WiFi Service

Executive Summary
The goal of this proposed project is to provide Hayden Island residents, visitors and shoppers with “free” WiFi while providing a “hot” communications back-up for emergencies. Here’s a shorter PDF version of this paper.

This paper proposes four WiFi hotspots on Hayden Island that can provide “free” internet access. No cellular or land-lines are used. No landline power required. The hotspots are solar powered.

This $12,000 system is designed to work after a major earthquake, using satellite internet connected to four hotspots. Ad revenue is anticipated to cover operating expenses of $200-$300 a month.

It’s modeled for maximum sustainability. It does not require ANY subscription revenue. Advertisers pay for the service. It uses Portland-based Open Mesh hotspots which are cloud managed and mesh-enabled for seamless roaming with maximum security.

Project Overview
This proposed project would use Hayden Island’s RV park WiFi network as a model for the whole island. It differs in that it provides internet access using a consumer satellite rather than a DSL or Cable modem. Primary service would use Gigabit fiber to the central hub. The satellite would provide back-up. Digital signage provides the revenue.

A satellite-based internet terminal on the roof of a hotel (or a tallish building like Best Buy or Home Depot) would enable 4-6 hotspots to be connected to the hub wirelessly. The system would operate without land lines, cellular connections or AC power and provide emergency information and communications after an event like a large earthquake.

The $99/mo Exede satellite internet service can deliver up to 150 GB/month, making it more cost/effective than cellular data plans.

Satellite internet access is still relatively slow and expensive compared to virtually all other options, including Comcast, CenturyLink, Stephouse Wireless or Google Fiber. Those providers would supply our primary feed.

Satellite internet has one thing landlines and cellular don’t offer: it will work if there’s no power or landlines are down. Satellite internet is used as a backup to the landline. A load-balancing router automatically kicks in the satellite link if the landline goes down.

A second innovation is cloud control of remote access points and seamless roaming. This provides users with the highest level of security, and the convenience of freely roaming between hotspots without having to “sign in”.

A third innovation is digital signage. It provides the revenue that makes the free WiFi network sustainable.

Combining satellite hubs, cloud-controlled hotspots, and digital signage, this solar-powered WiFi hotspot system should be available through the apocalypse AND make money.

Self-sustaining operation
Ad revenue is expected to pay for monthly operation. Monthly operating costs are estimated to be $200 – $300. With a dozen different advertisers, each paying $20-$80/monthly, revenue is expected to cover operating costs. A 501(c)3 WiFi operating company would provide the public service elements and enhance funding opportunities for a demonstration project.

Ad revenue would be provided with digital signage using screens like Samsung’s $450, 19-inch Android Table which might generate $100-$200/month for each hotspot kiosk. Additional revenue might be generated through an associated website.

Digital signage might be a better model than a touchscreen tablet operated by users. Let people use their own phones to search for information.

Ad revenue in this budget system is estimated to generate a total of $200-$1000/month for four hotspots ($50-$250 for each hotspot per month). A modest goal.

This proposal combines a high capacity satellite terminal with four, solar-powered hotspots at strategic locations around the island. Perhaps two Kiosks would be located on the West side (at the shopping center’s street furniture and near Jantzen Beach Moorage or the business park), while the two on the East side might be located at the Red Lion or Oxford Suites, with another near Columbia Crossing or Lotus Isle Park.

More Kiosks could be added after their cost/effectiveness is proven.

The total front-end cost for the satellite terminal and 4 self-contained hotspots is estimated at $12,000. If spread over 3 years, that would amortize at approximately $350/month. It would demonstrate sustainability if, after 18 months, each Kiosk generated approximately $150/month ($600/month total). A grant covering initial front-end costs would show the viability of the self-sustaining model.

Hayden Island, Portland’s only island community, is strategically important. It is the foundation for the Interstate-5 bridge, connecting Vancouver WA to Portland OR. It is also home to some 2,000 permanent residents and some 10,000 people (when visitors to the Jantzen Beach Shopping Center are counted).

Numerous studies suggest the island will be cut off, without access to transportation or communications for a period of weeks in the event of a Cascadia Subduction Zone earthquake. More prosaically, an iPass survey found 74 percent of business travelers would chose Wi-Fi over cellular data while 87% of hotel users were frustrated when Wi-Fi is not available.

Needs Addressed with this proposal.

