorite Web and print writers, has engaged me and others on the issue of Boingo. He's concerned about the Windows-only focus and the potential proprietary nature of a single client that locks out other platforms and other uses.
My reaction is that the deals that Boingo forged will allow other players to enter the market with a lower bar. In fact, not every traveler or businessperson or home user or hobbyist wants everything that Boingo provides. This opens the potential for other companies to build their own virtual networks using what will become for-hire infrastructure run by Wayport, Surf and Sip, and others, and aggregated by hereUare and others.
The next logical step in evolution will be an independent group that helps aggregate without offering ISP service. That is, they will sign the contracts to link networks and provide a uniform API or software interface for developers who come in Boingo's wake.
Another development on this front is 802.1x, a uniform and secure authentication protocol in progress from the IEEE. 802.1x will allow the same method of user login from many different systems and platforms, and may allow those who come after Boingo to avoid Boingo's huge cost in building a Swiss army knife of authentication. Their software can talk to anybody's login procedure, but 802.1x may sweep all of that away. It just takes time and updates: all APs, operatings systems, back-end systems, and Wi-Fi cards would need to be upgraded to support 802.1x in large or small ways. (XP supports an early version of 802.1x, but not the ratified version because it hasn't been ratified yet.)
This is all grist for the mill: we'll see if other players emerge and turn Boingo's great idea into a teeming marketplace in which Boingo was first and possibly becomes premiere, instead of flying solo.
NASA to add Wi-Fi into space suits: reduces cost from $20 million to $19,990,000.
Happy holidays, you all. As I sit in deepest, darkest West Hartford, Connecticut, I send my best wishes for the holidays and the new Gregorian year to all readers of this newsletter. And, in fact, to everyone in the world, if I could only find them.
I wanted to address a few more issues related to Boingo before we return to the maelstrom of the rest of the industry. There are several distinct problems that average travelers face - even if they're sophisticated business users - when they try to access wireless service on the road. Part of my excitement about Boingo is that I believe if they properly implement everything they plan that they are poised to solve these problems.
1. Finding out where service might be. Currently, branding is relatively hidden and confusing when you enter a facility that may have Wi-Fi. Although a few hot spot finders exist (see list at left) and software like that supplied by iPass provides a list of their partner networks, there's no comprehensive directory. Boingo's location finder includes free community networks along with their partners. I would predict very easily that after MobileStar's current situation is salvaged, Boingo will partner wtih more than 95 percent of public for-fee network locations.
2. Discovering networks in the vicinity. Although a variety of client software supplied by different Wi-Fi card vendors offers discovery, or a scan to see which networks are visible, Boingo's sniffer is more sophisticated. It's even better than Windows XP's huge leap forward in managing Wi-Fi connections. Better still: if Boingo knows the network, it identifies. Not even XP can do that now.
3. Managing multiple locations. Although modern operating systems recognize that people travel, I still find it a hassle to move from place to place and maintain the settings I need. Apple's Mac OS 9 had a superb Location Manager which allowed you to create collections (time zone, TCP/IP, sound, etc.) that you would select based on location. OS X and Windows XP have some options, but they're not terrific. Boingo's profile manager is full-fledged for Wi-Fi, including WEP encryption key management.
4. Storing WEP keys. This seems like a minor point, but the client software I've used for Wi-Fi cards doesn't typically offer enough profile space, nor does it truly password protect the WEP keys: it just doesn't make them visible after they're typed in, but a cracker could easily extract them. Windows XP is certainly better at this task, but Boingo, by integrating it in the profile manager, makes it even simpler.
5. Securing transmissions. The number one fear among those with any anxiety about wireless Internet use is other folks sniffing their private correspondence and other secrets. Because public hot spots are almost universally run without encryption (or if they are, the WEP key is easily available when you want to get on the network), your data is essentially being posted on the local bulletin board. Boingo offers a tunnel out of the local network using strong encryption out to their public operations center far away on the Internet. This ensures your data doesn't shoot out in the open, as it were, until it's in a more remote environment. (This service is initially free, but will be part of a paid plan for non-monthly users.)
