May 16, 2008

Wee-Fi: Go, Go, Wires! Go, Go, Cablevision!

By Glenn Fleishman

OSnews explains why wiring a house still makes sense in the 21st century c.e.: A very well-reasoned article from OSnews explains why the site still backs residential wiring. They’re involved in the build-out of a Utah home partly as a technology demonstration, and they’ve put coax and Cat 5E Ethernet cable throughout, as well as conduits for future wire pulls. Fundamentally, wire has more capacity; I’d argue it does across several dimensions, too. You can run 1 Gbps raw across a Cat 5E or 6 Ethernet cable in both directions at the same time versus best performance of unidirectional nearly 100 Mbps in my testing of Draft N. But you also get switching with Ethernet—multiple simultaneous symmetrical 1 Gbps—and if you need more capacity you simply pull more wires and put in more switches. Wire is cheap and switches are now, too. It’s a good read if you’re thinking of rewiring (or unwiring) your home.

Cablevision’s already started its rollout: An observant tri-stater at the Cable Rant site spotted Cablevision installers putting up BelAir gear on their cable line. He took some photos.

Posted by Glenn Fleishman at 11:12 AM | Permanent Link | Categories: Wee-Fi | Comments (0)

May 15, 2008

MetroFi Plans Market Exit: Sale or Shutter

By Glenn Fleishman

MetroFi will sell its networks, but plans to shutter if there are no buyers: Ah, folks, the trifecta has arrived, and I’m nothing but sad about it. MetroFi’s chief Chuck Haas emailed me this evening with the news that his firm has decided that they will sell their networks in nine cities, including their first cities in the Bay Area (Cupertino, Santa Clara, and Sunnyvale), and their largest muni deployment in Portland, Ore. If no buyers emerge—including the cities in question—Haas said that MetroFi would have a shutdown plan for gradually unlighting the networks.

MetroFi was one of the three most prominent pure play metro-scale Wi-Fi firms, if you count EarthLink’s municipal wireless division as a separate operation, and Kite Networks, which was a subsidiary of a larger telecom firm. Each company had made a unique network hardware choice—MetroFi, SkyPilot; Kite, Strix; and EarthLink Tropos plus Motorola—and each had a sort of specialty. Interestingly, a fifth firm, BelAir powers Toronto (a small but super-fast Wi-Fi network) and Minneapolis (the only putatively completed large-city Wi-Fi network), and will be behind Cablevision’s nearly $350m New York Wi-Fi plan.

MetroFi was the only major firm to back ad-supported no-fee access, coupled with paid, no-ads service, and higher tiered commercial offerings. They built mostly smaller cities, with Portland being their only real big city win. The firm began with the notion of building Wi-Fi out gradually as a way to provide broadband in communities that lacked service, with no municipal involvement. That plan required sparser networks and typically a home signal booster designed by SkyPilot. (Kite mostly focused on the Southwest; EarthLink on big cities.)

EarthLink was in many ways largely responsible for the mess that all Wi-Fi providers found themselves in last year by offering to build Philadelphia’s network back in 2005 at no cost to the city—in fact, paying the city and the local utility fees. That set the stage for nearly all the RFPs that followed where, if EarthLink were a bidder or the city was aware of the alternatives, the notion was that no city dollars would be spent, even if taxpayer money wasn’t “at risk”—that is, even if a city could save money by switching current line items in their telecom and data budget to a wireless network.

Haas noted via email that MetroFi has been working towards anchor commitments by cities for nearly two years, but the inertia of those early networks led municipalities to reject those options. In Toledo, where MetroFi had negotiated an anchor commitment, a change in administration led a new mayor to retreat from the plan.

Is there a future for metro-scale Wi-Fi? Yes. With thoughtfully constructed, outdoor-focused deployments centered on municipal purposes, with public access a secondary issue, it seems like these networks could still provide an inexpensive way for relatively high bandwidth compared to the alternative of cell data networks.

However, that advantage is likely short lived in larger markets. The near-future certainty now that there will be multiple provides offering wired broadband speed service starting later this year with Sprint/Clearwire’s WiMax, and continuing through into 2012 with significant network buildout by Verizon and AT&T in several bands (including their new 700 MHz holdings).

