Thousands of people read this page every day. You can reach them in this spot by sponsoring Wi-Fi Networking News

The above could be a paid, sponsored link. Email for more information.

Subscribe to essays from this site via email. Email to subscribe, or sign up via your Yahoo account.

Red Herring leaves unusual smell: It's my unfortunate duty to shred an article written by a colleague in The Red Herring about the wireless hot spot business. I don't know Dan B____y, and my apologies to him; this won't be pleasant. (Update!: I swear it's not my fault, but the Wall Street Journal is reporting at 11.30 am Pacific today, that Red Herring's March issue is its last.)

Third paragraph: Four years ago... Larry Brilliant put together AerZone, which was another big idea with dubious prospects in a sea of bloated expectations. AerZone was a division of SoftNet, which had also purchased Laptop Lanes. AerZone, by late 2000, had contracts with several airports and airlines to provide service. In mid-December, four days after I had interviewed the CMO, SoftNet closed the division and put Laptop Lanes on the block. They couldn't raise additional capital. Part of AerZone's model was, as with other wISPs at the time, to pay for everything.

But, the company and the wi-fi market it hopes to exploit are starting to look like a hype bubble -- some much-needed skepticism injected into the overall Cometa coverage. Much of the coverage has focused on the scope and nature without drilling into costs and practicality.

Wi-fi is yet another example of a grassroots technology for which early adopters have grown accustomed to paying little or nothing. It really depends on which audience, and they're hardly early adopters any more. People who use Wi-Fi consistently on the road are entirely accustomed to paying for it. I'm not sure who gets it for free unless they're conveniently near a community node or an open node.

Now we get into muddy waters; put on your waders. ...David Hagan, president of Boingo, one of the pioneer wi-fi service providers, with 800 hot spots in hotels and airports around the United States, and a staunch believer in the build-it-and-they-will-come philosophy of customer acquisition. Actually, Boingo believes in the you-build-it-and-we-will-come model, because they aggregate not build infrastructure.

Now the author starts sinking into the mire. Boingo's expensive philosophy is precisely what Cometa intends to avoid. (Boingo raised $20 million in venture capital a year ago and says it has two years' worth of cash left.) Boingo is developing software and partnerships, so their philosophy is quite cheap compared with infrastructure build out. Boingo is relying on distributed infrastructure in which no one company winds up building out massively, like Cometa, and bearing huge costs, but rather builds strategically and relies on aggregators to fill their pipes.

Cometa is relying on the widespread corporate adoption of wi-fi, a trend that is still nascent. It might be my InfoWorld mentality here, but the penetration of Wi-Fi into the corporation is nearing mid-term, not nascent. Within two years, every laptop that a corporation buys will either have Wi-Fi preinstalled or prebuilt (a la Centrino), or will be a mandatory part of configuration. Many corporate surveys show most corporations have already deployed wireless LANs or are in the middle of testing; most of the rest, a small number, plan to install WLANs soon. Security remains the stumbling block, but not much of one.

This next part is terrific, however, and delightfully expressed. Cometa's stated goal in creating its 20,000 hot spots throughout the United States is to provide wi-fi access within a five-minute walk of any urban point or a five-minute drive of any suburban site. But experts say it will take far more than 20,000 sites to accomplish that ambitious goal, especially since it requires several hundred hot spots just to cover a decent-size college campus.

Now, we get into mixed market issues. Given the players, it seems appropriate that Cometa is making a bet on the enterprise market. After all, home networking enthusiasts have been known to be stingy. But analysts predict that growth in the small office/home wi-fi market will far outstrip the enterprise wi-fi market. But that's hardware sales and installation, not service plans. How would Cometa serve a home audience? It makes no sense. Intel is providing hardware for consumers and businesses, so their part of Cometa is following that market.

The author now provides an example of why listening to the company you're interviewing too closely causes you to drink the blue zombie soup and repeat their tropes. With IBM (site installation), Intel (wi-fi chip manufacturing), and AT&T (broadband access) running the show... (Just by the way, this illustrates how Cometa isn't the core of these companies but rather IBM Global Services, not IBM proper; Intel, not Wi-Fi chip manufacturing, but Intel Capital; and AT&T's bandwidth division.) ..., however, corporations are likely to feel more comfortable exchanging sensitive data wirelessly than through, say, companies called HereUAre or Surf and Sip. Dude, hereUare has been out of business for months. And Surf and Sip runs as tight a ship as any of the other ISPs; they're even Intel Centrino approved (see the bottom of this article). What, the brand name makes corporate execs nervous? I'm not sure marquee names make CIOs less jumpy given what happened to Worldcom.