6. Sending email on the road. Can I see a count of hands for people who have had to spend more than a few minutes on every trip reconfiguring or messing with their email settings to just simply send outbound email from any connection they're hooked up to, even if it's just a dial-up account? I've spent about 15 minutes on this trip already over a few days. This isn't part of the client software, but it's just as important. Your Boingo account logs you into an authenticated SMTP server, which is available from any point on the Internet. Period.
7. Paying a consistent price. Sky Dayton, Boingo's founder, said their focus groups showed that price point for service was highly elastic. That is, $5 vs. $10 for a business traveler is irrelevant if the service works and meets other needs. But it's certainly true that businesses want predictable pricing: your company might approve a $24.95 per month account with Boingo, but they almost certainly won't approve $200 a month through 30 different providers unless it's mission critical work. Especially because each connection and account is separate: separate credit card choices, separate Wayport and MobileStar accounts, etc. Even better: the unlimited rate offers a unique opportunity for salespeople. Connect at the airport in the morning, the arriving airport when you get there, the coffee shop on the way to a client, the hotel's network, and then back at the airports again. That could be six separate connections or as much as $50 to $60 in today's market. Boingo's offering that for $74.95 per month.
8. Keeping track. Then there's just the issue of knowing when, where, and to what you connected. The current generation of software falls down in this support.
We'll move on to broader Wi-Fi issues in the coming weeks as Boingo's service comes online; their beta software went up a couple of days ago. But I wanted to define the set of circumstances under which this company comes to market and solves the main problems that I've seen with on-the-road Wi-Fi.
I'm currently accessing the Internet by sticking my Apple iBook in the window of the hotel my fiancee and I are staying at in Manhattan. We're on Lexington and 51st, so we're out of the densest areas of coverage. But there is an open AP: no authentication, no information about the company running it. They have at least 768K DSL, maybe a T1. Since I'm here over a weekend, I haven't worried about affecting their operations. Still, it's an odd thing to connect to a stranger's network. Kind of like driving into a neighborhood you don't know, and plugging your RV into an electrical outlet you find at the curbside in front of a house.
I've received a fair amount of pushback on my Boingo article from Wednesday from readers and fellow Webloggers on how superfluous Boingo seems. Why not just do what they're doing in software via a browser window, they ask? Why lock into a specific proprietary software package thus creating the potential for a non-standard network?
This misses the mark due to what I would term a completely understandable blurring of the lines between the Web and the Internet. The Web runs on top of the TCP/IP stack, a layer of protocols that allows programs to break data into pieces and send it to a known address over any kind of medium (Ethernet, Wi-Fi, dial-up, ATM, etc.).
Boingo's software dips down below where the Web (even with Java) can go into protocol layers below applications. By using these lower layers, Boingo is employing standards to tie together disparate wireless network operators. The point is: anyone can do what they're doing; they just have to do it.
My understanding is that Boingo's agreements are non-exclusive with each carrier. Sky Dayton, Boingo's founder, said to me repeatedly that the goal of any network should be distinct from the goal of a service provider: networks should load traffic; service providers should acquire customers and give them the most ubiquitous footprint.
Further, Boingo's software doesn't require any proprietary software installed at the service provider. In fact, that's part of its charm. The service provider continues using its existing authentication system; Boingo interfaces with that through its magic innards which can talk to many, many different kinds of login systems as well as its own. Boingo's software is a universal translator.
Boingo is not a software platform locking users in. In fact, it's a standards-based tool that relies on only standard protocols to ease the process for its users. Other companies will be able to come along, using different or identical protocols and still transit TCP/IP data on the Internet. They'll have to negotiate their own contracts with wireless infrastructure providers, but that will be the case in any vision of the future of Wi-Fi.
