Blog Post

Why Isn’t Wi-Fi Better?

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

We talk a lot about Wi-Fi offload as the next big thing for mobile operators, but a recent conversation with an AT&T executive put that into doubt. The executive noted that AT&T (s t) didn’t see Wi-Fi helping the nation’s No. 2 carrier offset congestion because in most cases people don’t use Wi-Fi unless they are sitting still in a hot-spot. And apparently, there are plenty of people still wandering around watching YouTube (s goog) videos.

Sure this person was selling the AT&T and T-Mobile deal based on AT&T’s need for spectrum, but the statements also ran counter to what AT&T itself has said back when it launched the iPad. At the time, an executive with AT&T said that Ma Bell expected a “substantial” percent of usage to occur over Wi-Fi networks. I couldn’t get details on actual usage in the last few months from the executive I met recently, although AT&T loves to talk about Wi-Fi connections in laudatory press releases. But it started me thinking about the perception of Wi-Fi and the reality. How can we make the reality match the perception of Wi-Fi as this fabulous place for offloading mobile traffic?

Making Wi-Fi Mobile.

One of the problems with Wi-Fi is that it’s for fixed mobile access, meaning a person is sitting still. But it doesn’t have to be this way as several startups such as FON or Anyfi have shown. It is possible to create a Wi-Fi network that allows a user to remain on the Wi-Fi network, even when moving. Other good work on this comes from chip firms where engineers are discussing standards to enable seamless hopping from hot spot to hot spot or perhaps even a handoff between a cellular and Wi-Fi network. The bottom line, is that for Wi-Fi to be mobile, the device must hop from network to network and authenticate without the user being aware of it. The Wi-Fi Alliance is working on a standard and authentication with operators to make this happen, with a suggested standard to surface later this year and an actual standard some time in the first part of next year.

Look at the Cable Companies,

The next step after a way to hop seamlessly from network to network is to actually get hotspot owners to let that happen. The standards mentioned above will help, but folks should take a page from the cable guys that already do allow roaming on their networks. In New York City, the three cable providers share a Wi-Fi network, essentially turning the city into a giant hot spot for their customers. Plus, this week Deutsche Telekom signed a deal with iPass to create a carrier Wi-Fi market that essentially wholesale Wi-Fi access for carriers and enables them to allow their end-users to roam onto this combined network. Another option are services such as Boingo that folks can subscribe to and that offer apps to make hopping fairly consumer-friendly.

Muni Wi-Fi 2.0

Once Wi-Fi is everywhere, seamless and one can roam and can be mobile using Wi-Fi,the next step is to put it in more places. The AT&T executive pointed out that someone enjoying lunch in a public park likely can’t find a Wi-Fi network, so will surf on the cellular network. But more and more cities, shopping malls and other developments are adding Wi-Fi access as a lure to get consumers to hang around. Some of these take the form of public-private partnerships where a business development district might implement a network in partnership with a corporation located nearby, and some are just public networks. In three or four years, if the Wi-Fi Alliance, Intel or another large player can set up the infrastructure to make seamless roaming and authentication happen then having those networks in more places helps solve the final piece of the puzzle — finding networks everywhere.

There are plenty of other startups and efforts aimed at making Wi-Fi more usable and better, but it will likely never replace cellular when it comes to convenience. But I think these would be a start. What else might help?

Image courtesy Flickr user Adventures in Librarianship.

26 Responses to “Why Isn’t Wi-Fi Better?”

  1. It does seem possible to create wi-fi mesh networks to cover entire regions or areas at a fraction of the cost of adding additional capacity to a traditional mobile network. In my experience coverage of a square km can be achieved with between 15 to 25 nodes, depending on terrain. It seems we are just waiting for one of the large operators to take the bull by the horns and deploy a nation wide overlay of built up areas, in fact i believe trials and small deployments are in operation.

    A good example of city wide coverage would be Google’s wi-fi network in San Francisco (Mountain View), its secure (as Wi-Fi can be) allows devices to roam from hot spot to hot spot with no or very little handover delay. Coverage Map –

    The same technology can also be used for smart metering and supporting a mobile worker type network. The sooner a large operator realises a wi-fi network could support more than just one business case, the sooner we will see a mass rollout.

  2. Gary LaPointe

    Lots of roaming spots have awful doped and reliability. Most ATT McDonald’s wifi is very slow (in the dozens I’ve tried in Michigan) compared to 3G. And 10% of the tine I can’t connect.

    Many times I’ll just turn off the wifi for a better experience.

  3. Mikael

    Why is it better to off-load traffic from the mobile network to WiFi in the first place? Because ATT and the carriers don’t have to pay for back haul. The radio equipment would be equally expensive if mobility was added to WiFi. And the Spectrum licenses purpose is to remove interference, which could become a huge problem if more people started using WiFi.

