Faultline has said many times that it believes that cellular operators are too busy and under too much price pressure to spend time building out WiFi networks despite the best intentions of base station and femtocell equipment providers that are encouraging them to go down a HetNet route which includes WiFi.
We think, that certainly initially, that there will be a scavenge for WiFi capacity which is already in place, highlighting Homespots, which are available in the millions, as even more important than established outdoor Hotspots.
A Homespot is where domestic WiFi is adapted with a software download, to be outward looking with a second, commercial SSID.
But so far we have given over little time considering how this model comes into being, whether this is a series of roaming agreements, with each one requiring a separate piece of software, or if it will be a single WiFi trading exchange. After talking to BandwidthX this week we are now tending to believe that it will go the exchange route. Whether or not there will be a single such exchange is another matter.
Back in June the first WiFi trading hub called BandwidthX came out of stealth and chose the WiFi Global Congress to launch. Most reporters were taken aback and as usual wrote up a piece that quoted the press release and moved on, but our take on this is that it has launched at a pivotal point in time, and that its lead to market, while not unassailable, could see it rise in short order to the pre-eminent player in WiFi offload roaming.
While we expect rivals to emerge, we think that by solving the most critical problems in the right order and to the correct depth, BandwidthX will trigger mass uptake as a trading proposition, and change the WiFi offload landscape overnight.
The company plans to operate a cloud based Bandwidth Market, which enables Mobile Network Operators to buy WiFi capacity when they want and where they want from participating ISPs and WiFi service providers who have capacity in their networks.
The idea is that Bandwidth Market establishes an equitable clearing price for each Wi-Fi capacity transaction, separately taking into account bid and offer pricing in an automatic negotiation. No changes are needed in existing WiFi access points, which throws a curved ball at Hotspot 2.0 and other protocols, and the mobile networks don’t have to make any changes either. Only a small Bandwidth Market application needs to be added to the mobile devices to participate in the trading and so far it has written one version of this app for Android.
We spoke late last week to CEO Pertti Visuri, who was probably previously best known for founding AirGain, the embedded antenna specialist, which as far as we know remains private and successful. You can see the Airgain influence in how an understanding of onboard smartphone antennas is needed to make this system work.
What Visuri told us made us rethink our view on WiFi considerably. We previously thought it would make or break operators which already have large swathes of WiFi capability and create something of a problem for cellular operators who are without any WiFi capacity. We still agree with that, but further we reckon this is going to totally change the overall Capex Opex conversation at traditional cellular operators. This is brought about because sudden overnight access to existing homespots will make it economically improbable for cellular operators to create rival networks, and could permanently make ISPs and cellular operators interdependent.
Beyond that, cellular operators may eventually move more and more services onto WiFi, as unlikely as it is to offer voice in that manner, we have to remember lessons from other cellular operators that have already got some experience of working a joint cellular/WiFi operation, because they were lucky enough to have access to both.
Free in France built out 2,700 base stations, and to cover the rest of the country used either France Telecom’s cellular network via roaming or its 5 million Homespots. Most of the data traffic went over the Homespots.
China Mobile went out and built 3.8 million Hotspots to give it cellular offload, but it’s going to be tough to see any other operator going ahead and doing this, once they realize they can have that type of capacity and more, for a small amount of continuing Opex.
Visuri told us, ‘About ten years ago we did some work (at AirGain) improving the Freebox gateway for Free in France. At the time it didn’t occur to us, but now we know why it needed such a good WiFi performance. It was always planning this.’
But if a cellular operator can gain access to WiFi for a lot less than the lost opportunity cost of tying up tons of cash, then its desire to build out network and eat Capex will go down. Cellular operators under financial pressure will either slow down their equipment spend, or approach WiFi in a far more gradual and strategic way ‘ for instance building out networks in areas where there is either the greatest need or the capacity cost of buying it (on a market like this) are the highest. On the other hand, cellular operators will once again look at ISPs offering them WiFi offload in this manner and buying them outright, to reduce capex costs and inherit ready-made networks which are rising in value.
Visuri explained, ‘In WiFi several devices cohabit the same spectrum in a smooth and reasonable way. If you walk downtown in any town, anywhere there are 20 or 30 WiFi SSIDS, so it makes sense to use them.’ Not only are Homespots under used when at their largest capacity, but capacities are going up and when homes are empty they are totally unused.
‘Ten years ago you could bet that some of these SSIDs were open, but today security options ship switched on, so they are almost always closed,’ he said
‘What we have done is build-in negotiation and authentication and accounting into a single application without requiring anything from either of the networks.
What we have launched is an App for Android which cooperates with our cloud based service.’
He then walks us through the security options, which Visuri feels had to be strong, with all communications authenticated using the International Mobile Subscriber Identity (IMSI) plus a shared secret.
The app makes devices scan for access points and then share the results of the scan with the cloud based marketplace, which tells it which SSIDS it can see are participating in the exchange and what its asking price for bandwidth is.
But the beauty is the understanding that this system can be used for working with operators that you are friendly with and want to work with, and also with players that are highly competitive. This is because not only can prices vary by time of day or by where you are or what type of device is attached to the WiFi network, but also based on which MNO those devices are owned by or are connected to.
Both types of operators (WiFi and Cellular) set a price profile, and as Visuri points out, ‘not all Gigabytes are created equal,’ meaning that those available in the night time are virtually worthless, because cellular networks are less congested then.
So once a phone has offers of WiFi there is a discussion in the cloud server, where a cellular operator might effectively be saying ‘I am prepared to pay $15 per GB of data when by network is busy at this location,’ and the marketplace will match the offers from buyers and bidders and determine a clearing price
‘We have Alvin Roth, who won a Nobel Prize in Economics, on our advisory board,’ says Visuri. Roth jointly won his Nobel Prize for theories in market design and is known to believe in the real world application of his theories.
