Muni Wi-Fi Redux
Muni Wi-Fi, or Metro Wi-Fi has been riding the hype cycle for a while, and has recently fallen into somewhat of a backlash, for which I’d like to take some credit. Not credit for being a reactionary naysayer, but for at least trying to force some reality into notions of using a Local Area Network technology to cover a whole town. This post is triggered by a piece by Craig Mathias over at Computerworld. That article is singing the praises of Metro-scale Wi-Fi networks, powered by multiple radios. Mathias seems to think that a new multi-radio approach will cure the maladies that have plagued many networks thus far, but this isn’t the case. And this cure isn’t new: I met with Bel-Air Networks’ VP Networking Phil Belanger in 2004, and believe me they were pushing multi-radio solutions all along vs. the Tropos single-radio approach. Problem is more radios/node drives up the cost of the network. Funny, this reminds me of the last time I publicly disagreed with Mr. Mathias. It was 2004, when the future of Bluetooth was unclear to some, and I said, “there is not a credible market-ready solution that truly threatens Bluetooth.” and Mathias said, Bluetooth “is largely a vestigial technology now that 802.11 is broadly accepted.” That’s right. He said Bluetooth would be replaced by…Wi-Fi. Wrong about Wi-Fi then, wrong now.
I’ve decided to whip off a laundry list of thoughts on Muni Wi-Fi, but to spare those not interested, you need to click ‘Read More’ to see them.
TDMA stands for "Time Division Multiple Access" and it is what GSM uses to share a frequency among many users. The frequency is cut up into time slices, and each user is allocated a slice where their compressed data is transmitted. Similarly, I think of Wi-Fi as WDMA, or "Wall Division Multiple Access". In WDMA, unlicensed spectrum is used, and the notion is that every home or office building can share the same frequency, because a wall divides them, attenuates Wi-Fi signals, and reduces interference. When you think about it, that's what the FCC hopes for when they set the RF transmit limits for unlicensed bands - the goal is to make sure one person's use does not interfere with another's. And Wi-Fi fits WDMA so nicely, because it has very poor wall-penetration characteristics. Taking the walls out of WDMA is like taking the time-codes out of TDMA or the guard-bands out of FDMA (frequency): there's going to be interference.
- It's Not Already In Your Laptop
Mathias and others often allude to the fact that Muni Networks can use "off the shelf" Wi-Fi components, and that people already have Centrino chips in their laptops. But the reality of most live Muni Wi-Fi networks is that most people will need additional antennae or repeaters mounted on the exterior walls of their homes if they want to receive the signal.
-The Elusiveness of Ubiquity
There is a huge difference between having many hotspots all over a city, and having ubiquitous coverage over an entire city, including in-building. Most muni projects offer mostly the former. Many of the early deployment towns were disappointed at the abundant coverage holes, and were forced to increase their infrastructure expectations and costs, but still do not have ubiquitous coverage. But here's the thing, ubiquity means just that. Either you have it or you don't. If you sell ubiquity, and don't offer it, you will have angry customers, support calls, and churn. If you don't offer ubiquity, you have a lower value proposition to the mobile users, who may thus prefer cellular data.
- Municipal Government Use
The best ideas I've ever heard for Muni Wi-Fi, and the only cases which I believe have been successful, are those that are using their Muni Wi-Fi networks in order to lower costs, and increase efficiency of operations for city staff. Building inspectors, meter reading, cameras, police access, etc. are all cases where staff can save time by not having to return to base to enter/access data. But also, employees are people that can take advantage of hotspot access, and don't necessarily need ubiquity. You can teach your staff where your ~12 access points are throughout the town, and send them there. By removing the need for ubiquity, and removing public access, a town can greatly reduce costs while reducing complaints and customer service hassles, too.
- Public Access
As soon as public access is added, expectations are higher. Full-coverage is expected, and the service will automatically be compared to DSL for home users, and EV-DO for mobile users. How does Muni Wi-Fi stack up? Against DSL, poorly, and DSL providers will simply lower their prices to compete (which is still a good thing for consumers, but not for the Wi-Fi provider). Against EV-DO there is a huge price advantage for Wi-Fi, but without ubiquitous coverage, EV-DO still wins. Few mobile professionals are willing to sacrifice the reliability and coverage of cellular data for the religion of Muni Wi-Fi. Cities like Philly are learning that their free Muni Wi-Fi will cost $20/mo, reducing attractiveness. Even though it drives the hype, Public access is a bad reason to deploy Muni Wi-Fi, but may work as an additional service for covered homes once the business plan is amortized by govt use.
- Enterprise Use
Are enterprises likely to sign up for a Muni Wi-Fi service as their main network provider? Forget about it. The reliability of Wi-Fi just isn't there. On the 10 laptops that I have managed as "IT guy" and the dozen or so APs that I've owned, signal drop-outs are standard fare, even though the clients are yards away from the AP. This isn't enterprise-grade stuff. It's OK for laptopping in the conference room or Starbucks, but don't get carried away.
