Sometimes the chirpy cheerfulness of the Wi-Fi vendors (even when they are laying people off ) gets us yearning for a bit of pessimism. The up-beat mode can be a bit wearing and a bit transparent. Vendors need Wi-Fi services to work to save their companies and service providers want them to perk up comms spending.
Enter the Wi-Fi prophet, a Jeremiah for the unconnected age, a strategist for a medium-scale telco who wants to remind us of the downside of Wi-Fi.
Steve Kennedy, head of product futures at Thus, is not against Wi-Fi itself but, like an Old Testament prophet, is always ready to warn us of troubles ahead - if we don’t listen and repent our sins. Sure, Wi-Fi may be useful, it may be working fine and there may be plenty of untapped demand, he says, but just wait! Wind him up and you get an hour’s closely-argued sermon on the dangers.
Thus, previously known as Scottish Telecom, is happy to use wireless when appropriate but has no plans for hotspots. Its fixed wireless access in the Highlands and Islands of Scotland is wireless-based, because “digging fibre into granite is not really a very useful thing to do,” says Kennedy. “But we are an infrastructure provider and the hotspot model hasn't been proven.”
On the business model, he is dismissive: “To make hotspots profitable, you have to sell an awful lot of Wi-Fi access, or coffee,” he adds. “Are people really going to choose their coffee shop on whether it has Wi-Fi or not?” Kennedy reckons it takes 10 to 20 users a day, each paying for an hour’s usage, to make a hotspot break even.
Even though it is now quite cheap to get a “hotspot in a box”, people should be wary of running them and expecting to make money. It would break the terms and conditions to use one on a residential DSL line and stretch the business model to use a commercial one. If you are already putting in a network for another purpose, you can use spare bandwidth for Wi-Fi backhaul (as Cloud is doing with networked gaming machines in pubs). “But in those cases, the infrastructure is already partially paid for and infrastructure is the major cost,” says Kennedy.
What’s the hurry?
It’s the hype he objects to: “Wi-Fi is like the dotcom boom of a few years ago,” he says. “People are seeing it as a panacea technology.” Promotional campaigns like Intel’s Centrino are really aimed at selling more mobile computers and Intel processors, he says: “People like Intel have a vested interest.”
Difficulties and costs are being overlooked, he says. For instance, billing and audit trails are an issue even for people running a free access point. “Billing is not just about getting money out of people,” he says. “If someone goes into a hotspot and uses it illegally, you need to prove that [even though that traffic came from your IP address] it wasn't you.”
“When they come with a seizure warrant, you may not care,” he says, in full-on Jeremiah mode. “You may be running your business on that computer and happy to have it sitting in a warehouse for six months. You may be fine about that.” There are cases, he says, where this has happened to people on networks that share broadband access.
Even a free hotspot – or a company that offers visitors access to the Internet over its own network – should require users to agree to terms and conditions before offering them a conduit to the Internet. “If you offer Internet service through an ISP connection, it is your legal obligation to pass on the ISP’s terms and conditions - assuming the ISP lets you share the bandwidth in the first place.”
You should also get MAC addresses of client systems, to provide some sort of audit trail, he says. If the hotspot is deemed to be a service provider, it will also have to comply with the RIP provisions and be prepared to keep logs and allow authorities to intercept traffic.
Of course, service providers do all these things more or less automatically. Kennedy is saying that in some ways there may actually be more costs to managing a free hotspot than a commercial one - since someone not geared to billing and logging will have to start doing those things. In practical terms, this means that home users should not share their bandwidth and coffee shops should factor in the cost of compliance before deciding if it is worth it.
Another cost is support: “If you are a coffee shop running a hotspot, and someone has problems connecting, do you offer them support? A trained member of staff is an increased cost, if you are trying to do this as a way of selling more coffee.”
