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.
The two researchers, David Maynor and Jon Ellch, who allege to have documented flaws in Apple's Wi-Fi security no longer speaking at Toorcon: The description is still up, but Securiosis notes that SecureWorks has released a statement that the firm, Maynor's employer, is working with the CERT Coordination Center and Apple. Neither firm will release more statements until CERT and the two firms agree, SecureWorks states. (Ellch ("Johnny Cache") doesn't work for SecureWorks, but has deferred to them because of his partner in research. I have been told that Ellch recently completed graduate school and took a job with another firm, although his Web page reveals little.) Update: Ellch told News.com that he wouldn't deliver the talk without Maynor, and also confirmed with Apple that it is working with SecureWorks. Later update (see below): Ellch gives a rant at Toorcon.
I won't rehash the entire issue here, but Maynor and Ellch have been prevented from actually speaking about this issue in a direct manner since they spoke to Washington Post online Security Fix columnist Brian Krebs. As a result, most information is secondhand, incomplete, incoherent, or off the record. I have piles of information received from many sources that I can not share in any fashion.
Where it stands now is that SecureWorks can effectively prevent the full disclosure of information that would either reclaim Maynor and Ellch's reputation, or demonstrate that Maynor and Ellch provided precisely what Apple has previously at length said that they did not. (See George Ou's publication of a very on-point response from Apple on the specifics of what Maynor and Ellch and SecureWorks provided.)
What's left at stake now with Apple having released patches that apparently include anything that Maynor and Ellch allege that they discovered is the reputation of those two individuals.
Saturday update: News.com reports that Ellch got up on stage and read a statement. (eWeek reproduces the rant here.) For some reason, Ellch and the reporters covering this are fixated on the "Mac community," as if that's a monolithic entity. On SecureWorks, Apple, and CERT working together, Ellch is quoted stating: "That's funny...I thought there was no bug, and I thought SecureWorks provided no useful information to Apple...If SecureWorks provided them with virtually nothing useful, then what...could they have to coordinate with CERT?"
News.com writes, "The two researchers had put pressure [sic] Apple by saying they would do a live presentation at ToorCon, Ellch said." That's a paraphrase, and I don't see this precise text in his reproduced rant, so I don't know if Ellch said words that match the rephrasing. However, "put pressure [on] Apple" could be better stated as, "violated the basic tenets of security reporting in which companies or projects are given advance word of flaws in their products and given reasonable amounts of time to respond and release patches." Just by the way.
Ellch also noted, "A few weeks later, one week before ToorCon, they patch it, and say we had nothing to do with it," News.com reports. Of course, "it" isn't just out of context in this quote, but lacks context in the sense that Maynor and Ellch have apparently not been allowed to release the documentation that would prove Ellch's point. There is no way except by taking Maynor and Ellch on faith to accept Ellch's assertion.
Some folks following this issue believe that such documentation doesn't exist, or that, if release, it would prove Apple's contention. In any of those three cases (pro, con, or non-existent), SecureWorks is the linchpin here. Ellch wrote in his rant, "One day before [our ToorCon] talk, SecureWorks and Apple get together and manage to stop Dave from coming." Apple isn't Maynor's employer; this is called displacement.
The News.com reporter summarizes, "While some in the Mac community see the cancellation of Saturday's talk as prove that Maynor and Ellch are frauds..." Yes, that would be Artie MacStrawman. Damn that hayseed!
Verizon is apparently considering selling its telephone lines in Vermont, New Hampshire, and Maine: The 1.6m local phone lines in those areas are expensive to operate, and as a large firm, Verizon doesn't receive subsidies that smaller carriers do, which allow small carriers in these markets to remain in business. Any buyer of the lines would not receive higher subsidies, according to this New York Times article.
Rural and town residents in these three states already have developing-world quality broadband service. There's a fear with a smaller firm with less capital that even less effort towards improving data services will be made.
One potential buyer of these phone lines is CenturyTel, however. The firm won the contract to build out Wi-Fi service in Vail, Colorado, and just signed up to deploy a 1,700-square-mile network in Pierce County, Washington, covering mostly rural areas and small towns, with Tacoma as the only large city in the mix. There's a slight glimmer of hope for the rural New England market in that outcome if Wi-Fi and other wireless data networking becomes part of CenturyTel's arsenal of tools in making these rural and small-town markets profitable.
The CIO of Philadelphia left the city to join a consulting firm that had previously provided services to the city and the Wireless Philadelphia project: Those consulting jobs were long over when Neff took the job, but ethics concerned were raised about past and future business. Neff had pledged, as is required, to not solicit business from Phila. for some period following her departure; she'll be heading up Civitium's international work, in any case.
KYW in Philadelphia reports that the ethics board found no violations because the consulting firm had no active contracts at the time Neff took the job. However, they suggested she should have consulted the board in advance to receive an all-clear before taking the job. KYW reports Neff's written response as stating that given there was no requirement to consult the board first, there's no reason to complain that she didn't.
Microsoft will sell its Wi-Fi-enabled Zune music player for $250: The device will ship Nov. 14 in the US, and feature a 30 GB hard drive and some preloaded content to try out, as well as an FM tuner. A Zune Pass subscription is $15 per month for access to millions of songs in their library during the period of the subscription. Songs are purchased for 79 "points" each, using the Microsoft Points system, which they say is a portable method of storing value across their products.
Zune has Wi-Fi built-in for transferring "samples" between Zune users. Songs can be sent to another user, who can listen to the song three times within three days and then flag it for purchase. I'm still a bit confused about whether Zune Pass subscribers, when they receive a song from another subscriber, then have to enable that song when they're connected to the Internet again.
Zune owners can apparently download and transfer unprotected (Creative Commons, self-ripped, etc.) content, although Microsoft will apparently not allow that content, when transferred over Wi-Fi among Zune devices, to escape the three-day, three-playbacks limit. A Microsoftee working on Zune says that this doesn't modify the file; it's a device-based limit, and he claims this is technically not digital-rights management (DRM), as DRM normally accrues to a file, not device. In related matter, as the Electronic Frontier Foundation (EFF) noted two weeks ago, the Zune will not play non-Zune protected music, even though it uses Microsoft's own Plays4Sure DRM.
Microsoft will offer a host of accessories, all oddly simply named. They did learn something from their internal "why iPod branding is great and Microsoft overlabels everything" exercise. The Home A/V Pack is $100 for a wireless remote, output cable, sync cable, dock, and ac adapter. The Travel Pack ($100) includes a remote, earphones, bag, sync cable, and AC adapter. The Car Pack ($80) has an FM transmitter and car charger. These items will all be sold separately as well.
A big mistake on Microsoft's part, however, is not defining what comes with a Zune--do I get an AC adapter or any charger cable with the $250 Zune? You have to read footnotes. The Zune comes with a sync cable for charging; the AC adapter is sold outside of the packs noted above for $30. The $250 iPod model comes with a sync cable, earphones, a dock adapter, and a crappy case.
Pierce County, covering nearly 1,700 square miles, plans to roll out a wireless network: The county of 750,000 residents, includes the city of Tacoma (pop. 200,000), the butt of my hometown of Seattle's jokes until the recent high-tech, low-smell boom down south. (Tacoma is known for the odor produced by forest product handling, when the wind is just right.) The pilot will roll out starting in November town by town initially. A consortium of municipalities is working with CenturyTel, which will bear all the cost, and which will use SkyPilot gear. Tacoma may opt in for a downtown pilot project.
While SkyPilot maintains their gear competes effectively against other metro-scale providers for urban deployment, I do believe they have a special advantage in rural/town networks. SkyPilot uses a single radio that constantly switches through a scheduled pattern on eight sectorized (45-degree) antennas. This allows them to run extremely high-powered directional links and low-power local links on a single device with simplified pointing. This would require multiple radios or nodes with other manufacturers' gear.
The network will sport free, time-limited public access each day; premium, fee-based access (pricing not set); commercial business-grade broadband; secured municipal and academic access; and free public safety access over what I believe is both the 2.4 GHz public and 4.9 GHz public safety bands. The network will also be resold on a wholesale basis to ISPs.
The "west-metro" area of Colorado--Lakewood to Boulder--collaborate on 220 square mile network: The '"largest regional Wi-Fi networks" phrase will soon need to be retired. This project is big now, but with 1,000-square-mile projects busting out, it will seem quaint in a year. The area in question comprises 10 communities and 600,000 people, and the goal is to have a vendor building the Wi-Fi network by early 2007. They'd like to charge $20 to $30 per month for basic access with low-income discounts.
Santa Barbara airport adds Wi-Fi: ICOA has installed service that launches tomorrow at the airport in the affluent southern California town, home to a bizarre newspaper conflict involving the ex-wife of a prominent McCaw; Doc Searls, blogger and raconteur, an author of The Cluetrain Manifesto; and Jim Sterne, the nicest e-marketing guru on the planet. Oh, and 200,000 other people in the metropolitan area. The airport handles nearly a million passengers a year.
The unsigned editorial at the Boston Globe says the FCC should favor Continental's side of the dispute: The Globe argues that Massport's contentions about interference and other problems are unfounded because "Massport would be foolish to assign important security information to these frequencies, whose chief purpose is to facilitate public access to the Internet." Right on. The Globe doesn't mention that no other airport in the known world is attempting Massport's appropriation of a federal role in spectrum policy regulation.
