The leading cable companies in the U.S. are banding together to set up 50,000 Wi-Fi hot spots. Any subscriber to one cable providers will be able to use the others’ services; for example, a Time-Warner Cable customer in New York, where the service, called CableWi-Fi, is set to launch, will be able to log on to to a Cablevision hot spot. This is great news for consumer. Maybe.
Unfortunately, I heard about the cable companies’ plans while traveling to New York on Amtrak’s Acela. The Acela and other Amtrak Northwest Corridor boast free Wi-Fi, but it is generally so bad that I usually turn it off to force my iPhone and iPad to use cellular connections. Otherwise, they can spin forever just trying to check for mail or load tweets. Mobile devices are generally designed to use Wi-Fi when available, but they aren’t smart enough to figure out when the access point isn’t really connected to anything downstream and they just keep futilely trying unless you turn Wi-Fi off.
Amtrak has an excuse for its horrible Wi-Fi; trains rely on flaky, low-bandwidth wireless connections that many passengers will try to share. But Wi-Fi service almost as miserable is all to common. Free Wi-Fi connections are almost guaranteed to be awful, especially if the hot spot location is busy, and even paid service, such as absurdly overpriced hotel connections (the more expensive the hotel, the more you pay), if often nearly unusable.
The biggest problem is that it is to set up a hot spot, but not so cheap to run one. A consumer grade access point/router can cost well under $100 and even a carrier grade Cisco access point goes for under $1,000. The problem is that you need a way to connect your hot spot to the internet, and this connection, known in the trade as the backhaul, is where the cost starts mounting up. Even my terrible Actiontec wireless router, supplied by Verizon, regularly delivers 20 megabit per second downloads, but that’s because it’s connected to FiOS fiber and there are almost never more than two of us using it at once. Have 100 people sharing a 2 Mb DSL connection and you begin to see the problem.
Inadequate backhaul is probably the biggest reason for the disappointing performance of public
Wi-Fi. Paid sites, such as those amalgamated by Boingo, tend to do better, but even there you’ll find tremendous variance in performance. It doesn’t strike me as good business for vendors to promise free Wi-Fi and then deliver a terrible experience. Let’s hope the cable companies provide backhaul over their own speedy DOCSIS networks, but there is no guarantee they will, especially if the hot spot locations are not passed by cable service.
Another open question is how easy it will be to connect with these hot spots. The cable companies have made it clear that this is a service for paid subscribers, so they will need a way to authenticate logins. This can be a real pain for users, since the username and password information generally must be entered manually on a mobile device every time you connect (and often, reconnect, when your connection is mysteriously dropped.)
A new standard called Hotspot 2.0 (also known as PassPoint and IEEE 802.11c) could make these connections much easier, at least with devices that implement it in either hardware or software. These devices would authenticate themselves to a hot spot without any manual intervention by the users. The cable companies, however, have not indicated whether they will be using this technology.
If the cable companies do this right, it could greatly enhance their ability to deploy their existing services to their existing customers wirelessly while avoiding the difficulties imposed by wireless carriers’ data caps. But unless is it well thought out an executive, CableWi-Fi will just be another chapter in the long saga of public Wi-Fi disappointments.