  • Provides a self-sustaining model for communications that operates independently of land-line power, as well as cellular and internet connectivity.
  • Delivers life saving communications for public service workers and the general public.
  • Utilizes state-of-the-art satellite and WiFi gear for greatly improved cost/effectiveness.
  • Provides residents, visitors and shoppers with a “free” amenity while providing a “hot” communications back-up for emergencies.

Fiber-fed Hotspots
The free LinkNYC WiFi kiosks in New York City are an interesting model. Flat screens provide ad revenue. Each Kiosk is supplied AC power and Gigabit fiber. The big screens provide only digital signage, while a small tablet on the side enables web surfing and usb plugs provide cellphone charging.

NYC’s WiFi access points use the Ruckus Wireless ZoneFlex 802.11ac wave 2 4×4 access point with Qualcomm VIVE 11ac Wi-Fi silicon providing multi-user MIMO technology, supporting hundreds of simultaneous connections. The kiosks will also incorporate a 3G modem to support 911 calls and emergency services.

The Ruckus Cloud Wi-Fi Management system, used in NYC, allows one person to manage hundreds of hotspots, such as their ZoneFlex T710, an outdoor Wave 2 hotspot, inside the kiosks. Cloud-managed WiFi provides seamless roaming.

Digital Signage
Digital Signage is the revenue generator. There are many different systems. You can use large screens, dongles like FireTV, or tablets. A large 42″ monitor would take lots of juice. Our restricted power budget might use a 19″ Android tablet or a 12″ Android tablet with an OLED screen which is bright in daylight and power efficient. It would sit behind heavy Plexiglas and power down in the evening.

The ScreenCloud Signage app is one of many options. It turns an Android tablet into a beautiful digital sign while the cloud management system lets you bring in social media displays and presentations. It costs $20 per screen per month.

A small LED might might illumiminate a large, colorful transparency, while an inserted 27″ monitor powered by USB Type-C or 19″ tablet could provide signage behind a plexiglass cover. How much revenue would each kiosk generate? Let’s say $125-$150/month ($500/month for 4 hotspot/kiosks).

Cellular-fed Hotspots
Cellular has monthly data caps, making it an expensive option. Since the average home broadband subscriber uses 100GB per month, mobile hotspots just won’t work for a primary Internet connection, let alone a shared public hotspot.

Bigbelly’s Wi-Fi-enabled, solar-powered trash cans are a relatively cheap and useful solution. Stash a mobile hotspot (with cellular backhaul) in the lid and slap a poster on the side. Done. But a cell-fed public hotspot is likely to blow through 20GB in a week.

Netgear’s AirCard 790 mobile hotspot, for example, could be enhanced with a $25 external LTE antenna. The whole thing could be powered by a $50 30Ah Portable battery and kept charged with a $60 5V/20 watt panel. That totals about $350. Cradlepoint’s mobile hotspot can also use the 700 MHz First Responder LTE network, FirstNet

But if each cell-fed hotspot cost $150/month to operate (at 30GB/mo), it would not be sustainable. It would cost ~ $600/month to supply internet access for 4 hotspots. That’s more expensive than nearly all the cable, DSL or satellite options. And you wouldn’t have any emergency back-up.

Bottom line: satellite data (at least for Viasat-2 and Hughes Jupiter-2), will cost significantly less than cellular. The digital signage generates monthly revenue, and it works if cellular goes down. The incremental cost of adding satellite data is negligible, adding perhaps 10% more to the system cost. A public/private partnership model might benefit everyone.

System Description
The satellite hub links to four remote solar-powered WiFi hotspots using a wireless bridge. Hotspot coverage area can also expand with additional meshed hotspots.

Today, ViaSat’s Exede Internet service delivers about 12 Mbps to individual users with a 25 GB monthly allowance for around $99/month using the company’s first generation, 140 gigabit-per-second satellite. Currently, satellite has about double the monthly bandwidth cap of cellular for the same cost. Cost is headed down.

ViaSat-2, scheduled for launch in spring of 2017, will have 2.5X the capacity of ViaSat-1, making satellite-fed hotspots practical. ViaSat’s follow on, ViaSat 3, scheduled for launch in 2019, is said to have as much bandwidth as all the rest of the satellites in the world combined.

ViaSat competitor Hughes (above) successfully launched JUPITER 2/EchoStar XIX in December 2016, with 50 percent more capacity than JUPITER1. EchoStar 19, also known as JUPITER 2, may have a 6 month headstart over ViaSat, and should be in service by March, 2017.