The Globe also notes that Boston's Partners HealthCare system, a hospital operator in the area, is supporting Continental's efforts to run its own Wi-Fi network at Boston-Logan. The healthcare operator worries that a shift of control of Wi-Fi's spectrum from the FCC to property owners would reduce a tenant's ability to set its own security policies.
The New York Times offers some strange cover of Fon in Wednesday's paper: I don't quite understand this article, which I think fails to emphasize Fon's actual strengths and weaknesses. Fon is designed to be a Wi-Fi extension to fixed wireline networks. The company and its founder have emphasized time and again that if they are successful in their approach, traditional fixed wired broadband operators will sell a much greater number of pipes to consumers and businesses, not fewer. Fon is not mesh, and its pricing (for all but Foneros who agree to share their own active networks) is designed to discourage daily use--at about €2 or $2 per day, the service is intended to cost more when used every day of a month than having one's own wired broadband connection. (Or, in the US, an unmetered 3G hookup.)
When I was first writing secondhand reporting on Fon, when I was writing about the coverage and what Fon founders and enthusiasts were saying, I mistook some of these statements as Fon trying to challenge established hotspot networks. Some press releases and blog posts did and some continue to compare Fon to Boingo Wireless, noting that Fon had some tens of thousands of Foneros, and had exceeded Boingo's hotspot count. But these comparisons weren't apt, because most of Boingo's locations are centers of business, including places like convention centers, hotels, and 1,000,000-sq. ft. airport terminals; Foneros are at homes and smaller businesses, at least in this initial wave. It's very unlikely in the current model to see Fon locations set up in business traveler venues. (I later interviewed founder Martin Varsavsky, and learned much more about Fon's real efforts.)
Counting hotspots is always a tricky proposition, too. The Times article cites the 85,000 registered users and the "thousands" of enabled routes. Over at tech.am, Mike Puchol has posted an extensive analysis of data he pulled from Fon's mapping system and estimates 3,700 routers are online at any given time worldwide in Fon's network. Fon didn't respond to my request for comment, but Fon's founder responds in the comments that only newer firmware installations have a "phone home" option that shows when a Fon location is active; older firmware would need to be upgraded to appear as active on the map.
More problematic is how the Times article seems to equate public, for-fee hotspots solely with two failed companies, Cometa and Joltage, failing to mention the 12,000 hotspots in the US under management by Wayport, which include several thousand operated by Wayport for AT&T FreedomLink, and about 8,000 in McDonald's that Wayport provides services for the fast-food giant and to which Boingo resells access to AT&T and Nintendo. T-Mobile's 7,000 locations aren't mentioned, nor the thousands of other locations that charge for access or the thousands that charge nothing for access. These locations all compete in some fashion for Fon's mindspace, although the free sites are more "competitive" in that Fon will still charge a price--either dollars or participation.
I have to disagree with an assertion in the article, not credited to any interviewed subject, that 125 routers in the East Village constitute complete coverage. There's no way in which 125 routers of any kind but high-powered, externally mounted Wi-Fi nodes could provide any form of non-spotty coverage over that area. That's distinctly different than providing coverage in the right place.
The article also misses a point that I certainly also misunderstood in my early coverage. Fon is not encouraging its members to violate ISPs terms of service. A Time Warner Road Runner spokesperson trots out and says the usual thing about theft of broadband being like theft of cable--it's entirely regulated in a different manner, and it's a poor comparison--but Fon doesn't want its Foneros to use ISPs that don't allow or encourage sharing. They've been working over the last months to find either ISPs that allow it (like Speakeasy, based in my hometown of Seattle) or that explicitly will partner with Fon to promote it. They're finding more non-US ISPs, which I predicted would be the case, than those here in the good old Homeland.
Sascha Meinrath's criticism of Fon in the Times article is valid from the standpoint that Meinrath is trying to create robust software to bypass the control structures of existing monoliths to create more power in the community, and, in fact, faster overall networks in which the Internet path is not the gating item for bandwidth. Fon is trying to extend the power of wire in opposition, one might say, to mobile networks, which have been extremely expensive outside the US. (That's changing rapidly, as convergence of all kinds have caused home-country and roaming rates to drop everywhere.) Fon's founder, Martin Varsavsky--not mentioned by name in the article--has expressed interest in mesh networks and other means of spreading Fon's coverage, but it's not a focus in these early stages, and it's not the part that they are apparently putting their efforts into.
Finally, there's an odd reference in the Times article to Less Networks, stating that Rich MacKinnon, a normally sane businessman, having 100,000 hotspots in the Less network. Now, that's clearly outside the universe I live in. I have a query into Rich as to what he actually said--the citation in the Times is a paraphrase, not quote--because while Less has good software and is a growing business, this number would indicate that institutionalization is necessary for poor Mr. MacKinnon. Update on Oct. 10: MacKinnon was traveling and replied to my query. They have 150 hotspots, but 100,000 users, and hit 500,000 sessions recently.
[Disclaimer: I contribute regularly to the New York Times, and have written for them in the past about community Wi-Fi, Wayport, and other wireless data related subjects. I have also written for the Economist about aspects of metro-scale Wi-Fi.]
All about ultrawideband (UWB), including its imminent arrival in products: Stephen Wood coordinates Intel's ultrawideband (UWB) strategy and is the president of the WiMedia Alliance, a broad industry trade group that develops standards for UWB. In this podcast, he talks about the technology, its potential, and the timetable for products with UWB appearing on shelves--as early as the end of 2006. [36 min., 17 MB, MP3]
UWB is busting out all over this week, by the way, with demonstrations all over the Intel Developers Forum. Artimi, for instance, is showing off its system-on-a-chip UWB solution for low-power devices. Their equipment will find its way into cameras and handsets. Also this week, an announcement from the USB implementers Forum of the certification program for wireless UWB that will be labeled Certified Wireless USB. And, as Wi-Fi Planet notes, several firms have released UWB references designs--prepackaged board designs meant to be turned into final products--including six designs from Staccato alone.
Philadelphia offers a map of its pilot area: Wireless Philadelphia is starting to gear up. They've got a CEO and a staff, and are starting to push out dollars and other information. In a newsletter distributed locally today, they included a map of neighborhoods covered by the 15-square-mile pilot project that EarthLink will build to prove its system works and to test parameters for the system; and rates. Retail residential pricing is $21.95 per month, with a Digital Inclusion rate of $9.95 per month for those earning at or below 130 percent of the federal poverty level or participants in certain "supportive programs," such as the free lunch program. (The map doesn't appear to be on their Web site, but a description of covered areas is.)
Sacramento receiving bids for network: The city re-issued an RFP after asking its first bid winner, MobilePro, for a higher level of free service than originally requested in the RFP and higher than MobilePro felt was economically sustainable for its model. Sacramento is asking for 300 Kbps of free access, what Google is underwriting in San Francisco, but far below MetroFi's ad-supported free speed. They also want seven percent of advertising revenues donated to digital divide access programs. MetroFi will bid on Sacramento's proposal. A firm that unwired nearby Galt with BelAir gear will expand its WISP offerings into Sacramento and Lodi, but won't answer the RFP. They charge $25 per month for 2 to 4 Mbps, and reject the economics of ad-supported service.
Groups circulate plan for San Francisco to own its network (PDF): GigaOm's Katie Fehrenbacher reports on a plan making the rounds from the Institute for Local Self-Reliance and the Media Alliance which proposes that SF hire out construction but own its metro-scale network rather than continue on the EarthLink/Google path. The document spells out the scenarios, but the groups believe SF could easily recoup a modest investment in a short period of time, with surplus income following thereafter. Their model is similar to that proposed in Boston, in which a non-profit coordinates the network's buildout and infrastructure, and may sell only wholesale access to providers who resell to customers. (Update: SF will study this alternative at a county supervisor's request.)
San Francisco defends progress on talks with EarthLink, Google: In front of the Local Agency Formation Commission (LAFC), a county agency overseeing cable TV franchises and other issues, a policy analyst from the city's executive branch said that talks had progressed, Google's pique notwithstanding. The analyst said that talks have covered duration of term, privacy issues, the building of a test network, and fees. SF would like 5 percent of gross revenue plus utility pole fees, and would use this money for digital divide initiatives. You have to love this part, though: "San Francisco Supervisor Jake McGoldrick said he feared the city was undervaluing the city's assets by seeking just 5 percent of revenue. [Policy analyst Brian] Roberts said that percentage would be very competitive with what other U.S. cities have gotten." Right--here's the problem. This is a network that will be built for the city within the city. It doesn't really matter what other cities have gotten. It matters what San Francisco can negotiate with its preferred bidder, whether that's 1 percent or 20 percent. "Competitive" implies that you're trying to have parity with other cities, which only makes sense when comparing tax rates to attract businesses, I'd argue. "Consistent" makes more sense to me--ensuring that you're not asking something that's totally ridiculous, illegal, inappropriate, or...too little.
In an interview this morning with Panasonic Avionics's David Bruner, he said the firm could hit its target of 500 committed aircraft: The avionics division of Panasonic said last week that they would consider deploying a satellite-based high-bandwidth broadband system in aircraft, which would be similar to a system developed by Connexion by Boeing. Connexion is being shut down in stages, shuttered due to massive revenue shortfalls. Panasonic's Bruner, their head of strategic product marketing, told a trade journal that the firm wanted a critical mass of commitment of 500 aircraft within 60 days to make the deployment viable before launch.