As a quick rundown, Boingo uses the following standards: NDIS 5.1 (talking to cards to sniff the network), RADIUS and related authentication protocols, VPN (not sure if they're using PPTP or IPSec, but I would bet on IPSec, the better of the two), authenticated SMTP (to login to send outbound mail whether using their software or not), and TCP/IP.
The face of public space wireless service changes Thursday as Sky Dayton, founder of the dial-up Internet service provider Earthlink, launches Boingo Wireless. Boingo will build no hot spots. Instead, they are aggregating the network infrastructure of other companies and wrapping it up through a single user account, a single bill, and a single set of pricing. Dayton summarized the new firm's thrust: "Boingo Wireless is a non-infrastructure wireless ISP."
Their initial launch includes over 750 hot spots; Dayton estimates Boingo will encompass 5,000 by the end of 2002. The company did not announce partner networks, but Dayton said that their partners currently represent about 90 percent of hot spots outside of the MobileStar network. (MobileStar filed for bankruptcy in early December 2001; VoiceStream had a proposal to acquire its assets in early January 2002.)
The Boingo model places a software interface on top of the myriad of login and authentication systems used by individual wireless ISPs (wISPs). In interviews conducted over the last two weeks, Dayton said that the software handles and hides the interaction with its partner networks to provide a seamless login. The company has been working on the software since spring.
By requiring client software, initially available only for Windows, Boingo offers a variety of features in one bundle: single user login, WEP key management, Wi-Fi network profile management, preferred network priority, VPN (virtual private network) service to Boingo's public servers, quality of service (QoS) tracking, and connection logging.
Boingo also throws in authenticated SMTP mail service, which allows outbound email service anywhere on the Internet.
A Macintosh version of the connection software is planned for 2002, but the company did not want to issue a prediction for delivery. (Dayton himself is a committed Macintosh user.)
Pricing has initially been set in three tiers: a pay-as-you-go model, which requires a free account setup, at $7.95 per connection (up to 24 hours in a single venue); a medium-usage package offering 10 of these connections for $24.95 per month, and $4.95 per additional connection; or $74.95 per month for unlimited connections. Dayton said that there is no bandwidth cap nor any surcharge for use at any partner location, including hotels.
Phil Belanger, VP for marketing at Wayport, alluded to this kind of charge as part of the panel I moderated on public space Wi-Fi at the 802.11 Planet conference in November. In response to a question about working with iPass's pricing model, Belanger said that Wayport had renegotiated all of its hotel contracts in the last several months to allow a variety of pricing and revenue sharing with hotels.
This new pricing model should ripple through independent networks and partner networks who also run their own account plans. Dayton said, "The pricing that we helped establish in the last nine months as we've been putting all our deals together is very simple: it's not per bit, it's not per minute, it's 24-hour periods."
Although other network aggregators and resellers exist - including hereUare and NetNearU in the specific Wi-Fi space, and iPass and GRIC in the corporate roaming area - Boingo occupies a unique niche. hereUare and NetNearU are working to aggregate small networks, including single hot spots, into a larger whole, then offer back-office billing to those venues. iPass and GRIC started in the dial-up business, making it easier for corporate travelers to find a dial-up, and later wireless or wired connection wherever in the world they are, and pay a specific metered rate for access.
Boingo, on the other hand, wants to brand itself on top of infrastructure, leaving the building to its partners, and taking customer service, billing, fee settlement and roaming, and software development onto itself. The software is what differentiates Boingo from any other firm offering access, including iPass, which has the most sophisticated runner-up. Dayton said that Boingo would certainly want to work with aggregators, although he would not specify which firms Boingo initially had agreements with.
I have not yet had a chance to use the software, but I have seen a preview of its features and functions. It is worth a walkthrough of each of the elements. Each part of the program solves one or more problems for the roaming Wi-Fi user.