- HotZones<br> This is one of my favorite uses of Wi-Fi. When you are discussing a finite area, like "200 yards of Main St.", "City Library", or "Grant Park" then you are talking about a manageable area, with fewer unexpected obstacles, and a situation that can effectively be covered with a Wi-Fi signal. Wi-Fi doesn't scale well, but in this scenario we're not asking it to. I'm of the "too cheap to bill" school, so this type of deployment subsidized with a splash screen, advertising, a sponsor, a Chamber of Commerce, or taxes (ugh) is my preferred case of Muni Wi-Fi. You can attract a few people to a city center, or at least keep them there longer with such a network. Some merchants could use it for credit card POS machines during sidewalk sales or markets.
Wi-Fi still has a role in hotspots, but not so much in a muni sense. Hotspots, for the most part, should offer free connections as an ancillary benefit of using their core product (coffee, ice rink, auto repair...) Some locations have a trapped audience, and can extract a premium for access. If you are a vendor comfortable gouging customers, you can at: airports, hotels, conference centers, airline lounges, airplanes, trains, and such. Beware the backlash, customers who use Wi-Fi also are exactly the customers who choose their vendors based on Wi-Fi. Thanks to pricey T-Mobile Wi-Fi pricing, Starbucks is losing coffee customers.
- The Anti-Carrier Becomes a Carrier
If a municipality does offer public access Wi-Fi, as the service grows and scales, the Wi-Fi technology has trouble scaling with it. But on other levels, the operation starts to look a lot like a telco. Billing, operations, customer service, specialized CPE, truck rolls, redundancy, 24hr maintenance, and the core network all start to look like a telco. Just because Wi-Fi is simple and cool in your office doesn't mean it scales to city-size the same way. This point is one of few in which I agree with Mathias, who says in a white paper, "Operators of metro-scale Wi-Fi networks are increasingly recognizing that such large-scale deployments need to be carrier-class. There's really no difference between operating a metroscale Wi-Fi service and any other broadband infrastructure. This means that it's particularly important to have a carrier-grade operational support system (OSS) to handle provisioning, billing, and customer service." Congratulations, you've replaced the telco with a telco. Way to fight the Man.
- Running Interference
Interference is a huge problem for Muni Wi-Fi networks, and although there are mitigation strategies in the works, I'm always baffled that the first generation of Muni Wi-Fi solutions didn't address this. Now the same Polyannas that told us the first generation would work are telling us the second will. Oh, but now they need multiple radios...oh, and CPE. Wasn't the appeal in the first place that this would be sooo simple? Anyways, back to interference: despite mitigation strategies, the amount of interference is only going to grow. More and more radio devices are being sold on the 2.4 and 5.8 bands. And Muni networks themselves are the worst purveyors of interference, broadcasting at the maximum FCC allowable signal strength from high on lamp-posts. "Sir, there's good news and bad news: the good news is you're one of the people in range of Muni Wi-Fi; the bad news is your home Wi-Fi network doesn't work anymore."
I've been quoted by some astroturfing think-tanks in the pockets of telcos, and yes, I work for telcos all the time. But that doesn't mean that I, or Techdirt, have any political stand in the Muni Broadband debate. If anything, we just love technology, and broadband is good no matter how it comes. If a town and its citizen want to provide broadband through government, knock yourselves out. We just think that it's likely that there are better options than current generations of Wi-Fi.
You, the reader, deserve good predictions that prove to be correct a few years later, so here some are: there will be live, operational Muni broadband networks in 2010. Why? Because people won't listen to me. There are many more voices on the other side of the debate - some idealists who don't understand the technology, and some others biased by their personal interests, and some who just have a different perspective. Many Munis will be using some form of Wi-Fi. But Muni Wi-Fi will not replace the existing commercial wireless solutions, nor will it command the majority of the market. Many municipalities will find the costs over-budget, the results were less than expected, performance and coverage never met promises, and the fixes of the vendors who made the promises included buying more equipment from them, and upgrades too. But with sunk cost economics, it makes sense to continue. In 2010, many of the municipalities who were considering Muni WiFi in 2007 will have chosen NOT to launch the public access networks based on the experiences of other towns. Many towns will shift their plans down to Hotzones for public/private access, and hotspots for government operations, and see success. By 2010, cellular data like EV-DO, HSDPA, LTE (maybe 2012), and (dare I say it) WiMAX will remain much better for mobile/nomadic use than Wi-Fi. Cellular data prices will come down from 2007's low of $60/mo to tiered service levels starting at $20-30 for a capped throughput or speed (Like 5GB/mo for $30). Wide-Area Wi-Fi will be too little, too late.