    WiFi in Starbucks, hotels etc are using back haul that is also carrying startbucks own data communication, credit card verifications, cash register transactions etc. So the back haul for the coffee customers are “Free”. With increased customer WiFi traffic, someone would have to start paying for that back haul, and the advantages to cellular goes away again.

    I think the carriers realize that some off load to WiFi is good, but that the bulk of the traffic must be carried on the licensed network, fully within the control of the individual carrier.

  4. TCamilli

    Two things missing from this argument: Cost and use case.

    From a cost standpoint, too few companies provide users with an economic incentive to use WiFi. If the use case is based on offloading smartphone data, I’ve already paid for my mobile data. Why would I pay again to use WiFi? In the US, I didn’t appreciate how spoiled I was with Starbucks, Panera, etc. offering free WiFi. Here in Australia isn’t not uncommon to see WiFi hotspots that cost 3-4x the cost of mobile data on a $/GB basis. An extreme example: The last hotel I stayed in charged $60/day for WiFi access.

    The primary use case challenges have already been touched on with the user experience comments. If I’m standing in line at Starbucks and want to check email or Twitter, it’s simply faster to use my mobile data instead of an arduous login process (which almost always requires agreeing to terms of use statement that I’m pretty sure have taken both of my children from me). On the other hand, if I’m hunkering down for an afternoon of working from a cafe, it makes more sense to go ahead and login to WiFi.

  5. Stacey nailed it. Most of the problems with Wi-Fi aren’t really related to Wi-Fi. They are related to interworking 802.11 with with backend authentication and policy systems and inter-network roaming technology and agreements. This is all being sorted out. May not happen fast, but there’s not technical obstacles in the way. There’s also no one single solution. Good things just take time. And often times these things are more political than technical. But carriers ARE interested in Wi-Fi….big time. None of them expect it to operate like the licensed spectrum but they all see it as a good complement to what they have today to add capacity and coverage quickly (and cheap). Vodafone, Virgin, O2, KDDI, BT, the Cloud, DT, SK Telecom – almost every tier one has a strategic plan to better integrate Wi-Fi traffic with their evolving packet core cell networks or using Wi-Fi to backhaul small cell traffic, or mobile data wholesaling or simply better, faster public access. What’s going to actually make Wi-Fi more reliable is someone touching the RF spectrum (the air). Most Wi-Fi suppliers do NOTHING to address the “layer 0” aspects of Wi-Fi. They just give you fancy and complex knobs to turn once users get connected (if they can). To make it better we need to continue to develop technologies that control signal path integrity and Wi-Fi transmissions over the unlicensed spectrum.

    • bobfrankston

      How do you do something as simple as a communicating pacemaker when accounting is level zero in the ISO stack?

      Do you authenticate yourself to sidewalks when you take a stroll? Why do you need to authenticate just to communicate? ( for more on authenticated sidewalks)

  6. One possible hurdle to this kind of technology is the recent trend towards making wifi network operators liable for any traffic passing over that network. Well done anti-piracy legislation…

  7. We need to be careful about the terminology we use. This is not a problem with the technology of Wi-Fi. It is entirely a problem with trying to maintain a business model that no longer makes sense. In the 1920’s we transitioned from private pikes to public roads. We shifted to infrastructure funding because we recognized that the value is external. This is even more the case with exchanging bits where value is 100% external because ones and zeroes have no value in themselves. We don’t even consume bits – we merely exchange them.

    As Skype has shown there is no technical problem with mobility – we just need to use protocols based on relationships between users or end-points. In fact the carriers’ own UMA is another example of the ease of mobility. Yet we accept the idea that two phones using UMA must generate billable events even when there are no carrier facilities are being used. That’s the real problem – a funding model that requires limiting our ability to communicate.

    Why must a connected pacemaker deal with the complexities let alone the cost of generating billable events? And why must it fail to connect as the default?

    Exchanging bits is simple and we can take advantage of any available path at essentially no marginal cost path as I explain in

    Public roads and sidewalks are very expensive yet we make them available without complex authentication/billing schemes. The wires and radios we use to exchange bits can cost one millionth what roads do but we spend a lot of money to erect barriers which threaten our lives and harm the economy.

    The question isn’t “Why Isn’t Wi-Fi Better?” The question is why are we trying to so hard to limit our ability to communicate?

    For more –

  8. After living abroad for the last two years (mainly europe) I think we’re a little spoiled in the US with wifi availability and where and when we expect it. In most major cities in Europe you have to pay for this – at starbucks, hotels, trainstations, airports, etc., and it ain’t cheap. I’ve paid up to $6.50/hr to use wifi at a number of public and private places…

  9. All this talking is vapourware, and you know it:
    “The Wi-Fi Alliance is working on a standard and authentication with operators to make this happen, with a suggested standard to surface later this year and an actual standard some time in the first part of next year.”