In an interview this morning, Bruner clarified two points made elsewhere. First, the 500 aircraft would be equipped over five years. Second, commitments of 150 aircraft by airlines so far include no Connexion customers. They hope to make a deal good enough for those airlines, such as Lufthansa and SAS, to agree to be part of the new network. "If we take what we have together with the number of aircraft that were committed to CBB [Connexion by Boeing], we're getting very close to the number we want to have to launch," Bruner said. "We set this target for ourselves to ensure that there really is strong airline demand for this service. We believe that there is, but we need people to commit to it."
Panasonic's approach to offering this service to airlines differs from Boeing's in several important ways. Panasonic will be deploying 2006 technology, not 1999 technology. "We've got a significant advantage, probably two generations forward, just in the modems that are used in these types of satellite communications," Bruner said. Also, with a firm commitment of aircraft in hand in advance, Panasonic won't face uncertainty. Boeing had a certain kind of commitment from domestic US airlines prior to 9/11 after which all bets were off, and the company had to retool its model without having the money to invest in refreshing generations of gear.
The new satellite modems and antennas weigh considerably less and have nearly three times the throughput over the same bandwidth, Bruner said, with 12 Mbps down and 3 Mbps up (versus 5/1 for Boeing) expected. Bruner also noted that his firm can add a second antenna without reducing drag that can substantially boost download speeds--he wouldn't say by how much--for on-board media services, like many channels of television. (Their TV service has appeared on Delta's Song, and they have commitments from two international airlines coming on line soon. The TV system can have broadband easily added to it should the broadband plan go forward.)
Bruner emphasized that the "return channel," or satellite uplink, was critical for onboard broadband. The 3 Mbps they expect to offer should provide the necessary pipe for rich email, VoIP, and other services, which could include video surveillance and other security and safety telemetry by the airline.
While Bruner couldn't project fees passengers would pay, the goal is to get their wholesale price as low as possible to encourage the greatest number of customers, which would include airlines and airline partners, such as Aeromobile, which will deploy mobile phone picocells, and has previously signaled its interest in working with whatever backend transport was available on planes. In contrast to Boeing, which viewed Connexion back in 1999 as a way to bring in billions in revenue, Bruner said that "I don't think anybody's trying to get rich off" their potential new offering, and that some airlines are interested in offering broadband as an included amenity, while others look to charging on a break-even basis.
Because Panasonic Avionics already integrates many kind of onboard gear, they will work with Boeing and Airbus to allow their system to be installed while a plane is being built, which can substantially reduce cost.
The new system, by employing the Ku satellite frequency band, will give Panasonic a fixed yearly expense for transponder rentals; they won't be paying for satellite time by the minute or megabyte, which is how Inmarsat charges its customers. Panasonic will contract with a satellite operator to coordinate satellite and ground station operations, removing yet another cost that Boeing incurred in building its own in-house expertise and own worldwide network of ground stations.
The Ku band is limited over the polar extremities, because the satellites point at populated areas, although they do also cover the Pacific route thoroughly for aviation purposes. Bruner said that Panasonic expects about 90 percent coverage of all existing air routes measured by miles or hours in flight.
Inmarsat's system covers most of the globe in its 3rd and 4th generation satellites, a key advantage for airborne communications of all kinds, but at a high cost. Bruner said that Panasonic installs Inmarsat equipment on plane's today, and that Panasonic's custom software works extremely hard to trim any excess bytes over that pricey wire. Inmarsat's 4G system will offer 432 Kbps per module; planes could have one to four modules installed.
But Bruner notes that Panasonic's system could be coupled in a single aircraft with Inmarsat and even AirCell's upcoming domestic US broadband airborne service so that an airline could choose the cheapest or best method of broadband. Bruner said the company is very interested in working with AirCell as one of their integrators. A combination of Ku band and AirCell gear would make sense for certain wide-body planes that fly over the US and internationally, Bruner said, while narrow-body and regional jets would be poor candidates for satellite access.
Esme Vos details the just-released RFP for the city of Chicago's Wi-Fi network at Muniwireless.com: The highlights are that the city requires wholesale resale of the network; the city won't provide dollar one for the network; it must be ubiquitous and egalitarian in coverage; there is no access to the city's fiber infrastructure; access to city facilities for mounting is non-exclusive; any form of free access to residents or visitors isn't required for a winning bidder.
Muniwireless.com also has the PDF for downloading, as Vos points out that the city requires registration as if you were to be a bidder on the network just to read the RFP.
South Florida will try out rail-based Internet access: Local firm Wi-Fi America, part of the ATCI consortium building an on-board trial for California's Capitol Corridor project, will pick up the $150,000 cost. It's a 60-day trial with free access during the test.
Southern Calif. train line tests onboard service: This line runs around San Diego; the test will cover about 10 miles of the 42-mile line. The cost will be paid by Homeland Security, and used for video surveillance transmission from the train and other fixed points on the line. They're also looking into general passenger access.
The Washington State Ferry system has been operating a trial of onboard Wi-Fi for over two years: The system linking ferries and waiting areas with Internet access was supposed to operate for several months under a US Department of Transportation (DOT) grant that allowed Mobilisa, a firm located up in the tiny, marvelous town of Port Townsend, to test and deploy a network. Then, this firm would help write the RFP; they'd be barred from bidding, however. I wrote about this as part of an article on Wi-Fi on buses, trains, and boats for The New York Times back in July 2004. (That article led into the rail-based Internet access article I wrote for this week's Economist, too. Fruition of these services took two years, apparently.)
Parsons has been awarded the contract, according to a mailing from the WSF today, and it's confirmed on Mobilisa's ferry site. The service has been operated at no cost to date, but Parsons will convert it to a fee-based service. The ferry system had no interest in underwriting the service, although they said two years ago they thought the marketing value for a major telecom firm might have been worth offering no-cost access. During the transition, service will remain free, but there will be interruptions as Parsons installs its own gear. Parsons also won the VIA Rail contract in Canada several months ago, and they are actively involved in a number of transportation-related Wi-Fi operations and bids, given their international focus on the transport industry. The email from Parsons promises to increase the service's backhaul bandwidth, and to offer voice over IP in the future.
The WSF handles about half of all passenger trips on ferries in the US--a total of about 25 million passenger trips each year--and a considerable but smaller fraction of all car ferry trips. Half of the WSF passengers are commuters. (Staten Island Ferries has 20m yearly riders, and while they added Wi-Fi to terminals some time ago, I have heard nothing about on-board access on the short hop it covers.)
Typical commuter runs from Bainbridge Island and the Olympic Peninsula's Kingston port take about 30 minutes, but wait times can be one to two hours or longer during rush hour. Because the service was designed to cover the extensive car waiting areas, someone could remain in their car and get work done during those long periods while they're just parked.
Hey, Maynor and Ellch were right! Sort of: Apple released three major vulnerability patches for its AirPort networking system today, but noted that no known exploit is available. The security bulletin describing the weaknesses indicates that an Apple adapter or a third-party adapter on an Intel-based Mac using Apple's Wi-Fi framework need only be turned on, not connected to a network. And the attack only need be "in proximity," but there's no mention of a requirement to be associated with the network.
The patches fix separate weaknesses that could allow properly crafted frames to cause an escalation in privileges, execution of arbitrary code, or system crashes. The PowerPC patch (for Mac OS X 10.3.9 and 10.4.7) mentions just arbitrary code execution; the two for Intel-based Macs (10.4.7) correspond to built-in AirPort support and third-party hooks for Wi-Fi support, and mention all three potential outcomes. (The PowerPC patch likely only affects 2003-and-later AirPort Extreme technology rather than Apple's 1999-2003 original 802.11b AirPort Card-based adapters and base stations.)
David Maynor and Jon Ellch confused the world by telling Brian Krebs of Security Fix at The Washington Post that they had discovered a weakness and developed an exploit against Mac OS X. They later amended this statement, although Krebs continues to claim to have seem a demonstration of it and has a transcript of Maynor stating that. A few weeks ago, Apple released a statement that they had been provided with no evidence showing the weaknesses, which would allow non-associated attackers to hijack computers without accessing a network, and without a Wi-Fi adapter being actively connected with a network.
I wrote about the confusion about who said what, later statements by Maynor and Ellch, additional detail from Brian Krebs and ZDNet's George Ou, and so forth in a Rashomon post.
Apple's Anuj Nayar told Macworld that Maynor's firm SecureWorks "did not supply us with any information to allow us to identify a specific problem, so we initiated an internal audit," which is how these vulnerabilities were uncovered, he said. He also told Brian Krebs, "SecureWorks approached Apple with a potential flaw that they felt would affect the wireless drivers on Macs, but they didn't supply us with any information to allow us to identify a specific problem." I spoke briefly with Nayar this afternoon, who confirmed the accurate representation of his statements.
If you're a Kremlinologist of Apple and Wi-Fi, as I am, you can read two important facts out of Apple's security bulletin. First, they state for each of the three patches they released that there is no known exploit. That translates into plainer speech as, "We have seen no software code that can take advantage of this." Second, the bulletin omits any thanks to the reporting sources. Apple always publicly acknowledges--as most firms do--the people and/or organizations that provide direct information about security flaws.
Krebs describes this as Apple and SecureWorks differing over "which side found the flaw and how exploitable it really is." SecureWorks hasn't commented on this directly at all, and Krebs said as I post this that they hadn't responded to his request for comment. An IDG News Service story that appeared later note, "SecureWorks declined to comment further on the Apple matter." That sounds awfully final.
