July 2007 Archives

Well, we all know the winning model for muniWiFi is not here yet. Here is the current litany of pain from Fast Company: Fast Cities Struggle to Go Wireless

People are discovering that WiFi cannot compete side by side with the telcos to provide universal broadband access over a large geographic area with anything like the QoS that people expect. Cell gives people universal coverage. They want their muniWiFi to provide it, especially if they are being asked to pay for it.

Mesh would be the way to go, except the attenuation (degradation of signal) between hops makes the technology — so far — not nearly as robust and cost effective as it needs to be. Earthlink/Philly/Tropos is what is cited as the main example.

What everyone seemed to forget as they were laying out their plans for a wireless municipal network is that WiFi by FCC regulation, and given where it is on the spectrum, doesn’t penetrate well — into buildings, through trees, down the hall, etc. People also forgot that this is open, unlicensed spectrum, subject to interference from cordless phones, microwave ovens, baby monitors, fish tanks (Wi Fi can’t penetrate water), other WiFi networks, etc.

WiFi itself was created by people who took the thin slice of free or junk spectrum alloted by the FCC and went with it well beyond what anyone could have anticipated. That said, there are continued limitations with WiFi that correspond to laws of physics. More robust spectrum at a higher power level is what is really needed.

In the meantime, let’s take what WiFi does well — provide local broadband connectivity. Let’s create a local wireless broadband experience within a neighborhood public space or commercial corridor.

Forget city networks. Too big, too bold, wrong paradigm for the spectrum you have been alloted. Dig “community wells” rather than trying to lay all the pipes necessary for “indoor plumbing.” Don’t go toe to toe with cable and the local telco and try to be the third player. You will lose because you will have a lot of the headaches and overhead of a telco — the in-house wired infrastructure, a large sales, marketing and customer service force — and not nearly the means given WiFi’s limitations to deliver a service that can compete in terms of price and quality, not with DSL prices continuing to drop, and I daresay $99 voice, cable and internet triple plays to be had at internet speeds far far higher than what WiFi would provide indoors.

We should try to instead create Community WiFi, as opposed to Consumer WiFi. Establish Hot Zones that are highly local, not mobile or municipal. WiFi Salon believes WiFi should be established the community’s centers — the schools, libraries, parks, public squares, the business districts. If you try to bring something to everyone everywhere, most certainly you will spend too much and still come up short because in the end you won’t be able to deliver enough to individual homes and offices. As a location-based service in key areas — well that is another matter.

The Wall Street Journal today had an interesting article on one potential business model that would support MuniWiFi: Advertising.

Here’s the link:).

The article discusses how the ad supported muniwifi model has not taken hold because large advertisers cannot buy ads in bulk or across large areas because WiFi is a very local and small scale thing. Each municipality is a separate negotiation. If you are selling or promoting a national brand, that is a problem.

One potential answer, according to the article, is to aggregate the locations, get a hundred providers to sign up with your ad service, and then turn around to the major brands and sell that space. This is the JiWire strategy. They are a WiFi directory service, a provider of WiFi security solutions, and now in conjunction with Microsoft a provider of advertising network services.

The value proposition is this: You are a Wireless Service Provider (WISP). You run their ad platform on your network. They have many dozens of networks signed up. They are able thus to grab ad dollars from national advertisers because they now have the reach and scale necessary, and they split the revenue with the WISP. Free WiFi, ad supported.

Will this work? I would say that unless this otherwise top-down platform allows also for a means by which to create highly local ads, and support user generated content for reviews, recommendations, new locales, this alone will just will not work for (free) muni WiFi.

A director from Digitas noted that people would be very likely to tune when watching the obligatory ten second video that would pay for the free WiFi. Maybe so, but the medium — broadband wireless internet — is going have advertising possibilities — interactive, location-based, IP based — that this re-purposing of desktop ad technology just lacks, and which is now, as noted, real tired.

We respectfully submit that WiFi’s strength is that it is the internet, localized. Local content, services, and yes, advertisements. Advertising is relevant to the extent that it is actionable. With WiFi, the customer is the point of sale, and a WiFi Zone and a Commercial Zone can be one and the same.

Integrate local advertising with a local interactive map, geolocate the content, enable user created content. Keep it hyper local, aggregate. The Long Tail, if you will.