The software includes both a network sniffer and a location finder. The sniffer requires card drivers that support NDIS 5.1, the newest version of a protocol that abstracts communication between an operating system and drivers, allowing more robust, consistent, and simple interchange. Vendors are rapidly deploying updated drivers because of XP's support for them. The sniffer can show all networks and their signal strength in the vicinity, and brands any partner networks with a Boingo label (see screen capture). The location finder includes both Boingo partners and free community network hotspots that have allowed Boingo to list them. The software will constantly update its list of locations, checking at each connection, and uses XML-based structures to make downloads and installs quicker and more modular.
Connecting to a Boingo partner requires clicking a button. And that's it. Real-world experience will verify this ease of use, but this is the quintessence of the software and Boingo. Solving the login problem was their first and biggest challenge.
Christian Gunning, the firm's director of product management, said in an interview last week, "Every one of those carriers does authentication differently. Some of them use RADIUS [a standard ISP user login system], some of them don't, some of them use certificates." Dayton added, "We've got this very complex authentication token methodology that we've worked out" that handles each network's individual requirements.
The software also includes a Boingo VPN (virtual private network) client. This is a killer feature, and one I've been hoping a company would introduce. The VPN uses strong encryption to protect data before it leaves the user's machine until it arrives at the VPN server, which decrypts it. Boingo's VPN server sits on the public Internet, so that encrypted traffic is sent over the local wireless network, and through the intermediate network points before it emerges from Boingo's operation center.
This VPN client overrides security weaknesses and compromises that might allow a wireless sniffer or a network sniffer to extract user information before it leaves a local network node.
A future version of the Boingo client will also provide a pass-through VPN service for customers whose home networks already use VPN. This service will connect the user's machine with a static IP from Boingo's operation center so that a VPN user can connect even from networks that use NAT to assign non-routable Internet addresses for local machines.
The Boingo client uses Windows security to password-protect WEP keys for separate networks, and bundles a profile manager as well, similar to that found under Mac OS 9 or Windows XP. The password protection doesn't just protect a plaintext file; it uses the built-in Windows messaging-style encryption to lock the contents down in full.
The integration helps manage these profiles better, and the profiles can be reordered by priority, so that in an area with multiple networks, the priority affects the order in which networks are connected to. Gunning said, "If your office is right next to a Starbucks and you're competing with them for a signal," you just re-order your office network to a higher priority.
The company also decided to solve a traveler's most annoying dilemma: configuring outbound mail service. Most ISPs don't offer authenticated SMTP service or fully support the XTND XMIT option for POP that allows a login to transmit outbound email. Boingo includes its own account-based authenticated SMTP server, which can be used whether on or off its network of partners. This, along with the VPN service, will be available initially for free as part of even pay-as-you-go accounts.
The software displays a log of connections and network information to the user, but it also records a variety of parameters for quality of service monitoring. Dayton said that Boingo will use this information unconnected with user profiles "to first get quality of service on all the various networks." He said, "We learned this at Earthlink" that you don't know what's happening on the system side when people can't get through. "In this case, we're logging everything that happens from the user's perspective."
In focus groups Boingo conducted to better understand business traveler frustrating, Boingo found an almost universal sentiment that broadband would influence a travelers choice of flights and other venues.
For instance, because San Francisco International has no Wi-Fi network but San Jose has full Wayport coverage, Dayton said that 97 out of 100 polled said they "would fly to San Jose if they can." Also, "they'll stay in hotels if they can get high-speed access."
Dayton drew deeply from his Earthlink experience in building Boingo, including the notion of a separate software client. Earthlink early on shipped a tool called TotalAccess, an automatic network configuration tool that also had a list of phone numbers that could be updated - remarkable in a day of manual SLIP and PPP tweaking. (Before that, Dayton recollected buying copies of Adam Engst's landmark Internet Starter Kit book, removing the floppies from the back, individually configuring the software on them for a particular user, and then sending it out to that new user.)
Dayton said that he learned in creating Earthlink that building a physical network is different than running a data layer on top of it. He wanted to split the infrastructure from billing and customer interaction. "Earthlink didn't go out and build modem networks and POPs - we left that to the UUNets of the world," he said.