Perhaps that's because SecureWorks just merged with another firm yesterday? Could Maynor and Ellch's public complaints about lawyers refer, in fact, to the lawyers at SecureWorks who were trying to close the merger deal and didn't want outside distractions that might affect due diligence? The implication was that Apple was behind legal threats, which seems unlikely given the release of today's patches.
The next step here, if Maynor and Ellch are still maintaining that they had discovered a vulnerability as related by Brian Krebs's reporting on it, is for the two researchers or SecureWorks to release everything they have on this to show that Apple is being disingenuous. Because SecureWorks is now off the hook, right? I don't think there's a chance that we'll see that happen.
Were Apple to be lying about any of this isn't credible; that's a huge risk for a multi-billion-dollar public company to take, and one that, if it were a lie, would clearly result in lawsuits due to the security implications. I believe this might be the last we hear about this.
Update: I'm wrong about this being the last we hear about it. Maynor and Ellch are now scheduled to talk on Wireless Drivers at Toorcon at the end of September in San Diego. Note that while Maynor and Ellch have consistently said they're bad at dealing PR, that the description of the talk uses a particular rhetorical technique in describing the kerfuffle. "Since the first details of our demo were reported two camps instantly formed, people who thought the work and research was good and people thought we faked everything and we are horrible people."
In fact, there are several camps. You know the old joke--there are two kinds of people in the world: those who classify people into two groups and those who don't? There are those who believe that Maynor and Ellch are perfectly fine people and overstated the impact of their research, too.
There's also the problem that there is a very small camp of people who have seen "the work and research." Because none of this has been released, only a few non-disclosed people have ever seen what Maynor and Ellch allege is the vector of attack and the related (if any) exploit code.
Maynor and Ellch have continually tried to recast the aftermath of Black Hat as those of us reporting on it being a bunch of tech newbies who can't see the overall importance of their generic fuzzing approach which can reveal weaknesses that otherwise prove resistant to other forms of testing. Of course, that's never been what's reported on. The issue isn't whether a generic technique results in new methods for improving security; that's fantastic. Rather, whether the two researchers discovered anything in particular.
My prediction is that Maynor and Ellch continue to be evasive at the talk and fail to show any code samples or anything that provides convincing proof that they had an actual instead of theoretical weakness in hand at the time of the Black Hat talk.
Update to update: Okay, with a little more insight that I can't provide details about, I now believe Maynor and Ellch will provide a lot of detail at Toorcon. We'll have to wait and see.
And another update: George Ou runs through his timeline on the exploit/weakness debate. Ou has a variety of information that he is not allowed to disclose--he discloses that he can't disclose it--that lead him to state definitely that Apple is not giving Maynor and Ellch due credit. He'll be at Toorcon and offer coverage of that event.
Trains get unwired (subscr. req.): Several hundred trains will be equipped with Wi-Fi for access and WiMax, satellite, and cell transceivers for backhaul by the end of 2007, I report in The Economist magazine's Sept. 21 issue, now online. In talks with rail operators and Internet access providers in the US and Europe, it's become clear that the phenomenon is bigger than I realized. When I set out to write the story this summer, I thought there were dozens of trains equipped with Internet access worldwide. In fact, my count is at least 100 (as of a few weeks ago). That's trains as in an engine plus cars--not train lines (there are over a dozen lines covered), nor individual train carriages. The Dutch railroad alone expects to equip 1,400 carriages representing hundreds of assemble trains.
An item there was no room for in this article is that electrical outlets are being added to existing trains and have become a requirement on new passenger cars. With outlets, the train becomes a far superior place to work than any alternative for commuters. Many commuter trains offer wide seats, trays, and business-class coaches or amenities, too. Let's see--drive by myself, get no work done, burn gas by the ton that costs a fortune, or, an inexpensive and comfortable trip with Internet access the whole way? I foresee big changes in commuter rail.
Today also brings an expected announcement, which is that Virgin will work with Nomad Digital, the folks behind the Brighton-to-London line, to equip all of its English West Coast trains with Internet service. Nomad uses fixed WiMax at regular intervals to provide service, and notes in this article that they will have the third-generation cellular data standard HSDPA as a backup to cover gaps in service. Top speed will be about 2 Mbps on this line; they can achieve 6 Mbps intra-network on the Brighton line, their head told me.
A set of public agencies in the Sacramento Valley considers plans for a 12,000-sq. mi. network: WiSac would cover 30 communities, and be used for public safety and monitoring levees, among other purposes. Sacramento would be included. The plans aren't far along, and there's nothing in place yet for financing. This would dwarf any other project worldwide except Singapore's ultimate goal of unwiring the entire country.
Reuters reports the FCC will rule in favor of Continental Airlines in a dispute over the right to use unlicensed spectrum at Boston-Logan International Airport: The dispute is now well over a year old, and stems from Continental choosing to offer free Wi-Fi in its membership lounge rather than sign on to Boston-Logan's airport-wide system, which is a for-fee network. Massport, which operates the airport and other transportation infrastructure, said that interference would occur and security would be compromised, and that Continental was violating its lease.
As I have written several times in covering this ongoing matter, the FCC has maintained--and one could reasonably maintain--that spectrum regulation is rightly the domain of the feds and that the FCC is the only body that can control spectrum's use. They issued an order to this effect in June 2004 that could hardly be clearer. A landlord usurping that role acts against the federal interest, especially in unlicensed bands, which have requirements for certification and operating parameters, but no restrictions on the party operating services in that band as long as they are within legal guidelines.
Reuters says an order is written and approved by the FCC chair Kevin Martin, and is awaiting likely approval from at least two of the four other commissioners.
Massport's contentions have rested at times on two different grounds. First, Continental was causing interference. That's problematic for a band in which interference is an expected part of operation, and not a cause for shutdown of offending transmitters unless they act outside the legal signal limits. Second, introduced as a later argument, was that vital security purposes would be subverted by having a non-Massport-controlled network in operation. Quite specious given that one would hope that a band that could be disrupted by a balky cordless 2.4 GHz phone wouldn't be the basis of our airport safety. And that hundreds of other airports have no such restrictions or concerns.
The Oregon Public Utility Commission solves power problem preventing city-wide network plan: The PUC will allow new rules so that Portland General Electric and MetroFi can proceed in their agreement for MetroFi to mount Wi-Fi nodes on utility poles. Once again, it's about power! Although, in this case, the utility doesn't sound reticent, just blocked by particular rules.
Connexion is shutting down, but Panasonic's avionics division may step in: This is not a sale of assets, but rather Panasonic Avionics Corporation may use Connexion by Boeing's general approach in a way that would allow airlines that have already deployed Connexion to make a simple and relatively quick transition. In an interview with Shephard Inflight Online, David Bruner, the strategic head for the 2,500-strong Washington state-based division of Panasonic North America, lays out the plans. Airlines will receive varying amounts of cancellation fees from Boeing as a result of Connexion's service being pulled, which may be used by airlines to fund the transition.
Panasonic wants commitments within 60 days from airlines for 500 planes to be equipped; they say they have 150 committed now. Lufthansa has over 60 planes with Connexion and SAS with about 20; I suspect both airlines have signed on. They'll still use Ku band leased transponders, but with much more efficient equipment that will allow 12 Mbps/3 Mbps versus Boeing's 5 Mbps/1 Mbps rate.
The antenna will be more "compact"--read: not weighing 800 pounds--using a new system from L-3 Datron Advanced Technologies, as well as a modem from their Linkabit division. Weight was a considerable portion of airlines' concern over Connexion. It costs thousands of dollars per long-haul flight just to move Connexion's equipment around, regardless of passenger count or passenger usage. The system seems to be their Exconnect platform. It's not clear whether already-deployed Connexion airlines would have to pull their gear or whether they could use the new system with existing equipment, and then upgrade during scheduled maintenance periods in the future.
Bruner didn't mention a specific price in this interview, but the target is a wholesale rate comparable to terrestrial access, which means about $10 to $15 a day, like you pay in some hotels. This lets airlines pass it on with little markup, Bruner says. Airlines might also eat $10 per user per flight for business and first class--that's a little easier to swallow than $27, or whatever Connexion wholesaled service to airlines at (if they wholesaled in this fashion).
Panasonic said that they set the 60-day deadline in order to not leave existing Connexion airlines hanging. Panasonic is the equipment integrator and operator for Qantas's in-flight mobile calling test scheduled for next year on a single aircraft over Australia.
Interestingly, Aeromobile, poised to launch Inmarsat-based service in Asia possibly late this year, could also buy service from Panasonic. Panasonic could avoid being a competitor because they're seen as an avionics integrator, and could make their money from providing the pipe, not the service. [link via iag]
Peplink offers load balancing, redundancy with combo package: The firm is coupling its Surf 200BG wireless bridge, designed to connect to metro-scale Wi-Fi networks, and the Balance 30, a load-balancing network bridge that accepts broadband from both a wired connection and the Surf 200BG. If one connection goes down, the other is available. Requests can traverse either network, preferring the less-congested one, I'd imagine.