While Sprint/Clearwire is talking about 120m to 140m homes passed by 2010 with their network, obviously focusing only on major markets, many of the 700 MHz licenses purchased by AT&T and Verizon carry buildout requirements with penalties. So cities outside the top 100 population markets and rural areas will still see some benefit. In those mid-tier markets, there’s also the 3.65 GHz band for shared licensed use, which is a model that Azulstar is pursuing with new WiMax deployments, as I wrote about recently.

Competition will likely push the cost of mobile broadband far below its $60 per month 2-year contract rate of today, which then would beg the question why a city or county with good commercial coverage would need to build its own Wi-Fi network. There are still plenty of reasons to build dedicated, first-responder 4.9 GHz public safety networks, of course.

I’ve always described Wi-Fi on a metropolitan scale as the best, worst technology. The best, because everyone has Wi-Fi in their laptops and increasingly in handhelds and gadgets. The worst, because the technology is absolutely not designed for the purpose, unlike CDMA and GSM evolved cell standards and mobile WiMax.

It’s possible that in the long term, looking five years out, that Wi-Fi on a metro-scale will only be needed in small towns, odd markets, and for highly particular purposes. Or, perhaps in a bit of irony, where companies like Cablevision feel Wi-Fi is necessary to retain the loyalty of their highly wired customer base.

Posted by Glenn Fleishman at 9:47 PM | Permanent Link | Categories: 4G, Financial, Free, Metro-Scale Networks, Municipal | Comments (1) | TrackBack (0)

May 14, 2008

Phila. Gives Up on EarthLink

By Glenn Fleishman

IDG News Service reports that Philadelphia won’t pursue further efforts to keep the EarthLink network up and running: The last paragraph is quite classic:

Without going into details, city spokesman [Douglas] Oliver said there clearly were maintenance and upgrading challenges that came with the free infrastructure. “How many times has someone not taken $17 million worth of something without there being a pretty good reason?” he said.

Posted by Glenn Fleishman at 8:55 PM | Permanent Link | Categories: Municipal | Comments (0) | TrackBack (0)

Thalys Hits Glitch in Impressive Train Launch

By Glenn Fleishman

Thalys has launched Internet service on high-speed train routes between Paris, Brussels, Amsterdam, and Cologne: The service hit glitches in its big press rollout, but glitches shouldn’t be mistaken for actual performance. The satellite-backed service pulls down 2 Mbps of ruinously expensive backhaul, compressed to provide speeds that feel like 4 Mbps. (Read: faster for email, TIFF images, certain PowerPoint presentations, and Web pages with gzip disabled; normal rate for JPEGs, GIFs, compressed Web pages, and PDFs.)

The service will cost first-class passengers not a thing, but coach will pay €6.50 (US$10) per hour or €13 (US$20) for an entire trip. The train operator is initially equipping 7 trains, but will complete work on all 26 trains by October. Trip durations run from 1 hour 20 minutes to 3 hours.

Most impressively, the consortium that built the system is using a pretty modest antenna that moves automatically to stay in contact with the satellite. It’s 80 by 72 cm (31.5 by 28.3 inches), and plans are to shrink that to something 2/3rds the height when a new dish is certified. Ultimately, IDG News Service reports, the group plans to use 3 cm (1 in) high phased-array antennas that would cover the train’s roof. Very, very clever, as it jettisons any moving parts.

Three companies worked on the technology: Telenet, handling the billing and authentication, is a Belgian ISP that also runs hotspots; Nokia Siemens is a well-known systems integrator, and is providing some gear and handling installation and integration; 21Net, perhaps the least-well known partner, has the satellite technology.

This project dates back to at least 25-April-2005, a point at which 21Net and Nokia Siemens announced a successful test on the Thalys run from Brussels to Paris.

Posted by Glenn Fleishman at 3:50 PM | Permanent Link | Categories: Rails | Comments (0)

May 13, 2008

Mobile Post: Where’s the Price?

By Glenn Fleishman

Why don’t service providers tell you what it costs? In this mobile post, I inveigh against the practice of hiding one’s prices coyly, like the menu at an expensive restaurant.