A related issue was his analysis that no single infrastructure company could dominate. "Even if I had a billion dollars and set up thousands of locations, I could never in my network have a completely ubiquitous footprint."
Instead, he turned to creating partnerships with existing firms who were looking to increase traffic over their networks. "The economics are such that if you own a fixed asset like a network all you're concerned with is loading that network," Dayton said. "The No. 1 problem in the public 802.11 space in terms of financial viability is network loading: getting traffic, and getting revenue.
"Anybody who incurs the underlying infrastructure costs, no matter how much capital they have, big or small, whatever, at the end of the day, they have to rationalize this depreciation, they have to load networks."
Dayton brought over several former top-level Earthlink technology managers. His board includes Stewart Alsop (who gave me my first top-tier freelance writing assignment at InfoWorld back in 1994 when he was editor-in-chief), and John Sidgmore, the former CEO of UUnet. One of the company's advisors is Dave Farber, former CTO of the FCC, and a renowned expert on many technology issues, especially where spectrum and government cross.
Dayton said, "A lot of the little details which separate success and failure we've just spent the last year building. That's all we did; we're specialists."
Paul Boutin tried to suggest to Slashdot this Boingo story, but got beat by an unrelated Earthlink item of interest. Earthlink is offering wireless Internet starting in Atlanta. It's unclear how long Earthlink has offered this service; ironic that Sky's old company hits the top of Slashdot with a wireless story the same night Boingo news breaks.
RSA Security and Hifn along with the 802.11 Patch WEP: internetnews.com coverage, News.com coverage: If everything in these stories and the RSA press release are correct, we could have a quick and easy update to WEP that makes it actually robust enough to trust as a simple first-line of defense. It still won't be the right method to rely on, but it will actually meet its initial purpose. The information makes it sound as though most firmware will support the revisions, meaning current devices can become compatible with just a flash. This new method ensures that vastly more RC4 keys are used based on the same shared secret (the WEP key) so that the complexity of cracking the key goes up to some large order of magnitude more difficulty. The shared-key problems remain: updating keys, managing them, keeping them secure, etc. (Read more about WEP and other problems with Wi-Fi's security profile in Wireless which now needs to be updated.)
Slashdot thread on WEP update: the best Slashdot thread I've read in a long time. Well-informed discussion by a core group of knowledgeable people contributing to an overall understanding.
I'd tell you about it, but then I'd have to... Folks, I have some great news that I can't tell you about as much as I'd like to. I can't even hint at it, besides telling you I can't hint at it. Come back Thursday morning when I release a great story that should make everyone happy just in time for the new year.
Solstice Suspension: I'll be on vacation for several days over the next two weeks, and I don't expect much news, either, as everyone regroups from a difficult year. I appreciate the vast amount of advice and information readers and friends have contributed to me in running this Web site, and I am thrilled to be writing about this whole subject.
InfoWorld CTO asks Santa, Wi-Fi Wherever?: Chad Dickerson, InfoWorld's CTO and penner of his regular (and well worth reading) CTO column, asks the bearded man in a red suit for ubiquitous Wi-Fi because, well, because it works! I'll be good next year if you give me wireless access to the Internet from my laptop from all the places where I find myself sitting unproductively.
This humble Web log (or blog) has been nominated for the best technology Web log of 2001 . Feel free to vote. The award comes from Scripting News, an ur-blog and blogfather and connected at the hip with Userland, a content-management software maker that also hosts Web logs.
What she said: European governments suck capital from eager telcos resulting in wireless data disaster: I've been saying this since spring, and the groundswell of support continues to grow for those of like mind. The European telcos, suckered in by a burgeoning economy and the end of old rules, bid $140 billion on 3G (third-generation) cellular licenses for frequencies throughout Europe. When the bubble burst, the telcos found themselves with licenses for a technology that doesn't quite exist, an infrastructure that can't support it, and lots of debt.