At $600, it's a bit pricey for a home user, but I can see where SMBs (small-to-medium-sized businesses) would find it a good investment. I've wondered about how many subscribers to a city-wide network would use that Wi-Fi service as both a backup (for inevitable wireline downtime, however brief) and as a mobile alternative. I suspect that some businesses will arrange special rates with retail ISPs to equip a number of roaming employees and have a link for their office. That Wi-Fi link could even be the primary one, with some metro-scale operators offering business-grade service at 3 Mbps or higher, and a lower-speed ADSL line acting as a backup.
An important factor for a multihomed bridge like this is that even though it can virtually bond two separate networks, it can't actually split traffic at the packet level. Connection-based services, like a VPN, would be routed over one of the two broadband networks and be disrupted were that network to drop. Web requests could be shuffled--one image request over one network, another over another--but it requires the cooperation of an ISP at the routing level to allow true bonded, multihomed networks.
Dear readers, wondering what the interest might be in a regular, hour-long podcast that features call-in questions of myself and a guest? With 23 podcasts under my belt, I've been considering other ways in which delivering audio could answer the needs of you, dear readers (and listeners). One idea I've considered is a call-in show, in which various tools would be used to allow questions to be left (via voicemail) before the show and live questions "on the air." The show would feature myself and likely a guest with specific knowledge of some area of wireless data that we'd ask questions to focus on.
I've also been thinking of putting together roundtables, in which several people could participate in a discussion of a given topic, such as municipal funding and ownership of Wi-Fi networks, or the future of mobile WiMax as a viable U.S. alternative to cellular data networks.
Feedback? Post comments below or feel free to email me directly.
The folks at Fon add prefab La Fonera to their line-up: The company is selling a concept, not hardware, but hardware helps enable the concept. Fon wants millions of people to become Foneros, sticking a Fon-enabled Wi-Fi gateway on their separately purchased DSL or other Internet pipe, and then allow other Foneros and non-Foneros access. This process is accelerated by partnering with ISPs, who think they will sell more connections with Fon-enabled routers hanging off the end, and by making it simple to get a Fon-enabled router.
While you can flash relatively inexpensive hardware with Fon's modification and extension of DD-WRT, La Fonera is ready to go. At $5 or €5 (one per registered Fonero on their first purchase), it's an awfully good deal. The device offers two virtual SSIDs so that you can have both a private and public network with separate parameters on each.
Foneros who choose to share their connection at no cost with other active Foneros can, in turn, use other Foneros' connections for free. All others pay a small daily fee which is designed to be useful as a day rate, but not be cheap enough to be an alternative to paying a separate monthly fee. [link via GigaOm]
T-Mobile garners 120 licenses for $4.2b, many in key markets, in just-closed auction: Cingular and Verizon Wireless also spent heaps to buy more coverage. T-Mobile is a big winner, though, given its lack of spectrum portfolio. These wins gives them more space in New York, Chicago, and San Francisco. A consortium of four cable operators spent $42.b for 137 licenses covering a chunk of the Northeast, Washington, D.C., Detroit, Atlanta, and a number of other cities. Part of the joy in owning licenses is trading them with other firms and using them as chips to leverage roaming deals and other partnerships. Cingular and T-Mobile initially built out the GSM infrastructure in the US; neither could afford to do it on its own. I'll be curious whether Cingular and T-Mobile now collaborate on HSDPA buildouts, with Cingular falling behind Verizon and Sprint's footprint.
Clearwire competing for hearts, minds of less-served audience: With licenses that allow their broadband wireless service to operate only in mostly less-dense and less-populated areas of the US, Clearwire is clearly trying to address that market. This article from the Anchorage Daily News details a reporting trying out a 90-day program on one bus, in which Wi-Fi service is relayed to the Internet over Clearwire's network. On one route, the reporter experiences some carsickness, enjoys people watching, and sees no other laptops in use. (My only bout of carsickness also arose from trying to use a laptop in the backseat of a car.) On a straighter route, he finds less nausea, but also loses the signal.
The logistics plus Wi-Fi and barcode giant Symbol will auction itself off: The firm could sell for $3.2b, a 20% premium over its current market capitalization. Symbol was a very early Wi-Fi player, using the technology to disentangle cables from its barcode readers and other devices used to track items in warehouses, on trucks, and on shelves. It was also an early entrant into the WLAN switch market with very weird devices that eventually matured and took a sizeable percentage of that market segment.
However, the Wall Street Journal reports that the company has had problems in capitalizing on the markets it's part of or dominates, despite good but not well-growing revenue and marginal net profits. The company also went through an accounting scandal that led to eight executives pleading guilty and the CEO fleeing the country.
Update: Deal announced Sept. 19 with Motorola purchasing Symbol for $3.9b. Motorola has a variety of broadband wireless technologies and is involved with logistics, but Symbol's product line should fill a number of gaps for enterprise and warehouse operations.
Google exec frustrated at San Francisco contract negotiation tactics, delays: The San Francisco Chronicle reports that Chris Sacca, Google's lead in its partnership with Google to put Wi-Fi in that city, is frustrated with the delays that the city has introduced into producing a final contract. The article couldn't get comments from EarthLink or the city, and speculates that Sacca's statements are a negotiation tactic. The article also notes that contact progress is posted.
Sacca contrasts the Mountain View process (2 1/2 months) with the several months so far without the end in sight. He should have studied Philadelphia, where EarthLink was awarded the bid in Oct. 2005, had a contract finished in early 2006, and didn't receive city council approval and then signed contracts until about May. The summer involved permitting and other work, and the pilot network is only now under construction.
Also, Mountain View was getting a big, fat, free gift. EarthLink will charge for access at a higher rate in SF, while Google will offer lower speeds for free.
A group of private-equity firms will buy the giant semiconductor maker for $17.6b: A few days ago, when this rumor emerged, I said that Freescale's unique flavor of ultrawideband (UWB) will most likely face the scrapheap of history in a purchase with flinty-eyed green-shade-wearing types looking at a multi-year effort to produce technology that now has the backing of no one else in the industry of scale (except Motorola, Freescale's former owner). Read last week's post for more detail.
The numbers supporting free, municipal-backed wireless: St. Cloud, Florida, was the first city of any scale in the US to offer a completely free network -- no need to view ads, no extra fees. The city of 11,000 households has seen 77 percent of those homes register for service within the first six months. In this podcast, the city's consultant Jonathan Baltuch discusses what those numbers mean, and why cities should consider paying directly to build a free network rather than allow an outside ISP to build one and charge for it. [37 min., 17 MB, MP3]
Microsoft will allow songs to be sent via Wi-Fi: People can sample each other's purchased song libraries on the upcoming Zune music player by permission. Songs are allowed to be transferred--another story mentioned that not all songs would work this way--can be listened to three times over three days before being purchased for what's expected to be 99 cents. There will also be a subscription service, the price of which hasn't been announced. It will also let you share photos among Zune users. You can turn off this feature or not accept shared music and photos, too.
One might suspect that the subscription service would allow any song downloaded by subscription to be transferred to any other Zune user who also had a valid and active subscription, no? This is a feature of MusicGremlin, which can transfer songs downloaded via its subscription service to other MusicGremlin subscribers. (While the MG can play other DRM'd content from other services, those songs can't be transferred among users.)
In fact, the answer is no. The Electronic Frontier Foundation read the fine print, and Zune apparently cannot play other Microsoft Play4Sure protected content. Holy mackerel, if that's true, that breaks Microsoft's model for media protection--that all Plays For Sure devices play stuff for sure. Further, BoingBoing analyzes this Microsoftee blog post to realize that Creative Commons-licensed content could be wrapped in protection in order to be played on the devices, which would be a violation of their copyright terms. Medialoper has more detail. (The EFF also notes that the Zune's creator suggests that users could rip video using methods illegal under the Digital Millennium Copyright Act, an act that the EFF finds offensive, as do I.) Update: The Zune won't wrap DRM around unprotected content, it turns out, but it will prevent this media from being transferred over Wi-Fi among Zune players without stapling on the three-day, three-playbacks restriction.
Pricing on the Zune is expected to be $250 to $300 with a 30 GB drive, full-screen 3-inch display, FM tuner, and Wi-Fi. It won't work with Macs.
Pinocchio gets its 3G spectrum, probably: T-Mobile looks likely to win 119 licenses for advanced wireless service (AWS) spectrum, paying about $4.2b (at the moment) for a portfolio that includes New York City and Chicago. T-Mobile has two problems in the US: It's the No. 4 carrier with a significant gap below No. 3, and it lacks 3G spectrum that would allow it to compete for next-generation streaming audio and video as well as data services. T-Mobile uses GSM, so they'd deploy HSDPA for 3G data, and that might allow them to do a deal with Cingular to more rapidly share and expand coverage. Sprint and Verizon have substantially more exclusive EVDO coverage in the US; they don't offer EVDO roaming.
A cable group aligned with Sprint has high bids on 137 licenses; there's some speculation that this could tie in with Sprint Nextel's mobile WiMax plans, in that Sprint says their 2.5 GHz licenses for mobile WiMax only pass 100m people in the US, or 1/3rd the country's population.
Update: A commenter says that Sprint's 2.5 GHz licenses cover 85% of the US population, and that 100m people is just their initial launch plan. Does anyone have a reference for that?
AirCell said today that it expected to launch its service in 2008: A few weeks ago, they were saying late 2007. The firm said that it is finally approaching the receipt of its formal FCC license, sometime in the next 30 days. They repeated their earlier expectation that they will be able to use regulatory harmony and partnerships in Canada, Mexico, and the Caribbean to launch throughout those countries and region shortly after their US launch.
