Muni Wi-Fi 2.0: smaller targeted networks, flexible business models
Glenn Fleishman has written an analysis of municipal wireless networking, tracing its development from large citywide “free” networks (which did not get built out) to smaller, more targeted networks (Minneapolis, Oklahoma City). He focuses much of the article on mesh vendor Meraki’s strategy: unwiring small neighborhoods, apartment complexes and downtown areas. EarthLink’s pullout from the municipal wireless market, the failiure of wireless ISPs such as Kite Networks and MetroFi, and more importantly, the financial crisis which has drained cities and businesses of sources of capital, it is highly unlikely that we will see any large scale municipal wireless broadband networks in the next six to twelve months. Nevertheless, the demand for Wi-Fi access continues to grow. How will service providers meet that demand?
Regular readers of Muniwireless are familiar with what went wrong: service providers that underestimated the cost of unwiring large areas and overestimated the revenues they would derive from subscriptions and advertising; lack of financial support from cities and large anchor tenants; significant opposition from incumbent telecom and cable operators, wireless equipment that did not work as the vendors and buyers believed, and so on. But Glenn also points out that when cities issued bids for these networks in 2005 and early 2006, the iPhone and similar Wi-Fi enabled devices had not yet been sold. Thus, the need for outdoor and indoor Wi-Fi coverage (including on public transport), especially where there are a lot of people looking for information on websites and Google maps, was not as keenly felt as it is today.
Many people were surprised to learn that in the third quarter of 2008, the iPhone was the best selling phone in the US, beating out the Motorola RAZR and the Blackberry. iPhone owners have been sending and receiving massive amounts of data via AT&T’s 3G and Edge networks. Some claim that the massive data use has contributed to outages in AT&T’s cellular networks and that AT&T’s recent moves — giving away free Wi-Fi at Starbucks to iPhone users and its acquisition of Wayport (an operator of Wi-Fi networks in McDonalds and various hotel chains — shows that the company wants more people to use Wi-Fi more often in lieu of the 3G network.
Business models are also evolving and becoming more flexible. For some providers of Wi-Fi, it’s an amenity. Public transport operators in Europe entice people away from cars, airplanes and their competitors by giving away Wi-Fi service. Lower end hotel chains have already been doing this for years. In some cities, Wi-Fi is not so much a business, as a service to low income families. For other firms such as iPass and Boingo, who charge their customers a monthly fee in exchange for access to networks around the world, the selling point is convenience and ubiquity. iPass goes one step further: they target enterprise users who are concerned about security on the networks.
Still, I see a problem with these small independent networks: if you are walking with an iPhone down the street and are looking up information online, there is no handoff between the networks. So as soon as you move out of the range of one network, you lose your connection.
Are we getting close to the day when these networks do seamless handoffs? When will all of these “small pieces loosely joined” (title of a book written by David Weinberger) become one?
_2.gif)
A key point that I take away from Meraki, not accepting their word absolutely on this (of course) but relying on the maps they’ve produced of active networks and other details, is that small networks get bigger.
With expensive gear that requires controllers and network operations centers (NOCs), and which have very high reliability promises, somewhere between cable/DSL and T1 service, a small network isn’t financially feasible, while a large network has very high overhead. These big networks required aggressive customer acquisition models, and ubiquitous coverage among other factors.
When you scale large networks bigger, costs don’t rise in a linear fashion. Adding more customers might result in a large increase in customer support budgets. Or if you go from needing one person 24 hours a day in the NOC to two, your costs more than double given absenteeism, sick days, etc. And so on.
Meraki’s approach has to play out, but their idea is to cut the NOC out, essentially, and remove the cost of controllers. I’m unclear on what their maximum throughput is–1 Mbps? 5 Mbps? It’s hard to know. But their equipment is being largely used in best-efforts networks, which have lower expectations, and often in free networks, which come with fewer promises.
The real question is that even though Meraki says that nearly every network they seed grows, sometimes quite quickly and to a large size. As these networks get bigger, can an operator (fee or free) keep costs very low and still serve users? Or do management issues, backhaul, and other elements escalate faster than the cost savings in these networks?
The Meraki model is interesting but the equipment is still very short range and requires a lot of hops or a lot of backhauls. I believe there is better outdoor equipment with greater range available at a better price that can provide a better cost model. The big issue is still finding applications that can utilize it. I believe they are there and we are designing for them.