Your browser is not supported. Please update it.

15 January 2008

With the metrozone bubble burst, does Meraki offer a new way forward?

The overhype that surrounded San Francisco’s municipal Wi-Fi project, especially during Google’s shortlived involvement, and the collapse of the scheme last year, have mirrored the rise and fall of the whole metrozone concept, especially in the US. Now start-up Meraki has raised funds to readdress the San Francisco challenge ‘ will it, too, prove representative of a more general trend, in this case towards community-based wireless, as already expounded by Spain’s FON?

There were key lessons learned in the rollercoaster of the US metrozones ‘ that Wi-Fi is not a robust wide area technology; that unlicensed spectrum does not deliver the quality of service required in a sophisticated user market; and that idealistic notions of bridging the digital divide soon founder amid the quest for a viable business model. The main models that have been mooted are full or partial public funding (usually a non-starter with taxpayers); conventional fee paying services, which needs high quality of service and in a hi-tech city like San Francisco is very competitive; and free, ad-supported offerings in the Google mold, a model that has not yet been proven in metrozones, and certainly plays against the universal access theory because of the need for advertisers to target only populous and/or affluent user bases.

Two other models remain with some hope attached to them. One is to build the network primarily to make local government services more efficient, or to support new ones such as enhanced public safety communications. This justifies public investment in the system and harnesses the advantages of the municipalities’ control of infrastructure such as lamp posts, and commercial services offered on spare capacity in the network, usually by a private partner, are a bonus rather than essential to the success of the project. This is the most common model adopted by US metrozones that are continuing to be built out or already functioning with some degree of success.

The other is a community approach. There are many variations of this and most have been small scale and run by local enthusiasts. However, FON blazed a trail with its system of offering free services to users prepared to open up their access point to third parties, thus creating a self-seeding ‘organic’ network. Meraki, a spin-off of Massachusetts Institute of Technology ‘ which has been experimenting with mesh Wi-Fi models for years in its Roofnet initiative – is proposing something rather similar, and for some reason has, like Google/EarthLink before it, chosen the over-competitive San Francisco to test out its ideas before spreading them to other cities and countries. It recently raised $20m in venture capital ‘ from Sequoia Capital, Northgate Capital, and DAG Ventures ‘ to fund its plans and now aims to instal a citywide free wireless network in the Californian city. As well as the VCs, Meraki last year attracted funding of ‘under $1m’ from Google, demonstrating the search giant’s continuing interest in new platforms that will expand use of the broadband internet and therefore its services, despite the failure of its own ideas of being a metrozone owner.

Instead of renting private or municipal sites on which to instal access points, as EarthLink intended to do before the previous city plan fell apart, Meraki will spend its cash on distributing about 15,000 Wi-Fi mesh devices to residents. These semi-open source devices, developed by MIT’s mesh programme, will be low power repeaters, that can be installed in windows or balconies, and will communicate with one another to form an organic mesh network coordinated by Meraki central servers. One in three or four devices will be backhauled to the Meraki servers using DSL links funded by the start-up (so users do not need to have a DSL line in place to use the service). This will be a large enough number of devicess, Meraki says, to seed a mesh that will blanket much of the city, and will bring free access immediately to about 40,000 people. The use of techniques such as supersmart routing techniques, also developed at MIT, will enable the network to offer better QoS than many meshes and to deliver almost 1Mbps of connectivity to each user, claims the start-up. Among the techniques is support for so-called Hazy Sighted Link State Routing, which de-emphasizes routers on which a large number of dropped packets is seen.

Of course, there will still be huge issues of coverage, quality, signal strength and interference,. This is inevitable in any system using unlicensed spectrum, especially the overcrowded 2.4GHz, and can be worsened by the somewhat random network planning enforced by organic mesh approaches. Meraki’s router technology will help mitigate the effects, and it has also done extensive anti-interference planning ‘ claiming it has “identified and worked around more than 20,000 sources of interference” in the trial deployments in San Francisco. This highlights another lesson of the early Wi-Fi mesh craze ‘ that rigorous network planning and often expensive anti-interference exercises are more important in a mesh system, not less, and can significantly hit the economics.

But though Meraki does genuinely seem to have cast off the overoptimistic ‘hang an access point on a lamp post and off you go’ approach to early metrozone planning, it will be even more important to its success to manage expectations. A network relying on users to keep their access points open to others, and in unlicensed spectrum, will never be able to deliver the quality, security and reliability that most people take for granted from their access services. This is fine, provided Meraki is clear about what it is offering ‘ decent bandwidth and a best effort system, not a full commercial quality service for free. If this proves successful, as an alternative to the commercial networks, not a competitor, Meraki may well find itself attracting partnerships with larger carriers, as FON has with BT, Time Warner Cable and others. In the first instance, though, its main aim is to rent out its mesh to small WISPs offering low cost or innovative services.

The FON effect:

With the apparent failure of the ad-supported free metrozone idea, operators and city authorities continue to experiment with other models.

In New York, broadcaster CBS is revisiting a model that was very popular before the Google-inspired municipal trend gathered steam ‘ sponsored hotzones. It is blanketing midtown Manhattan with Wi-Fi access and delivering news, sports and entertainment for free to WiFi-enabled devices. The ‘CBS Mobile Zone’ will cover a 20-block area from Times Square to Central Park and from 6th to 8th Avenues and the company is giving away routers to local businesses such as restaurants and coffee shops. The initiative is part of a pilot program with New York’s Metropolitan Transportation Authority and New York City Transit, and CBS has also partnered with mesh equipment vendor Tropos and free hotspot supremo FON. The system also uses CBS Outdoor, the media company’s network of billboards and marketing displays throughout the US, to house access points. This model has the advantage of multiple revenue opportunities ‘ paid-for third party advertising, free advertising of CBS’ own services, and the chance to partner with retailers of music and other content or products, with a revenue share.

FON, which includes Google and Skype among its investors, sells or gives away its own branded wireless router, called La Fonera. It has distributed about 370,000 of these worldwide. These devices split a Wi-Fi signal in two: one is used as an encrypted channel for the user (or Fonero), while the other is available for public use by anyone close by. If they are also Foneros, they can gain access for free, otherwise they pay $2-3 for 24 hours of access. Users who agree to open up their routers gain a cheaper home subscription as well as access to other people’s Foneras. FON’s argument is that consumers should not have to pay twice for access – at home and on the road – and that its system is a low cost and simple way to give people, in effect, the always-on access, and the ‘carry it with you’ broadband link, that mobile broadband offerings like WiMAX will eventually offer.

A year ago, CEO Martin Varsavsky said he would sign six major US deals within six months, with cablecos his primary target. ‘There is a natural fit with cable companies as we enable them to further expand their offerings,’ he said.

The FON service is available on a limited basis in seven countries – the US, Spain, South Korea, Germany, France, Italy and China. The company is subsidizing its devices in order to expand uptake rapidly though it will reduce costs next year by making its next generation client available as a software download. FON predicts it will have one million hotspots around the world by the end of this year, two years earlier than its initial projections.