Keith Higgins explains IMS: You don't know what IMS is yet, if you're like most people, but you will soon, with a commitment from Verizon Wireless and Cingular Wireless to migrate to the fixed-mobile convergence technology, and interest worldwide as well. IP Multimedia Subsystem is a way to develop one platform for delivering services like voice, video, games, and music for mobile devices and fixed devices over many kinds of pipes--wireless, fiber, what have you. Stoke is developing a universal translator to help operators add IMS services, and Keith Higgins talks about what IMS is, what it can deliver, and how we'll move to an era of getting the same kind of thing everywhere, mediated by the device rather than the connection.
Another aspect of IMS is that the it uses SIP (Session Initiation Protocol), a now-old and well-used standard, for connecting calls and sessions. This is interesting because it shows the adoption of a widespread relatively open standard widely used in VoIP by the major telcos. Of course, they can wall off SIP as much as they want, but it's a nice big chink in the walled garden's fence.
The FCC has a process for allowing unlicensed use of unused TV frequencies by 2009: The process is widely backed by the consumer industry, including Intel, because the airwaves in question--white space in TV channels--are below 900 MHz and thus have much greater penetration and can glide themselves around certain obstructions. The FCC will have test results by July 2007, Reuters reports, that will allow it to set the technical requirements for devices by October 2007. Permits would be accepted in December 2007 for devices that could appear by February 2009 in the market.
The folks at Travelpost assembled availability, pricing, and other details for 135 US airports' Wi-Fi service: They've split it into top 20 airports and then all the rest, including pay-as-you-go rates for hourly and daily use (where available), the monthly rates for different network providers, and details about where service is available. Scroll through the list for two interesting facts. First, the number of free networks, noted in green. Second, the fact that only eight airports out of the top 135 have no Wi-Fi available. That's changed a lot in just a year.
The Lehigh University project turns commercial: The Wi-Fire is a high-gain, directional, external, USB powered-and-attached Wi-Fi adapter with Windows XP support. The idea is that this is somewhere between a high-gain PC Card or ExpressCard, which is anchored into a laptop and only works with laptops that offer a card slot, and a high-gain bridge that requires external AC power and an Ethernet port. They say they've seen 30 Mbps over 802.11g at distances where a normal adapter hit just 1 Mbps. They also claim about 1,000 feet of range, which, for this gain, is pretty reasonable via line of sight.
For $150, about the price of a bridge, you get a fairly high-gain unit. According to their numbers (15 dBm transmit power and 10.4 dBi antenna gain), it's roughly 25 dBm of effective power output (EIRP) or 350 milliwatts. The flexibility of this unit might make it a perfect match with metro-scale wireless networks. Bridges often run at 200 mW to 400 mW of transmit power with omnidirectional antennas that have relatively low gain. Wi-Fire also has high receive sensitivity at -98 dBm. The prototype could switch between omni and directional, but this approach appears entirely directional (and simpler).
Because a bridge is portable but requires line power, this is much more flexible. Combine a Wi-Fire with built-in Wi-Fi that you enable sharing on, and you have a bridge plus AP combination using two radios for best channel separation and throughput.
The folks behind this have done a great job ratcheting up from student project into a shipping product. Lehigh University has a small feature on how three students at the engineering and applied science college received a $2,000 grant in 2004 to create a prototype, assembled a team that built a project that won a $15,000 state grant to create and test 50 prototypes, which in 2006 turned into a $150,000 loan from a university incubator fund.
The company has 1,000 Wi-Fire units ready to sell, aiming at college campuses and bookstores. They plan to go into national marketing and additional products in 2007.
Let's give the Florida town some kudos because it was taking its lumps a few months ago: Jonathan Baltuch of MRI, a consultant for the city on this Wi-Fi rollout, writes in a guest column at Muniwireless.com that in the first six months of operation (as of Sept. 6, 2006), they saw 8,421 registered users or 77% of city households. The city has grown from an estimated 10,000 to 11,000 households during that period with annexations and developments, which tells you how fast towns grow these days. Update: I confirmed with Baltuch via email that the number of households is correct--the count of 8,421 represents at least one user in a household, not 8,421 individuals. They expect this could mean about double that number of individuals using the network, or at least in a home in which the network is being used, based on about two users per household in their survey.
They've logged 1.8m hours and seen over 500K sessions, transferring 10 terabytes of data, with an average session time of 3.55 hours. He's putting the money saved by residents at $3.7m per year based on an average fee previously paid of about $36 per month. (If you were paying $26.95 for AOL for dial-up, that's now dropped to $9.99, but I figure there's a lot of variation among dial-up and broadband monthly fees.)
These are pretty cool numbers because they support both Baltuch's case--the idea that free Wi-Fi paid for by a city could produce economic effects and high uptake--and the opposite. The opposite case being that there's such a pent-up demand for this that a combination of advertising-supported free service and paid monthly service at a rate below an average of $36 per month could have good uptake, too. Baltuch is absolutely correct that with predictions by EarthLink et al. of a 15-20% take rate of subscriptions in cities that will have Wi-Fi networks installed, seeing a six-month 77% take rate is fairly remarkable.
It may cause some rethink. The former mayor, Glenn Sangiovanni, talked months ago about how pretty much every dollar spend on Internet access in his city was leaving the city economy, and that by spending well under $3m, the city would be serving several different missions, including retaining dollars that would ostensibly be kept in town.
Chumby received a rush of blog-licity when the firm handed out these portable Wi-Fi thingamabobs at O'Reilly's Foo Camp to alpha-geeks: The device, in prototype, is small, designed for the "kids," and sports a Wi-Fi adapter, an AC power plug, a small, color touchscreen, and an open architecture. The company wants people to hack the software, hardware, and even the device's case with their own modifications. It's not precisely open source, but it's all open. They hope the device will ship in the second quarter of 2007 for about $150. They also expect that it could be licensed or replicated in many forms--they have released or shortly will release the parts list and schematics among other parameters--and they're curious what results.
In this podcast interview with Avalon Ventures partner and Chumby Industries chairman Steve Tomlin, we talk about how having a device that's designed to be open affects what gets developed for it. We also talk about how Chumby, as a general-purpose appliance, make available many kinds of applications--it's not just another picture frame, just another music player, or just another RSS display. In its current iteration, the Chumby has a touchscreen but no keyboard interface. Tomlin expects someone is already working on that. [27 min., 12 MB, MP3]
The chipmaker, spun off from Motorola, may be purchased: The firm makes hundreds of different kinds of chips, including a wide array of silicon designed for embedded automotive and cellphone systems, but our interest here at Wi-Fi Networking News relates to ultrawideband. Motorola acquired the original UWB developer, XtremeSpectrum, folded it into the chipmaking division; it was part of the spinoff. This UWB approach is now in the minority--a minority of one, really--where the rest of the industry, led by Intel, is part of the WiMedia Alliance, which is rapidly pursuing completion of another UWB technology.
Will UWB survive this buyout? Several years with no shipping products and a delay of shipments of manufactured goods from January to March to July to next year, reportedly, doesn't make it likely that hard-noses investors, on purchasing Freescale, will continue to authorize the division. Freescale is extremely successful in its other fields of endeavor, and I expect that following a sale, we will hear about a write-down for its UWB efforts, and then, eventually, how the firm has joined the WiMedia Alliance.
While multiple years of research is often supported for promising technology, as Freescale as continued to move away from its partners, drop its commitment from a trade group it formed with Motorola, and has not produced results, it's just hard to see a conclusion in which they stay with classic UWB, as they defined it. They may still own patents that produce a revenue stream, of course.
If that's not the case, I promise to buy the first consumer device produced with Freescale UWB chips.
Latest numbers of US municipal projects at MuniWireless.com: Esme Vos has released her latest report (free) on the number of cities in the US that have, plan to have, or are thinking about planning to have wireless networks. She counts 68 networks in operation as of Sept. 10 for combined public/municipal use, 43 public networks, 35 municipal-use only networks, 135 planned deployments (RFP/RFI issued or network underway), and 25 areas where networks are being considered. She counts 281 serious networks up from 247 in June 2006. (She gets 306 networks if you count those 25 that are pre-RFP/RFI.) Vos has the most definitive public information on this topic, and, as always, solicits feedback for corrections and additions.
Milwaukee reconsiders preferred contractor: Midwest Fiber Network's plans to fiberize and unwire the Wisconsin town are now up in the air--pun intended--as the city council pulls its support for a non-bidder deal. The council approved a deal in January, but the contract wasn't negotiated and submitted until June. The council is now in negotiation with EarthLink as a potential bidder. This might be a negotiating strategy on the city's part, of course.
Mercury News's Langberg handicaps the Wireless Silicon Valley plan: He has logic somewhat similar to mine, but with greater granularity and the added benefit of being local to the plan's territory. IBM told him the plan could cost over $75m, which is in line with my $50m-plus estimate. He notes, as I did, that IBM and Cisco have deep pockets meaning this project will be seen through to completion, and that the free service is quite respectable, making high uptake (paid via advertising?) likely. However, he points out that cities have to sign off on contracts individually, and we all know that cities have their own ideas about what they need. The assumption on the network's cost is that cities will be anchor tenants, shifting their data and telecom spending to the new network. Where cities can find this cost effective, they'll do it. Philadelphia has said in the past that moving wired, lease T-1 lines to broadband wireless (not Wi-Fi) links operated by EarthLink will save potentially millions of dollars.