We believe we are a ways from where you can show the right ROI to local businesses using traditional ad placement sales: How many devices / users would you have to have on a network to create enough sales to even pay for a $135 ad for a pizzeria. The question is, where do we see the right density of devices in use at WiFi Hot Spots —2008? 2010? — to command the ad rates needed to sustain the local WiFi network? Microsoft and JiWire have their projections. More revenue sources beyond advertising is required for now.

We believe strongly that lighting up a commercial corridor and seeding the area with wireless screens, kiosks, handhelds, and providing a local interactive map will create advertising solutions that will be all the more effective for being part of an “immersive” wireless experience. Do large brands even have a place here?

WiFi Salon has been asked often “What is the capacity of your network?” “How many simultaneous users can you have at any one Hot Spot?” Currently, WiFi Salon has 3 Mbps DSL lines to all its locations, with Columbus Circle and Sheep Meadow upgraded to 8 Mbps with Covad’s new ADSL2+ lines, and with Washington Square, Union Square, Summerstage, and The Delacorte/Great Lawn getting the same upgrade.

What does what we have now deployed, 3 Mbps, get us? If this DSL line was say serving a small office and several people at the same time were trying to download something, performance would suffer unless of course there was some kind of bandwidth throttling.

By bandwidth throttling, we mean apportioning to each user a certain amount of the pipe when they download — for us it is now 200 kbps. As a Youtube video only requires 55 kbps, we consider this a sufficient amount.

In this scenario, if say 15 people were downloading a video four times as bandwidth intensive ( better resolution, larger size ) as the standard Youtube video at the same time, that 3 Mbps pipe would accommodate only 15 users.

Real world, user behavior is quite different. Culling from years of data analysis in administration of the Bryant Park Wireless Network, Marcos Lara, my CTO, and network architect found that the most congested his network ever got was ~75 users on about 2.4 Mbps of the available 3.0 Mbps.

Bryant Park has, to be sure 2 T-1s, which are dedicated, versus our one DSL at each location, which is “best effort.” Shouldn’t 2 T-1’s then outperform? Not necessarily, so we manage the network, allocating bandwidth per user, shaping traffic.WiFi Salon believes that we will witness the same performance on its network. 3 Mbps well managed will support around 70 users, that we will see the same performance on the parkwifi network that Marcos was able to achieve at Bryant Park.

At ~$900 a month for two T-1’s it would not be financially possible to install T-1’s at all 17 of our locations. Byrank Park has the luxury of being a location that is at one physical address, we do not. This leaves us with DSL as our only consistently reliable solution.

So the question is can we provide a similar experience to Bryant Park’s dedicated two T-1s (1.5 Mbps each), versus our single DSL lines at each location.

The answer is yes, thru traffic shaping. Each of our DSL location employs network traffic shaping to ensure each user gets a quality experience similar to that Marcos was able to achieve at Bryant Park.

For example, in Washington Square recently, we had 23 simultaneous users consuming just 688 kbps of bandwidth, or less than 25% of the 3 Mbps capacity. This is due to the nature of web surfing and Internet usage. The truth is that web surfing behavior in the real world consists of people mostly reading and writing documents, which uses up no bandwidthand the time a user is actually transferring data is quite small. this is not the case with VOIP and Video, but those still represents a small but growing amount of Internet traffic at this time.

Over the next several weeks we will be nearly tripling our capacity with the new ADSL2+ lines in Washington Square and at five other key locations. this will further increase each users experience. however there are still other factors involved in how many simultaneous users can we support.

If more than 70 users were all trying to transfer data with the same wireless access point at the same time any wireless network would start to experience wireless network congestion that limits the all the connected clients ability to communicate effectively.

This “crosstalk” — too many trying to be communicate on the same wireless frequency can only be addressed thru the addition of more wireless access points to share the user load.

WiFi Salon would start to worry about crosstalk when usage reaches around 70 simultaneous users. Such as if there was an event we were hosting that might attract a larger number of users for an extended period of time. in those cases we bring out our big guns, very special wireless hardware, that can really handle a heavy load. it is expensive but well worth the cost if you’re hosting a special event.

One such piece of equipment is a world class radio manufactured by Altai. For distance, performance, quality of service, and we have seen nothing better.

Combine Altais with ADSL2+ in a commercial district, you really would have something. It works really as best as WiFi can within very noisy urban and congested environments.