‘The system always returns prices which are lower than the asking price, and higher than the offered price,’ said Visuri, ‘and then when the deal is made, we use the existing cellular connection to pass on credentials from the cloud. If it needs a password, User ID or if the system has EAP SIM or Hotspot 2.0, this then provides the credentials to the App in the device. That’s the App, not the operating system.’
Visuri explains that the connection uses SSL, and the system internally uses obfuscation and unit authentication to prevent the shared secret being leaked to the operating system. ‘This would only work on this device anyway,’ he added, ‘because part of the secret is extracted from the characteristics of the device itself, so it would not work on another device.’
The accounting is done in the App itself, because it already monitors how many megabytes are used from the network and it now knows the price which was agreed in the trading negotiation. The whole transaction is reported using the existing cellular Radius Protocol.
‘We have designed this so that it is able to plug into existing operator systems, like any connected exchange,’ he adds.
Visuri is not willing to say precisely how much his company charges for this service, but it seems that like any market maker, BandwidthX’s charges are almost invisible. If every bid is only accepted if it is between the asking and offered prices there is a margin between what is bid and what needs to be paid, and BandwidthX can pick up this difference in the same way that a market maker in a stock exchange does this, by buying at one price and selling at another, slightly different price.
We immediately talked about how it was like the power markets that sprang up around Europe for trading electrical and gas power during the 1990s when utilities such as gas and electricity boards were sold off by local governments and they wanted to spur local competition. ‘We have an electricity power market that is very like this in California, but the transactions are carried out once a day,’ Visuri agreed.
But it is then that we begin to see more far-reaching implications for this. ‘It has a very good effect on churn reduction for the cellular operator’ Visuri says, ‘because data is not blocked and doesn’t fail to get across a congested cellular network.’ And of course the cellular operator is still charging for all data used whether it goes over the cellular network or the WiFi network. Cellular operators could fill their own networks and then attempt to fill the relatively boundless capacity of the fixed networks by signing up more customers for any given cellular network configuration.
And if the cellular operator charges more than it is paying out in WiFi costs to ISPs, then it is still making a profit on that bandwidth. It could even charge less for WiFi if it wanted or give back credits against caps when people’s usage profile meant they spent more time on the cheaper WiFi network. There are an almost infinite number of new models that cellular operators could try out with this arrangement including ultra-cheap mostly WiFi serviced phones.
‘We made some calculations on production cost for network upgrades. You take the way Ziggo in the Netherlands has created 1 million Homespots with a software upgrade. The incremental cost of putting the remaining unused bandwidth to use is close to zero. All the money they get for it goes straight to the bottom line,’ said Visuri.
‘But for the mobile operator to build more network, it has to calculate site rentals, tower building, backhaul, power. We have calculated the cost of building all this and think it can almost never be recouped. The cost of the WiFi usage would be more than paid for by improved churn, and by not having to spend on Capex.
Visuri then talked about how well accepted his system has been, ‘We are talking to ISPs who collectively have 82 million broadband lines in the US, and 80% of these have WiFi connected and could be used in this market,’ Which makes the Faultline prediction of roughly 60 million Homespots ready to be turned on in the US perhaps slightly low, maybe it’s closer to 64 million.
But this is not all about cable companies. Attach rates of WiFi per broadband line are higher in telco providers than in cable providers, with virtually all new DSL and Fiber lines being offered with WiFi. Cable has perhaps only been installing in this way for a couple of years.
So when a Telco does this, does it replace FON, the service for sharing SSIDS between customers of telcos? ‘We can do a deal with FON,’ said Visuri, ‘We can do this with every Hotspot owner, like Boingo. It works with anyone.’
As we said earlier we think this changes capex calculations completely ‘ it discourages cellular operators from building their own WiFi capabilities based on economic considerations. It will be years before they have exhausted the spare capacity in these networks. But if cellcos want more influence over WiFi, they can build out with that purpose in mind. The most effective way for cellular operators to deal with WiFi is to take a leaf out of Vodafone’s book and buy a fixed line operator as it has in Germany with Kabel Deutschland. Right now this extra capacity has yet to be turned into revenue, and once it is, it will all be in the share price of the fixed line ISPs. Better to buy now before the financial markets realize what a goldmine the ISPs are sitting on, and any cellular boardroom where this is explored, a business case can be made for annexing WiFi in volume right now.
Going back to the trading exchange method, the real genius is that for instance Verizon can be offered one price profile by its partner Cable operators and AT&T a much harsher, almost profiteering high price profile. If AT&T phones don’t use the WiFi offload from US cable companies, they may choose to lower them just enough to make money out of their rival. Eventually both sides would do business, but at the right price for companies that are sworn enemies, which is far better than just blocking rivals.
And this too will change handset design. In order to maximize handset WiFi sensitivity, handset makers will need to upgrade the power amplifiers they use on WiFi, and pick those which best reach through the walls of homes. The range of the handset WiFi is the limiting factor in all Homespot use.
‘We already have signed about 50 to 60 NDAs, and we can’t talk about which suppliers they are with, ‘ said Visuri. ‘We have been in discussions in the US with nearly all the tier 1 suppliers and all the tier 2 players and nobody has stopped talking to us yet. But these companies move slowly and establish long trials, while tier 3 players are likely to sign more rapidly and we hope to have our first deals later this year.’ And during 2014 we would expect this to shift to tier 1 MNOs and show in their capex falling and opex rising, in subsequent years.
What happens to WiFi once this is common knowledge is anyone’s guess, but the next spectrum allocation from a regulator, post 2015, might show a preference for unlicensed spectrum, added to the pile that WiFi already has.