An op-ed article in the Toronto Star notes that profit from the new city-wide network launched last week goes into the city's coffers: Graham Longford and Andrew Clement argue that charging for the network after an initial several months of free usage as the network is built out makes little sense given where the funds go. If Toronto Hydro Telecom were privately held, then perhaps putting profit in shareholders' hands would make sense. The op-ed's authors maintain that conserving the fees that individual users would pay would have a beneficial impact on the city beyond the revenue that this indirectly causes. They also note that a smart meter program was mandated, requiring a network to built at the utility's expense anyway.
They miss one common American argument here, perhaps not as relevant in Canada, is that by charging users, only users pay for the system. By not charging users, all Toronto residents will pay, although much smaller amounts. Thus, Toronto Hydro's method is a fairer form of taxation--which the op-ed authors call "covert taxation"--unless the benefits of this network could be projected to be larger than the costs to those residents not participating in its use.
This is the public good model, of course. We all pay for roads, not just the ones we used, because we all get a benefit from goods and services traveling over those roads. Cyclists like myself would like to argue that we need more bike paths and tools for multimodal transportation that include bikes, because the social good is fewer cars on the road and greater health. Less cars means less smog, less congestion, and less gas used (less demand). Greater health among bikers means less social demand on medical resources, funded or unfunded.
We'll see how this plays out. I don't know what control citizens of Toronto have over their government or the government over its 100-percent ownership of the utility.
The airport authority will offer no-cost Wi-Fi in the food court (news not yet online): The rest of the airport remains a for-fee operation at $7.95 per day, operated by AT&T (formerly under AT&T Wireless, before the Cingular acquisition). The head of aviation for the city gave away the goods when he noted that connections are now 10,000 per month, double that of a year ago. A maximum of $80,000 per month in revenue for a system that surely cost millions to build and probably tens of thousands to operate each month is fairly paltry. And revenue is probably far below $80K because of roaming, in which AT&T will see quite a bit below $7.95.
Free makes sense in this equation because reducing the cost of collections and marketing reduces overall recurring operating expenses by a large factor. Increasing passenger satisfaction brings in more travelers who would then choose air over rail or car because they could get work done while waiting for flights. An addition passenger brings in several dollars (or more) of fees for the airport where their share of Wi-Fi revenue is probably substantially lower. Placing free service in the food court also increases revenue for concessionaires, which in turn produces more tax and franchise revenue for the airport.
(On a math-related front, the city's aviation chief said usage was "growing exponentially," but doubling in a year is an arithmetic progression by most definitions of both words unless it doubles again this year or doubled from the year before last.)
The city and a WISP both plan Wi-Fi service: The city wants to find a provider to install Wi-Fi across the town at the provider's expense, while Bright House Networks will expand a small existing footprint to more of the downtown. Five companies responded to the cities RFP, but not Bright House; the area to cover is 60 square miles.
Bright House offers cable TV and broadband locally, and will provide Wi-Fi free to its existing cable customers. Walk-up service is a buck for 30 minutes up to $9 a day. I have a great sound bite at the end of this article. When asked by the report whether the jury was out on metro-scale Wi-Fi, I rejoined: "The trial hasn't even started."
We discuss how the hotspot market has evolved over six years: Rick, the founder of Surf and Sip, was one of the first people I interviewed back in late 2000 for my early article in The New York Times on public hotspots that ran in Feb. 2001. (Rick said he still has the article on his wall.) This was the first article in the mainstream media looking at hotspots, and it's what pushed me on the road to writing this blog. Rick had over 100 locations at the time; his network has now grown to nearly 600 in Europe and the US.
Surf and Sip started by actually soliciting business, pounding the pavement, paying for equipment and DSL lines, and offering high-touch service. The company has shrunk in size to just a handful of people while running a multi-national operation as the complexity as reduced, and their in-house software has been developed.
We talk about whether hotspot growth has stalled, where revenue comes from, and the future of VoIP at hotspots and municipal networks. Rick has some interesting insight into Wi-Fi handsets and music devices like the Music Gremlin which require specific network resources or lots of bandwidth. [42 min., 20 MB, MP3]
Airborne cell usage is coming; you can't stop it: For this article in the Economist, I spoke to many people involved in parts of the aeronautic world, including regulators, airlines, and operators, and found that the juggernaut for in-flight calling, messaging, and limited cell data is bearing down upon us. The US carriers appear to have little interest in voice, possibly because of the general negative attitude about it, and AirFone's lack of traction (albeit at their high price for poor-quality calls). European carriers, who typically haul people over shorter distances, appear to have greater interest. Service should commence on as many as six airlines worldwide by fall 2007, with some much sooner.
Ryanair told me their average flight is 90 minutes, and people are already high engaged in activities on board. They will like spend something like $20m--my very rough estimate--to add service to all their planes, but with calls priced at $2.50 per minute, they could gross something on the order of $15m per year from just 50 minutes of calls among all passengers made per flight and hundreds of text messages. Text messages should cost well under a dollar each, although prices haven't been set.
What we Americans fail to understand when we see $2.50 per minute as an onerous rate is that Europeans have been used to paying insanely high cross-border roaming charges, sometimes with the same carrier owning both networks. EU regulators have started poking at these charges, which companies immediately started dropping. Nevertheless, $2.50 per minute for the privilege of roaming in the air just doesn't have the same horrible resonance as it does for those of us who pay $80 per month for 1,000 prime and 5,000 weekend/evening minutes.
If you'd like some amusing reading, you can download comments made by individuals and other interested parties to the FCC for proceeding 04-435, which was the FCC's public comment query on the use of cell phones in flight.
My article didn't address the broadband side of the equation much because that's going to be a later development. I had one in-flight operator tell me that satellite bandwidth will be prohibitive to use in the way that Connexion resold it. Inmarsat's fourth-generation satellites may offer 432 Kbps over a single channel, but they charge $8 to $12 per megabyte for a ground-station end user's data. What that might be bundled for in the air is still unknown. GPRS will be likely from day one because the speed is low enough that it would be hard for someone to spend a fortune.
The US picture is much simpler because AirCell and JetBlue purchased air-to-ground licenses, which require substantially cheaper operations. They don't have to launch satellites, deal with long latencies, or pay middlemen for any service. AirCell expects 1.5 Mbps throughput in each direction, which is ideal for...lots of voice over IP conversations.
The long-delayed network starts up: The network was bid and being deployed in early 2005, but coverage issues and other undisclosed problems led to the town dropping its initial vendor and rebidding. This article from the local paper doesn't mention the new provider, but does talk about moving transmitters and upgrading software. Service costs $19.95 per month, with a bridge available for a $5/month rental fee.
The city of lights in Illinois gets first phase of Wi-Fi network installed: This network deserves a short shout-out because it's the first significant non-Bay Area system that MetroFi has installed; they may also unwire an adjacent Illinois town. This very first phase will comprise just five nodes, and cover only part of downtown, but it will provide access by Sept. 15 for the Literary Festival. MetroFi will eventually install 600 to 900 nodes within about a year.
MetroFi's big win was Portland, Ore., which is underway; they have also picked up a number of new Bay Area towns. They did not win the bid for the Wireless Silicon Valley project, which will include a few cities they now cover.
Toronto Hydro Telecom opened the first phase in its city-wide network: The utility owns poles, purchased in an earlier deal from the city, which may make it simpler and cheaper to deploy. They already own equipment and know how to put gear on poles, something which has been an issue and sometimes a large expense in other networks. And they own hundreds of miles of fiber, too.
The first segment of the network covers part of the business district, and will grow by December to six square kilometers. Service is free for the first six months, and then switches to Cdn$5 per hour, Cdn$10 per day, and Cdn$29 per month.
A few weeks ago, I noticed that a second broadband wireless ISP was announced in midcoast Maine: I wrote about this at the time, as it seemed remarkable that an area with just tens of thousands of residents could support two firms. The companies' founders both commented on that post, and that's interesting reading, but I wanted to hear some additional detail. I got Jim McKenna, the founder of the newer firm, on the phone to chat about the advantages of wireless mesh in areas where real estate rights are easier to obtain, mountains abound, and customers have few other options. That's featured in today's 20-minute podcast. (Yes, I kept a podcast under 40 minutes.) [20 min., 10 MB, MP3]
He pointed out in the interview that the numbers for Maine are pretty staggering bad for broadband adoption, and he thinks it's about price. He said that one survey showed that only 10 percent of households passed by DSL or cable subscribed to that service. Maine is not only a poor state, it's largely rural, and I would guess that out of 1.3m residents and over 500K homes, that perhaps only 30 to 40 percent are passed by DSL or cable.
The largest cities in Maine--Portland, Augusta, and Bangor--have about 20 percent of the state's population. Everyone else is spread out. In Knox County, where Red Zone Wireless and the 1995-founded Midcoast Internet Solutions have their headquarters about two blocks from each, there are only 40,000 residents, and about a quarter live in and around Rockland, the county seat.
McKenna says that $20 per month is the right price point, and Red Zone offers a residential Wi-Fi-based service with broadband rates for that price and $50 setup; no contract, no cancellation fees, either. That's 500 Kbps down and 128 Kbps up, or about 10 times the download and four times the maximum speeds of a 56K modem. (Although when you get to some of the towns in Red Zone and MIS's coverage area, you're not seeing modem rates that high, either.)