And Cometa executives are confident they can drive standards through the industry that will allay any security fears that corporations may have. Bzzz. Thank you for playing! A multi-billion dollar industry is going to roll over for a hot spot operator. Please pick up your consolation prize at the door. Oddly, a broad consortium of companies and individuals in the IEEE, IETF, and Wi-Fi Alliance, among other groups, have created a variety of reliable security standards that will allay the fears as they are deployed over the next nine months, some of them in the next few weeks.

I heard Larry Brilliant speak a few days ago, and he spent quite a while discussing how CIOs want secure, audited hot spots. He also explained how corporations are using VPNs. Because no VPN technology has been compromised that I'm aware of, there's no need to audit hot spots. With a VPN tunnel and an appropriately configured laptop, a hot spot's security is irrelevant -- which is the whole point of a VPN in the first place: to allow use of untrusted networks! CIOs may be asking for this, but it's education on the subject that they need, not audited hot spots.

What makes more sense is for Cometa and all hot spot operators to participate in the public drafts of the wISPr (wireless ISP roaming) proposal that came out of the Wi-Fi Alliance a few days ago. That proposal could become the working document for building well-run hot spots, requiring spot checks through an independent group instead of individual audits or separate standards for each network. If I can bring up the specter of MobileStar twice in two days, the CEO of that departed firm told me in fall 2001 how important it was for them to only partner with companies that matched their high standards.

Finally, we get this amazing quote, which is a coup for the author: Rose Klimovich, the Cometa representative at AT&T, says it best when she admits: "We have to figure out the right way to make money. We will see over the next year or two whether Cometa has figured that out. It hasn't been tested yet." Dear Lord almighty, this is how the telcos still work: throw money at the wall and see if it sticks! (The trick? Soak the money in your own flop sweat first while you think about the future of your career if it fails.)

I feel dirty.

The Red Herring article aside, except for the good parts, it's clear that Cometa has an interesting plan for low-cost hot spot buildout. Because they're using AT&T for bandwidth, they don't suffer T-Mobile's high T-1 costs. Because they're a bunch of giant companies, they get sweetheart highest-volume discount rates on the equipment they install in each location. Because IBM Global Services is involved, again, they're not paying full cost of freight; for all we know, IBM's investment is donated labor and materials for the hot spot buildout.

What we really don't know about Cometa are the following issues:

• How are they going to find 20,000 real-estate venues to deploy? Outside of major hotel chains (many locked up), coffee shops (Starbucks locked up), and airports (Cometa specifically disclaimed airports as a market, and executive lounges of the three bigs are locked up), there aren't enough venues of scale in a chain to get big wins where business travellers go.
• Are they receiving co-deployment fees from venues they go into rather than bearing the cost themselves?
• Since they will only aggregate their service to brands that will market them, will they also make their network available to iPass, GRIC, and Boingo? Or, for that matter, Surf and Sip, FatPort, and Wayport?
• Will they partner with cell operators to provide co-marketing or other services?
• Are, as the author of this article correctly points out, 20,000 locations enough to make a real impression -- especially if it doesn't include many hotels, airports, or coffee shops?

I sound like a cynical jerk, right? I'm not trying to dis Cometa, but rather explain clearly why the degree of detail they've released so far makes it sounds as though we're living through the late dotcom era again. It's unlikely Cometa will release more details because it doesn't benefit them to do so. But their potential partners and investors will continue to be have questions like the ones I raise--and not everyone can be disclosed under contract to learn the nitty gritty.

Later Thought: The notion of Cometa pushing security standards is actually at odds with Intel's own Centrino co-marketing program in which Intel will be partnering to lend their name with firms like FatPort (announced yesterday), Surf and Sip, and various hotel operations (this last week). (Centrino is Intel's new Wi-Fi-integrated, lower-power laptop OEM system.)

Intel's name means a lot more to CIOs than Cometa could in the short term. The other program is the Wi-Fi Alliance's Wi-Fi Zone. If that program is successful coupled with their wISPr proposal, the most recognized name in wireless networking, Wi-Fi, becomes the mark that achieves CIO approval.