WiFi Salon would start to worry about the crosstalk problem at around 70 users. If there was an event we were hosting that might attract larger numbers, we’d bring out our very favorite piece of hardware, the Altai.

This is a world class radio — for distance, performance, quality of service, we have seen nothing better. Combine Altais with ADSL2+ in a commercial district, you really would have something. It works really as best as WiFi can within very noisy urban and congested environments.

As it is, when we look at historical data on usage, we note at no more than 1% of park goers will log on a park’s WiFi network on a given day. In other words, 7,000 people would have to be on the Sheep Meadow at the same time for us to be concerned we were hitting capacity, at least in terms of crosstalk. Crosstalk, and not bandwidth, would be the first problem — and now especially at Sheep Meadow after its upgrade to ADSL2+.

So if crosstalk appears to be the limiting factor, and not bandwidth, why are we adding the ADSL2+ lines? What if we were able to give each user a larger ‘pipe’ than is otherwise available in public WiFi Hot Spots? How much better can we make the user’s experience? How much more in the way of multimedia could we support?

We want the very best speeds available, and we want to support as many users as possible. We believe that WiFi’s true strength lies in its ability to deliver rich local multimedia content, services and experiences, and local high speed wireless connectivity.

WiFi, a low power piece of open spectrum as regulated by the FCC, has many inherent real world limitations. The WiFi industry, born by accident and incredible resourcefulness, is engineering feat by engineering feat, eke-ing out further performance enhancements within the very restrictive parameters that constitute its regulation.

WiFi Salon’s mission is work with companies that are seeking to push the technology as far as it can go, where let’s say for now 70 people at a WiFi Hot Spot can all have a rich media experience. With the parkwifi network, we are providing park goers in New York with a free amenity that we will continue to scale as needs evolve.

KQED emailed me a link to an interesting story on how a consortium of four companies, headed by Cisco and IBM, are planning to build a 37 town muni WiFi network in Silicon Valley.

I still want to know about device density i.e. how many WiFi enabled devices are there now, how many will there be in three years. Device density is crucial. That is the pool of potential users.

It’s a $100 million dollar experiment for them. I like their chances more than I did the ill-fated Cometa. That consortium of Intel, AT+T and IBM went through $40 million back in 2003 in the attempt to create a wholesale backbone/backend for WiFi. No devices then, some now.

wrote a little piece on WiFi Salon’s upgrade of the parkwifi network through ADSL2+ lines from Covad.

Small point — we are in ten parks (not 17 — that’s the number of locations).

We are glad that we were able to provide a leading edge service like Covad’s ADSL2+ to New Yorks park goers as part of our free parkwifi networkWiFi. We have about tripled capacity at Sheep Meadow and Columbus Circle. That means we can support three times the users and much more in the way of video streaming. Washington Square, Union Square, Summerstage and The Delacorte / South Great Lawn are also now in process.

We are now seeking to populate the portals with local multimedia content from media companies, arts and cultural institutions, from The NYC Parks Department and various parks conservancies, and public entities.

For Immediate Release

New York, N.Y. – July 23rd, 2007 – WiFi Salon announced today that it upgraded its free 18 location parkwifi network in New York City with ADSL2+ from Covad Communications Group, Inc. (AMEX: DVW), a leading national provider of integrated voice and data communications. This upgrade will allow for significantly faster delivery of localized multimedia content from public and private sources, including the Department of Parks and Recreation. Covad’s ADSL2+ is a next generation broadband service offering speeds of up to 15.0 mbps.

“By upgrading to Covad’s next-generation ADSL2+ service, WiFi Salon has significantly enhanced its ability to provide New Yorkers with free high speed WiFi on our parkwifi network. People want video, they want fast downloads. We can now scale to meet the growing demand not just for WiFi, but for high bandwidth and multimedia in public spaces,” said WiFi Salon CEO Marshall Brown.

“We are very pleased that WiFi Salon has chosen Covad ADSL2+ to power its network of WiFi hotspots,” said Lisa Graham, Covad senior vice president of sales. “This next-generation broadband service is uniquely capable of providing the bandwidth to support the multimedia experience that technology-savvy New Yorkers demand.”

WiFi Salon currently operates 18 locations running in 10 city parks in four boroughs. WiFi Salon and Covad now provide ADSL2+ in Central Park at The Sheep Meadow and in Columbus Circle. They are in the process of upgrading the parkwifi network to ADSL2+ in Washington Square, Union Square, The Delacorte Theatre / Great Lawn and Summerstage.