Posted by Glenn Fleishman at 11:31 AM | Permanent Link | Categories: Financial | Comments (1)

Wee-Fi: iPass Flies; Riverside (Calif.) Approaches

By Glenn Fleishman

iPass announces roaming deal with Aircell Gogo in-flight network: Gogo isn’t yet aloft, though it’s well into testing, but iPass has a contract in hand to allow its subscribers broadband access. The press release sidesteps cost, and an iPass spokesperson clarified for me that pricing hasn’t yet been determined; additional fees will be required, but how much isn’t yet set. Given that Aircell has spoken about fees of about $10 to $12 for cross-country flights, iPass can’t include unlimited service at a fixed rate, I wouldn’t think. But many terrestrial venues charge $7 to $12 for 24 hours access and iPass, Boingo, and other retailer partners pay those venues as little as 50 cents per session. (Correction: This item originally stated that iPass wasn’t planning at this time to charge for Gogo service. That was an error: iPass hasn’t yet set the fees, but expects to charge something on top of their flat rate.)

Riverside network should launch soon: I recall a ribbon (or cable) cutting ceremony for AT&T’s MetroFi-built Riverside, Calif., network some time ago, but the full launch beyond a trial network in 2007 appears ready to go by the end of May. The network was originally billed as planning to cover the 80+ sq mi of the city; this article says just 55 will be covered. And AT&T’s local project manager told the audience at a training session, that the service is “mainly meant for outdoor use.” Huh. Service is free with ads at a rate that’s not easily found (512 Kbps?); a premium ad-free service at 1 Mbps is free to AT&T’s 1.5 Mbps or faster DSL subscribers and fiber users, as well as by paying a monthly rate that isn’t yet disclosed. The 24-hour rate is a crazily high $7.99.

Posted by Glenn Fleishman at 10:15 AM | Permanent Link | Categories: Air Travel, Municipal | Comments (0)

EarthLink Will Shutter Philadelphia Network, Company Says

By Glenn Fleishman

It’s the end of the cycle, folks: The first shall be last and the last shall, apparently, be first to sue. The Philadelphia Wi-Fi network will be shuttered under plans by EarthLink that they announced via press release today.

The company plans to pull all its gear from the poles starting 12-June-2008. The company’s press release said it offered to give the network at no cost to an unnamed non-profit, as well as to the city, but claimed that “unresolved issues” led to the effort falling apart. EarthLink offered cash and more equipment, as well, in undisclosed quantities. Wireless Philadelphia, the non-profit in charge of managing the network provider and administering digital divide programs, was apparently not the non-profit mentioned.

EarthLink filed a lawsuit to allow it to remove its Wi-Fi nodes and cap its liability at $1m. That’s a pretty hostile move, given that the city would have been the more likely party to feel aggrieved and file suit against EarthLink for failing to live up to the terms of their agreement.

EarthLink’s claims of offering the network to “a non-profit” or the city for free skirts the issue that EarthLink may have certain liabilities for electrical power and other fees that haven’t yet been paid; Wireless Philadelphia had agreed to pick up or defer certain charges as part of the deal that brought the network provider in. But without a completed network, and the contract therefore perhaps susceptible to being declared in default in court, it’s unlikely that this will play out nicely.

And I’ll say bluntly: If someone offered you $17m of outdated equipment on a network that never worked to specification that wasn’t completed, and that already had known high annual costs, and which a private firm gave up as a bad job that they couldn’t turn a dime on—would you take that deal? No. EarthLink will ultimately have to pay much more than $1m, I predict, and I suspect some of the settlement will leave gear in selected neighborhoods behind for more modest networking purposes. It’s not going to be as easy as releasing a press release, although I haven’t read the contract’s provisions for this set of circumstances, and I’m not a lawyer.

The failure in Philadelphia, and EarthLink’s exiting the entire muni-Fi business, represents the end of a bad model in which a company agreed to assume all risk and costs associated with building a public access network. When the assumptions were that networks would be cheaper and easier to build in 2005, and that citizens in many larger cities had few affordable broadband options, it made some sense to build a network on spec.

Three years into this, however, it’s clear that that capital investment is 2 to 3 times higher than what was anticipated to reach a level of service quality that people will expect; that, when presented with potential competition, DSL and cable operators will slash prices and offer cheap 1-year or “lifetime” rates with long-term contracts; and that wireless broadband delivered via Wi-Fi isn’t the best of ideas for indoor service.

Minneapolis may wind up being the only large city, if the network quality and subscriber rates play out, that has a public access network that works and produces a return.

Update: Wireless Philadelphia released a statement from its chief, Greg Goldman, that WP is still hoping to work out a resolution. They “remain optimistic.”