I have a query out to talk to Midcoast Internet Solutions's founder, too, to compare notes. They started as a dial-up provider in 1995, and added Breezecom (now Alvarion) wireless gear in 1997.
This is bad news and good news for OnAir, in-flight broadband and telephony: You know it's bad when Reuters misses the key point in a story. In this article about Inmarsat announcing a launch date for its third satellite--its Asia/Pacific geostationary orbiter--to complete its fourth-generation, broadband satellite network, two key facts were missed. First, the third bird should have been in the air as long ago as 2005. Last year, it was expected to launch in the first half of 2006. Now we're looking at late 2007.
The second critical fact is that while Reuters dutifully reported the fact that Inmarsat's 4G network covers "85 percent of Earth's landmass," the network does not cover the Pacific Ocean, in large part. While Inmarsat has tried to bill its 4G network as being the first affordable, targetable worldwide broadband network, using beamforming arrays that aren't fixed for the life of the satellite, it's pretty clear that aviation and maritime use have to form a chunk of the service's revenue. There are generally mobile broadband alternatives in all but remote areas that would be cheaper than what Inmarsat calls BGAN (broadband global area network), except at sea and in the air.
The third satellite's launch is tied, in this story, to Inmarsat's deal with Aces International to offer handheld satellite phones in Asia.
Now this affects OnAir because the Airbus/SITA joint venture, which incorporate the assets of Tenzing Communications, based their in-flight cellular and broadband platform on the 4G service. When I spoke to OnAir's CEO recently, I said he must be extraordinarily patient. He said they were in it for the long haul. Now that haul is even longer. The company will launch extensive European services in 2007, however--pending regulator approval--starting with an Air France plane factory equipped with Inmarsat's 4G avionics system and an OnAir picocell. And they scored the 200-plus fleet of Ryanair planes, too.
But they can't serve Asia or US-Asia trans-Pacific routes, which I believe were a big market for Connexion's service due to the flights' duration. OnAir's competitor Aeromobile can provide cell service over Inmarsat's globe-spanning third-generation network, which operates about 1/8th as fast, and thus limits the number of simultaneous calls per flight to a relatively small number. Still, Aeromobile is talking about a near-term launch over Asia, and just announced a trail with Qantas.
Dana Spiegel of nycwireless loses his temper with the Park department: The contract with the New York City Park Department nearly two years ago required Wi-Fi Salon, the winning bidder, to build out Wi-Fi service and pay the city some fees. Newer deals no longer require fees, because that appeared to be counterproductive to getting service running, and Parks put a deadline of about a month ago on having at least Central Park up and running. Wi-Fi Salon announced a deal with Nokia for sponsorship and technology. And now...no word, Spiegel notes. Visit Wi-Fi Salon's News page and you'll note a lot of links to the coverage of "upcoming" access, and no information on the site as to when and where service exists. The deadline of a month ago was an ultimatum, as correctly reported by the New York Times; it was reported most everywhere else as an announcement date as to when Central Park would have live Wi-Fi, a mistake. Update: See comments below; Klaus Ernst has tested six of eight Central Park hotspots and was able to get a reply from Wi-Fi Salon about launch.
Spiegel notes that nycwireless has been busy this summer without a city mandate: "Over this past summer, NYCwireless has brought online a number of new hotspots, including Brooklyn Bridge Park, Stuyvesant Cove Park, and Madison Square Park. We’ve launched (through the work of students at Monroe College) hotspots at a bunch of restaurants and gathering places in Brooklyn, the Bronx, and Harlem. We’ve upgraded some of our hotspots that provide free Wi-Fi for affordable housing residents at some Dunn Development & Community Access buildings."
He also discusses in brief a situation he had alerted me to via email where the Park Department has prevented Friends of Dag Hammarskjold Plaza from installing nodes in that public space without a spurious and poorly defined insurance requirement. The department also demanded that Madison Square Park's access be turned off--it was fully operational--so as not to embarrass the Park Department while it's own separately authorized hotspots weren't running.
Is there a reporter in New York who wants to run with this? This sounds like a good story, with threats, governmental interference in private and non-profit efforts, attempts to control unlicensed spectrum, and incompetence!
An errant press release seems to have slipped the news that Azulstar, Cisco, IBM, and Seakay's group--Metro Connect--will unwire Silicon Valley: The effort by the Joint Venture Silicon Valley Network to develop a Wireless Silicon Valley proposal and then find viable vendors appears to have reached the end. Seven companies and consortiums originally responded, and that number was winnowed to three. I predicted from the beginning that this four-firm consortium with two blue-chip tech firms involved would be the only one with the right combination of plans for deployment, technology, social purpose, experience, and financing. Cisco and IBM have a very large number of employees in the region under consideration; IBM specifically noted that they have 7,000 in the Bay Area, and is the second-largest San Jose employer.
The Metro Connect proposal will feature 1 Mbps free service and span 1,500 square miles, encompassing many different municipal entities. Fee-based services will include faster speeds, voice offerings, and video streaming. There are also plans to incorporate WiMax.
There are nearly four dozen cities, mostly in San Mateo and Santa Clara counties, and the counties themselves. Fremont and Newark in Alameda County and Santa Cruz in Santa Cruz County are also signed on. Fremont is one of the largest cities by area in California at 92 square miles, just by the way, so it's a significant addition. Other cities may join later. Yesterday, Los Altos and Cupertino said they would join the project.
The participation by cities in this proposal is, however, a bit complicated. San Mateo County's telecom authority (SAMCAT) issued the RFP and has Joint Powers Authority for 18 of the 35 original cities signed on, and there's an intent to create model agreements that each city could then sign off on. However, the addendum to the RFP, issued in June, notes on page 7 that "each city will independently ratify the winning vendor's proposal for their geography and jurisdiction." Also, antenna attachment agreements aren't included as part of the RFP. And the winning vendor needs to work with existing WISPs, not just displace them.
This may be rather complicated, but given the way this was structured, I suspect that most of the approval process will go swimmingly and some percentage of cities will have specific needs, concerns, or objections. The old 80-20 rule, or perhaps 95-5. The details will emerge over what I expect is several months.
While Silicon Valley is often thought of as a high-tech paradise, development centers along highway corridors and city centers. There are plenty of less-populated, less-served, and once-considered-rural areas in this giant region.
InspiAir says they can span kilometers with normal 802.11b in point-to-multi-point formations just via software: Yeah, and I remember Karlnet, too. Back about five years ago, Karlnet sold firmware upgrades for common reference designs, like Apple's AirPort Base Station, that would allow, wouldja figure, timing changes and other protocol tweaks that allowed 802.11b to span several kilometers in P2MP configuration without failing.
InspiAir seems to be implying that they have mesh technology, too, according to this Techworld account. I don't think so. One of the investors is quoted talking about the "not-yet-patented technology." Given prior art, I would guess that might be "never-to-be-patented technology."
I don't quite understand why they've made a splash. The "100 milliwatt" radio they talk about using would be the unamplified signal coming out from a card. Stick a sectorized antenna on that and work within point-to-multipoint rules and you're allowed (in the US at least) to push out pretty high effective wattages that can span many miles.
It is, in fact, a violation of the law physics if you were using 100 mW of actual output power measured at the antenna and expecting to reach a few kilometers. Or more than a few thousand feet line of sight in a non-interferring, non-reflective environment.
Such as in a park. In the Techworld article, an operator in New York which provides service in the Hudson River Park is quoted as saying that they were astounded by how two nodes from InspiAir could cover a three-quarter square mile park, where 25 to 50 nodes is often quoted for outdoor deployment for a square mile. Yes, but that's a square mile of dense usage with obstructions. It's not a big deal to use high-gain antennas cover 90 degrees of arc at opposite ends of a square to cover that range. That's only a few hundred feet from either transceiver. When the leaves come back, then we'll talk about coverage.
802.11n is lurching towards completion, and we talk about when, where, and why: Matthew Gast was my guest Wednesday for a live, in-office podcast in which we talked at great length about Task Group N, the group within the IEEE 802.11 Working Group, that's attempting to push faster Wi-Fi at the world. Matthew is a voting member of the group, and we talk about the politics, the technology, and, you know, how fast this stuff will actually be when it's released.
We also delve a bit into voice and why 802.11n is important for handheld devices and how handset manufacturers worked hard to get their interests served in the compromises that appear to have led to current state of the protocol.
Matthew is the author of O'Reilly and Associates's 802.11 Wireless Networks: The Definitive Guide, which came out in its extensively revised and updated 2nd edition last year. [44 min., XX MB, MP3]
Nokia looks to add location-based features to cell phones: While this is a Wi-Fi blog, I'd be remiss in not noting the coming closer-integration of global positioning satellite (GPS) receivers in a greater variety of equipment, including phones and cameras. (Ricoh and Pentax sell cameras with GPS integration, while Sony offers a GPS-timestamping device that can be synchronized via software later with the pictures you took.) And recall that Skyhook Wireless offers a Wi-Fi-based GPS-like service that could be integrated into mobile phones to work in concrete canyons in which GPS satellites are hard to spot, but cell towers and Wi-Fi transmitters are abundant.
AOL's Mapquest division, internetnews.com reports, released a GPS-based navigation system for mobile phones last April. Flickr, the Yahoo-owned online photo sharing and archiving service, just added geotagging this week, which allows photographs to be labeled with latitude, longitude, and level of map detail, and then clustered on maps.