  10. The user experience sucks because the people who roll out the networks don’t take into account who is using WiFi and how they are using it. Example: most city WiFi networks and WiFi in subways and train stations require you to log in all the time with a cumbersome username and password. In reality, most people want to quickly get on the WiFi network, check email and then go. The WiFi network in SF’s Muni stations suffers from this stupidity. By the time you are done entering these details, the train has arrived. By now most people logging onto outdoor WiFi and public transport WiFi networks use portable mobile devices like iPhones (to a lesser extent iPads), but not laptops. Yet, when you look at how the network interface is designed, it’s as if a bunch of people are sitting there using the WiFi network with a laptop or desktop. Another pet peeve: having to log on again and again when moving around in an enterprise or in an airport. Why can’t the network remember you? Heathrow Terminal 5 actually has a WiFi network (BT OpenZone) that remembers you as you move around. But this is the exception to the rule.

  11. Thanks for the mention Stacey. :)

    I think Bob is spot on; it’s the UX that needs to be improved. Wi-Fi works really well when you’re at home (and have WPA enabled) but in hotspots it just plain sucks. The fact that hotspots are used at all I think is testament to Wi-Fi’s strengths in other departments.

    You need AT LEAST this if you want a good mobile Wi-Fi UX:

    1. Automatic authentication. Enter credentials once, use everywhere. Extra credits for reusing credentials already in the device, e.g. SIM card.

    2. Encryption, and preferably with a trusted end-point. The Guardian recently illustrated the trusted end-point part of it (in the sense that over-the-air encryption would not have helped):

    3. Quality coverage. Blanket coverage is great but not if it’s 80% crap that makes you want to turn your Wi-Fi off. Quality needs to be actively monitored and managed so that devices only connect (i.e. switch over from 3G) when Wi-Fi signal strength is good enough for a quality UX.

    We’ll hopefully have the first major ISP up and running with all of the above (and some more) this fall.

    • Lucian Armasu

      I completely agree with all your points. For Wi-Fi to eventually replace GSM/LTE it needs to be hassle-free for the user, connect automatically and have a longer range, but also be more battery efficient. Right now it drains the battery pretty fast, though LTE may be even worse right now.

    • Boris B

      Yep – that’s about right with a few others that are even more important:

      a) You need a “distributed” network approach, because achieving adequate coverage commercially by one vendor is near impossible. FON is closest to doing it right, but they require you to buy a box, which is a non-starter.

      1a) so consumer WiFi companies have to be incented to build in support for third party access through a central system.
      1b) consumers need to be assured it is 100% safe and get a strong incentive – better coverage everywhere else could be that.
      1c) a party needs to take on role of a central system – maybe iPass as Stacey mentions.

      2) forget about handoff in poor quality signal situations. Won’t work (too hard to measure what’s sufficient and not and how quickly to hand off). Need the new standard being pushed by QUALCOMM, Atilla, where the hand-off is implicit – a user is connected and using WiFi and 3G in parallel so as one gets better it just supplements and as one gets worse it just drops off.

      3) everything Bjorn says.

  12. Interestingly when WiFi was first specified and developed, it was supposed to be the next thing to replace GSM/UTMS (etc) cell networks. However, obvious issues about range, security and property ownership complications ended up resigning the technology to a home networking one.

  13. If wifi is a bonus offered by a business to attract costomers. (Starbucks for a good example) I don’t think they will be too quick to share their return for this service. Blanket coverage is, to me the only marketable option. Then why would you need cellular?

  14. IF the carriers didn’t charge customers for calls made over wi-fi. more people would enable wi-fi on their phones. The cost of operating a gateway from the internet to the PSTN is near zero for carriers, so there is no reason why making mobile calls over wi-fi on your phone should use plan minutes. I realize it’s hard for these execs to understand people don’t want to sign over their paychecks to them, but maybe one day they will realize that they are not loved by their customers, just like Kadaffi is not loved by Libyans.

    For data purposes, people don’t use wi-fi as much as they should because (I think) of the security reasons. You need to be very uninformed about computer security to access any site that requires a log-in over a public wi-fi network. I’ll turn on wi-fi on my phones and tablets at home or in my office, but the networks are secure there. Everywhere else I use my 3G/4G connection.

    I don’t know if ATT expects people to use wi-fi while traveling in cars; even with good network hopping technology, it’s not going to matter if you are traveling on a highway.

    ATT seems to want to be able to charge people as much as they can, while providing as little service as possible. Like, they want to charge $60+/month for 3G/4g service, put caps on it, and then encourage customers to use wi-fi (that other people provide, who might have to pay ATT for the local network’s internet connection) as much as possible. ATT doesn’t seem to realize that if they push people to use wi-fi by making their 4G service as useless as possible, they will provide incentive for wi-fi networks to evolve, and ultimately eat away at their business. Not that any ATT exec really cares about the long term sustainability of their company.