Full list of locations and coverage areas can be viewed here:

http://www.nycgovparks.org/sub_your_park/wifi/index.html

About WiFi Salon:

WiFi Salon is a leading free wireless service provider creating neighborhood hot spots where people can connect, share, and build communities simply using the newest approach to next-generation wireless applications, services and experiences. The company’s premiere installation is in New York, where it is headquartered, having secured the exclusive concession from the New York City Department of Parks and Recreation to establish and operate 18 Hot Spots in ten parks in four boroughs. These local portals for these locations can be found at http://parkwifi.portalize.net

1. Identify a key location where people congregate, ideally a commercial district near a college or university, with some nearby arts and cultural institutions.

2. Do a site survey. Determine the equipment and costs, where people will let you mount antennas, where you could bring in ‘backhaul’ — DSLs, T1s, Fixed Wireless, etc.

3. Work with the local non-profits — Business Improvement Districts, Community Boards — to determine what wireless applications, services and experiences would prove useful for the area in question. Local advertisement, both in the sense of creating awareness of the network, and creating advertising opportunities for local entities, is crucial.

4. Cost out the development of the local portal; use a phased approach. The portal and its associated features should go incrementally. What goes up first? What is stage two or three? At each stage, you will learn what is working for your users and what isn’t.

5. What other hardware based add-ons would people / local businesses find useful or compelling? Interactive flat screens? Kiosks? What devices — handsets, tablets, game consoles, VOIP phones, non-browser based devices — should the network actively support and promote?

6. Produce an ROI analysis: What is the demographic you wish to draw? What are the assumptions for usership? What will each user bring per day economically? If a WiFi Hot Zone covering ten blocks costs $100K to first build and maintain Year One, but $20K for each Year Two, and Year Three, but we can project an average of 300 sessions a day over those three years, is it worth 43 cents per session to bring people to this Hot Zone to work, dine, shop? How much revenue could be derived from local advertising to around 100,000 people a year?

7. Identify sources of funding — sponsorships, partnerships, grants, advertising, e-commerce. Depending on the place you want to build the network, it’s target audience and aims, the funding sources will be different.

8. Get the timing right. This model will work only when there are enough locals who have WiFi enabled devices. You need critical mass. We are getting closer…

There have been two very large impediments to the development of Muni WiFi: Few devices, and few good open networks.

We have been then in very much a ‘chicken or egg’ situation — there has been no strong impetus to building muni WiFi networks because not many had such devices, but few had devices in part because who would buy them absent networks?

Well now that we have the iPhone, and they have shipped 500K of them, the pressure will mount. With Nokia (disclosure: the parkwifi network’s sponsor) shipping their NSeries devices and Sony, Samsung, etc all shipping WiFi enabled gadgets, with T-Mobile now selling a service that is dual WiFi/cellular, the pressure will grow all the more. Good public WiFi will simply become a priority now not for the 5% who love technology and are passionate abou WiFi, but for many millions.

It will also become a priority at some point for the carriers, unless the cell network will in fact be able to support the demand for multimedia content. For now, it looks as though WiFi will be a complementary means of getting high speed downloads and video streaming on a dual WiFi/cell device, with the Hot Spot a nomadic stopping point in an otherwise mobile environment.

WiFi Salon is building Hot Spots that are community wells. Rather than aspire to provide indoor plumbing and running bandwidth, WiFi Salon provides neighborhoods with places to connect — both to the internet, and to the neighborhood itself.

The parkwifi network provides NYC parkgoers information on the surrounding neighborhood — where to dine, points of interest, where to shop .

We will be adding video and news feeds, then enable “community created content.”

WiFi Salon is dedicated to bringing the latest in WiFi enabled applications, services and experiences to New York City. With the parkwifi network in 18 locations in ten parks in four boroughs, WiFi Salon is bringing free WiFi to NYC’s park visitors.

The network, sponsored by Nokia, is a showcase for innovation in wireless media and technology.

The NYC Department of Parks and Recreation posts a useful map of our network here.

First, a thanks to Dan Carty for a well put together set of articles. He did New Yorkers a service by reporting on this free amenity, showing people where to find coverage, and how to surf safely.

Work, Unplugged

New York City Leaves WiFi to Others

Dos and Don’ts for Safe Surfing

—