On the issue of regulations, he points out that the 2.4GHz and 5GHz bands are not “unlicensed”, they are “licence exempt”. To him, this is a big difference. “The Radiocommunications Agency publishes a blanket licence. If you stick to that, you don't need a specific licence and in fact the agency will never issue a licence”
In the licence exempt bands, people have to operate under the Telecomunications Act, not the stricter Wireless Telegraphy Act, but they do have to comply with strict power emission rules, says Kennedy. They will also have to follow rules for service providers if they are deemed to be one. “Is a coffee shop with a free hotspot a service provider?” he asks. “Even if you are not charging, but you add value to your business, you could be deemed to be a service provider.”
Telecoms regulations are heavy-weight stuff, he argues, and even the lighter approach of the Communications Act place burdens on people. “One of the penalties of having to pay a price for a telecoms licence was that telcos had to maintain regulatory departments,” he says. “Now there are guidelines, and anyone can operate as a communications service provider, You are OK as long as you stick to the provisions of the communications act, but you still have to understand what you are doing to comply with the act.”
“It all becomes very nasty,” he says. “You have to keep billing records for seven years, but if you keep audit trails for too long you get into violation of the Data Protection Act. How many coffee shop hotspot providers have regulatory departments? " he asks. "They cost money.”
In the 2.4GHz band, devices like microwave ovens can leak a lot of power, even though they are not actually designed to emit any. “Signal to noise ratios on wireless LANs decrease dramatically between 6pm and 7.30 when people are cooking dinner,” he says. “A microwave kicks out 850W to cook food. If it contains 849.9W of that, then there’s 100mW coming out which will splat all over your Wi-Fi.”
He also makes much of interference with Bluetooth, even though this is usually fixed very easily by moving the offending Bluetooth device, since that has a range of a few metres. “That’s OK now,” he asks, “ but how many devices are going to have Bluetooth in them?” When virtually every mobile phone has it, Wi-Fi networks will have holes punched all over them.
The problem here is that service providers can’t guarantee a service to users: “You can't guarantee anything will work in that area,” he says. “Wi-Fi is not a bad tech for wireless LANs, it's just that there are things you have to think about before going into it headlong.”
Other issues he brings up are the well-understood “hidden node” problem, whereby problems accessing the air arise because two devices on either side of an access point both attempt to talk and neither backs off because they cannot hear each other. “The standard just says, don't have hidden nodes,” he warns.
The 5GHz band, where 802.11a operates, has its own problems since there are sub-bands some of which are in use by military radar. A proposed scheme of so called “light licensing” would require users to register their 802.11a at a website, and request permission to use it in a specific area - permission that may be denied if a military base is nearby. “They are still arguing over the licence for 5GHz,” said Kennedy.
“You may say these are just details, but that is the point,” said Kennedy. “It is the details behind this that make it complicated, and people don't take them into consideration.”
And then there is the big bugbear, security. “Everyone knows that WEP is breakable,” he said. “If you are running a hotspot you don't care.” The WPA replacement is a good idea, but won’t interoperate with WEP, he says. “It's one or the other. Currently 90 percent of people use. WEP-capable WLANs. It will change to WPA over the years but WPA is not there now.”
And the argument of using VPNs to encrypt data only applies for users linking to a corporate network. “If you are on the Internet talking on a peer level, where do you VPN to? If I access my ISP email through a hotspot I'm sending that data cleartext.”
So what’s the problem?
In the end though, Kennedy is preaching a paradox. None of the things he brings up are unknown. Take it all on board, and he is just arguing two things, that we knew already. Firstly, that the Wi-Fi services market should go to mobile operators, set up to handle the regulations, bundling service alongside other offerings. Secondly, that anyone else, such as businesses offering visitor LANs, should watch their regulatory backs. We knew that already.
Why is this going to be such a big problem, then? Simply because there will be such a lot of Wi-Fi around. “If it really takes off, then interference will be an issue,” he says. “If every office has one, and offers guest access to visitors, then these things ripple into each other.”
In other words, if Wi-Fi is a massive success, then we’ll have to do lots of work to make it operate better. Well yes, we knew that too. It’s like the argument (circa 1992) that the PC would die because it didn’t have good enough graphics. Sell enough of something and you have a critical mass to fix those things; if you don’t sell enough you don’t need to, because it isn’t a problem.