Good coverage from News.com's Ben Charny on profitable wireless ISPs around the country: this article succeeds where other fail by finding the small providers who, with a handful of employees and dozens to hundreds of customers are turning a buck.
Bluetooth gain traction: Bluetooth starts to take off.
hereUare Communications's JumpStart program to end at year's conclusion: As hereUare had originally stated, its JumpStart program which offered free access to a number of its partner locations (including those run by WiFi Metro) will end Dec. 31. The company sent out this email this afternoon:
hereUare Communications would like to thank you for your participation in the JumpStart wireless pilot program. The program was a great success and we hope that you have enjoyed the free wireless service! Moving forward, we will be expanding the pilot program to selected JumpStart users during Q1 for a one-time fee. You will receive additional information in the upcoming week. Please note that your current JumpStart User ID and Password will be valid until December 31st.
One of the trends spotted at the 802.11 Planet conference was the end of free: most of the representatives of networks that weren't yet charging or were offering sign-up deals said that in 2002, they'd be charging their actual rates. Demand had risen, knowledge spread, adoption upticked.
GRIC to bring Wi-Fi roaming to hundreds of hot spots internationally: This article doesn't mention specific partners, but GRIC (like iPass) has created an effective business in helping business travellers have a consistent available dial tone or wireline connection at a known price. The U.S. part of this network comes from NetNearU a back-office billing and service aggregator. Aggregators are rising in importance rapidly as the industry coalesces into cooperation. One of the messages at the 802.11 Planet conference was certainly that more deployment by more vendors fosters greater adoption and usage rather than reducing opportunities for each WISP (wireless ISP).
Study shows no 802.11b, Bluetooth inteference: this contradicts other information I've heard from colleagues (who have put both Wi-Fi and Bluetooth cards in the same laptop) and people in the industry. Nonetheless, it's good news: perhaps newer devices test better. The next refresh to Bluetooth should include the IEEE 802.15.2 solution which will allow adaptive hopping to avoid in-use frequencies.
Microsoft adds Bluetooth support by mid-2002 in XP: This is probably the right timing. Many cell phones and PDA models will have integral Bluetooth. PC and PCI card prices will have dropped to a tolerable level, and new PCs might have Bluetooth prebuilt or as a simple non-build-to-order option on standard models.
Bluetooth and Wi-Fi will co-exist serving their own purposes: a nice message of co-existence from the Bluetooth Developers' Conference as reported by EE Times. It sounds like the message is finally getting through to both sides: different purposes, different devices. I'm on the fence. I saw Bluetooth makers in the last year veer away from their plans to try to make Bluetooth more heavily overlap in WLAN territory, while 802.11b vendors were talking about Wi-Fi as taking away some of Bluetooth's purpose.
It still seems to me that many of the purposes for which Bluetooth was designed, a fully enabled WLAN device would serve better. Cell phones already have TCP/IP stacks in them; why not produce a Wi-Fi chipset that put that cell phone on a network, not just make it available for ad hoc use?
Likewise, when I hear about Wi-Fi in every pot, I think, Bluetooth was designed for low power output, low power usage, short-range signalling. Wi-Fi was not. Cutting Wi-Fi down to size (as the article notes) would take time and money, and Bluetooth will certainly be entrenched by the time that could happen.
The bottom line issue for any device is: does it need to handle traffic or is it just relaying information? A PDA can do both. A cell phone, mostly relaying. A computer is sending data back and forth; a keyboard, just passing it along.
I know there are only 3 non-overlapping channels in the b specification, but find it hard to find out the "dangers" of having overlapping channels. At my university, some of the buildings required more than 3 channels, therefore we had to overlap some of the Access Points(ie. we use channel 1 in the basement of the libary and channel 2 on the 4th floor).
Is this degrading my network without me knowing it? If so, how? I haven't had any complaints from the student body.
If you want to read the absolute and accurate truth about how cell carriers and their ilk are coping with integrating Internet offerings in their phones, and how other companies are succeeding or failing at similar missions, read Alan Reiter's new Web log.