In a press release about FatPort and Intel's program to test FatPort's FatZones as Centrino compatible, an Intel exec said this: "Intel Centrino mobile technology is designed to enable a rich, wireless mobile computing experience. We're excited to be working with FatPort to bring these services to Canadian locations." Cometa is competing against arms of itself, lending credence to the notion that Intel's involvement in Cometa is farther away from the center, as I note above, than the press releases would indicate. Cometa is an investment; Centrino is a core initiative.

T-Mobile Pricing Update

T-Mobile offers $30 yearly commitment, $40 month-to-month: T-Mobile has posted its new pricing plan and it's not exactly as reported. Gone is the $30 per month yearly commitment regional plan and the $50 per month unlimited national plan. In its place are two national plans, both unlimited: $30 per month for a yearly commitment, $40 a month for a monthly commitment. This is a nice compromise for the road warrior versus occasional traveler.

The bandwidth limit has been removed: unlimited data transfers on all accounts.

Per-hour rate is $6, not the per-day rate, which makes more sense in terms of worrying about people camping out for the day. Interesting choice on the per-minute part of per-hour: you can prepay for 300 minutes for $50 and use minimum 10-minute increments. If you pay by the minute at 10 cent a minute, you have a minimum 60-minute billing or $6.00. So the least you pay with the prepay plan is 60 cents versus $6.00 on pay-as-you-go.

I read the fine print on the new service agreement. First of all, it goes into effect March 1, not today, although they might implement changes earlier. Second, there are some penalties for cancellation of the monthly plans: $200 after the first 30 days for the yearly commitment plan (which is only $360 per year total); $25 after the first 30 days for the month-to-month plan.

The minutes expire 120 days from purchase or refill.

Without the day rate, T-Mobile is still out of sync with the rest of the hospitality hot spot industry (except STSN), but the new monthly rate makes harmonization possible, at least.

Other News

Internet Exchange launches UK wireless service at prices undercutting British Telecom: This Internet cafe has added Wi-Fi at 30 stores for five pounds (about US$8) per day; BT charges 15 pounds (over US $25) for the same service. Other rates are comparably cheaper. This is the week to cut prices, it seems.

Toshiba to blanket Canada with 1,000 hot spots: As the North American market heats up, we need to look north to our less temperate--I'm talking Celsius!--brothers and sisters. Toshiba of Canada. The article mentions Spotnik (no points deployed yet) and Bell Canada (testing but not charging), but not FatPort. Of course, the Global and Mail is based in Toronto and the service discussed is mostly Ontario, while FatPort is in Vancouver, BC. (North of the border insider joke: What did you do to annoy the G&M, Sean? Praise Conrad Black?)

PC Mag interviews Dennis Eaton, sidesteps hard question: The interview is solid, but the writer fails to note Eaton's dual role: Chairman of the Wi-Fi Alliance, but also a marketing director at Intersil, a company that has introduced draft 802.11g products like many others in the industry. A new piece of information: Eaton expects USB 2.0-based 802.11g adapters to be popular.

Wi-Fi over power: Amperion is offering an interesting bit of technology. They have a system that lets an electrical utility encode bandwidth over powerlines, but here's how they solve the step down problem -- the problem of getting bandwidth from the powerlines to the home through transformers. They're using 802.11b signaling from the poles to a customer premises equipment (CPE) device called PowerWiFi. Where they're clever additionally is that the CPE could be individual computers if there's a strong enough or clear enough signal or to some sort of bridge. They mention HomePlug, an in-house data-over-power system, but I assume they'd still be bridging via an external device into the house's power.

Atmel integrates: It's a bit techie, but part of a larger trend in chipmaking. Atmel has stuck processors, two Ethernet controllers, a USART, and a Wi-Fi MAC onto a single chip. You'd still need a baseband/PHY chip, which Marvell makes as a single CMOS unit, just by the way.

Why is Dan's name not spelled out?: I posted the author's name so that I wouldn't be a coward in criticizing his work directly. However, it struck me yesterday that I'm branding this guy because of this blog's high Google ranking. If I spell his name out, that means that every time someone searches on freelancer Dan B, they'll get this page among the top listings, shredding his article. (I guess the last issue of Red Herring didn't get the kind of scrutiny that it might otherwise have.) This would be unfair, and thus I've removed his full name.