I heard only 45 minutes of Alan's talk at 802.11 Planet recently, and he has a unique combination of technical savvy and market (and marketing) comprehension that buoy each other up. Many marketers don't have the chops to see how the tech and the market interact; many techies don't have the insight to understand how consumers and businesses respond to product offerings. Alan has both.
One of his great statements from 802.11 Planet about the real achievable speed of some current technology: "If there are absolutely no molecules in the air at all and you are making physical love to the transmitter" you can achieve maximum full speeds.
USA Today's rather long and generally excellent piece on Wi-Fi: USA Today's two reporters on this piece did an excellent job of explaining Wi-Fi in words of three syllables or fewer, and avoided a lot of techie cant in explaining its utility and appeal.
One very weird error, though, based on what must be a misunderstanding. Another problem: There still isn't a Wi-Fi standard, so some of today's products might not work with others bought a year from now. That's expected to be ironed out soon, through action led by WECA and its 125 member firms. In fact, as readers of this forum know quite well, the WECA certification program has been in place for some time and is extremely effective. Virtually all devices for sale are certified as compliant with the Wi-Fi specification. Perhaps they confused the new 802.11a (Wi-Fi5) spec which is being finalized right now for certifying those devices.
Bravo to mass-media writers getting it right. And kudos to our friends in the community networking world for continuing to explain what they do and why (and apparently get their pictures in the paper).
Earlier in the week, I wrote about 3Com's announcement of a 4-device Ethernet-to-Wi-Fi bridge that required no special software and would work with any Wi-Fi access point. I queried the company for answers to a few questions about how it actually works, and decided that this is an awfully elegant solution.
The device doesn't use MAC masquerading, as I thought, where one MAC address would associate at the AP and the device would work like NAT in routing traffic. In fact, the bridge associates up to four MAC addresses via its single radio, effectively turning each Ethernet device into its own Wi-Fi device without any intermediate configuration.
This is extremely elegant, as it requires only minimal configuration of the bridge, and no real configuration of the Ethernet devices. Pricing was estimated at $350 and the unit is set to ship in January. I expect that this will supply one of the few missing pieces in enterprise/office migration to Wi-Fi, by bringing printers and older units into the loop.
The Wall Street Journal weighs in on free public Wi-Fi networks: I can't link to the article or reproduce it here, but suffice it to say that the analysis goes skin deep. It focuses pretty much on a single individual in Colorado, and doesn't - to my mind - fully capture what's going on.
The mainstream media is treating Wi-Fi the way the Internet was treated originally. The technical details coupled with scattered widespread and disparate methods of adoption and deployment lead to articles that try to exemplify a trend, but only illuminate a tiny aspect of it.
Here's the nut graf (the quintessence of the article): Mr. Selby is a wireless guerrilla, one of several hobbyists around the nation who are building shoestring wireless networks out of such materials as potato-chip cans and rubber hoses. They are doing so by piggybacking free of charge on the premium high-speed Internet connections that telecom and cable companies provide to many homes and businesses for as much as $1,000 a month. Even so, Mr. Selby, who eventually aims to charge for access to his network, says he hasn't encountered any resistance from providers of such high-speed links, who don't seem worried about his plans.
Selby isn't really the epitome of this topic, given that most of the community networkers never plan to charge - that's part of the point. I don't blame or discourage him from his goals, but the Journal picked an odd duck out of the line-up.
It also fails to discuss the obvious: many contracts with ISPs allow shared use of bandwidth among users. This is a frequent topic of conversation on the BAWUG wireless mailing list, which I recommend as a meet-and-learn list for anyone involved in setting up any kind of Wi-Fi network that has a free or fee public component. If your AUP (acceptible use policy) with your ISP prohibits sharing, then you're on your own. If you buy a T1, like one BAWUGer did, and it says you can resell bandwidth, you're golden.
It's clear the story was written to identify the coming convergence with cell carriers' interest: Other entrepreneurs are launching companies to offer small-scale Internet access via 802.11b in airports, hotels and coffee shops. And some think the technology could be harnessed to offer commercial high-speed Internet access to homes and offices. These developments could conceivably spell trouble for long-delayed "third-generation" cellphone networks, which are to offer high-speed data services in addition to voice. The small scale reference is odd, too: there are companies with a few dozen hot spots (which we'd all agree fits the definition), and then others like Wayport which have several hundred or have wired entire airport terminals. It'd be more accurate to say dozens of companies of all scales are trying to capitalize on the growing use of Wi-Fi in homes and businesses by building out networks.
This mysterious paragraph fails to explicate the security dilemma: Security is an issue, as some companies using 802.11b discovered when hackers tapped their corporate networks. Mr. Dayton says he can detect a neighbor's 802.11b network when he logs on at his Los Angeles home. You can't prevent people from picking up the signal, which is why Mr. Dayton sees his neighbor's network, but you can encrypt the traffic so they can't read it. Most experts think the problem can be circumvented. (Mr. Dayton in that graf, by the way, is Sky Dayton, founder of Earthlink.)
I would have written the graf this way: Wi-Fi offers built-in security through an encryption system, but researchers proved this was easily broken in summer 2001. Several free Internet tools allow non-technical hackers to break into networks that rely just on Wi-Fi's encryption. Corporations typically use a strong, unbroken method lumped together as VPN (virtual private network), while individuals are often stuck with little or no protection. ... and then into the rest of the graf eliminating that last, odd sentence. What problem can be circumvented? The lack of security, Wi-Fi's broken WEP system, or the presence of security (which I guess isn't a problem)?
SkyCross offers multi-band antenna for existing cell and 2.4 GHz: in another vision of things to come, SkyCross has released an antenna for embedded devices of all kinds that can handle the Bluetooth/Wi-Fi band as well as the popular U.S. cell bands. This means that a single device could handle convergence of data and voice across all the most popular frequences. Cell companies will want integrated devices to sell to their customers so that a single card or phone or appliance will work seamlessly.
3Com bridge links up to 4 Ethernet devices to any Wi-Fi network: this is all theoretical until we find out limitations (will it work with all chipsets, all firmware, all Wi-Fi certified devices?), but it's a great idea. The $350 device ships in January, and doesn't have an integral hub. It sounds as if it supports four MAC addresses of any kind, not just PCs running special software for instance. In other words, a real hub. The other side of this equation should be interesting. Wi-Fi clients use a single MAC address when associating with an access point; will the 3Com use a kind of NAT-for-MAC?
Dual-band radio announced by British firm: shipping samples by 2nd quarter of 2002. This is one of the first (or the first?) of many such announcements that will come on dual-band 2.4 GHz (802.11b/g) and 5 GHz (802.11a) radios for PC cards and access points.
Are you a public space and want Wi-Fi? A reader queried me about starting a page that would help match sites like coffee shops or hotels that want Wi-Fi public access for their customers with companies that want to provide said service. If there's interest on either side of that equation, email me, and I'll start putting it together, and link it from the menu at left.
Brits stuck on the fee part of for-fee wireless: although cell carriers are free to charge for use of licensed sprectrum, the UK regulatory authorities won't decide until at earliest Feb. 2002 how to allow commercial entities to offer for-fee Wi-Fi and Bluetooth service. They worry about being behind the Swedes. Hey! Look over the Big Pond, blokes!
1,000 Wireless ISPs: Robert Hoskins, the broadband wireless editor of Broadband Wireless Exchange contends that over 1,000 ISPs in the U.S. offer some form of wireless service for their customers. After reviewing his extensive listing of services with links to the ISPs, I am hard pressed to disagree - and I'm surprised. I know of a dozen or so firms personally, including an ISP in Maine that I wrote about for O'Reilly Networks, but I didn't realize the phenomenon was so widespread. Most of these firms are independent and even mom-and-pop size, which means but one thing: wireless ISPs are profitable.