Jupiter-2 and ViaSat-2 represent a huge increase in satellite capacity.

High Throughput Satellites (HTS) provide small satellite terminals such as the Hughes Jupiter and Viasat-2 with more bandwidth. High throughput satellite platforms should provide enough throughput to enable satellite-fed WiFi hotspots for emergency communications, even when the lights go out, as well as delivering a low-cost, high-value service for both visitors and residents.

Today, satellite data caps are the most important metric, not raw speed. If one satellite terminal is shared by four hotspots, a minimum data cap of 50-100 GB/month may be necessary. However, satellite data might be considered a backup. Landlines may be used as the primary feed for this system. Satellites data would only provide back-up. A load balancing router ($85) automatically selects the best internet source according to the load – which would be landline 95% of the time.

Providing both a cable ($100/mo) and satellite connection ($100/mo) to the hub would cost a total of ~ $200/month.

Satellite-fed Hotspots
The Hayden Island Network (HI-NET) uses proven WIFI technology and architecture. It has been used by RV Parks and campgrounds around the country for years. We use basically the same architecture but add a satellite terminal for redundancy. A wireless bridge links the internet/satellite hub to WiFi hotspots up to a mile away.

Since 25 Mbps divided over four hotspots is only 6 Mbps each, users can’t be watching a lot of You Tube or Netflix – without supplementing the satellite hub with fiber or cable. Satellites can now work – in a pinch. It enables emergency communications while delivering a low-cost, high-value service for both visitors and residents, even when the lights go out.

Satellite terminal
The $400 ViaSat Surfbeam 2 satellite terminal (datasheet) uses a 1 meter (3 ft) dish. It draws about 80 watts and feeds two sector antennas (East and West). The similar HughesNet Gen 4 system offers 50-100 GB/month for about $90/month.

The ViaSat SurfBeam 2 Pro Portable Terminal (above) is a much more expensive ($10K) and portable version of the consumer ViaSat Excede terminal you can get on eBay. Satellite terminals provide independence from landlines, deliver a “hot” communications back-up for emergencies, and can provide a “free” amenity for islanders.

Mount a Viasat antenna on a van and you’d have WiFi to go – anywhere. A few 5 GHz links to remote hotspots could provide area-wide broadband.

The total cost of the satellite terminal is budgeted for $600, with $400 for the dish and $200 for installation. It feeds two Ubiquiti backhaul radios with sector antennas. The satellite hub would be run from AC grid power most of the time with a small Honda 2000 Generator providing power through an outage. It’s very quiet and uses about 1 gallon of gas in 8 hours.

WiFi bridges
The satellite terminal hub connects to remote hotspots through point-to-multi-point (sector) antennas. The Ubiquiti-based WiFi network, built by TenGo for the Jantzen Beach RV Park, is a pretty good model. It can be extended to reach most of the island by using larger antennas.

To feed the hotspots, Ubiquiti’s Multi-point hub might use two Rocket5ac Prism radios with two sector antennas on a roof. That should reach several miles when paired with a Ubiquiti’s one-piece Nanostation M5 at the receiving (local hotspot) end. Ubiquiti’s airMAX PtMP Titanium Sectors feature advanced RF isolation and variable beamwidth.

Portland’s Invictus Networks is a national distributor with lots of expertise on Ubiquiti gear. Here’s an in-depth video walking through a Ubiquiti Point to Multi Point setup. A Ubiquiti 5GHz Nanostation ($80) receives the 2,500 ft backhaul. This proposal should probably budget at least $1000 for installation expertise, although that amount is not in the budget. In addition, the Ubiquiti backhaul gear might be dropped in favor of a one-piece Wave-2 MIMO hotspot that combines both local hotspot coverage with an internal backhaul radio to simplify installation.

The Ubiquiti bridge connects the remote hotspots to the central hub. The Ethernet cable from the receiver plugs into the local hotspot, such as an OpenMesh hotspot or Ubiquiti’s Amplifi hotspot.

WiFi hotspots
The freestanding hotspots would cost ~$1500 each (depending on solar and platform options). The free-standing hotspots would include a wireless bridge ($100), the local hotspot ($250), a $450, 19″ Android Tablet (for digital signage) solar panels ($300), charge controller ($50) and a battery ($350).

Outdoor hotspots might resemble a garden sculpture, propane-powered deck heater or street furniture with a solar panel providing protection from the rain. The base houses a 100-200amp/hr deep-cycle battery for power.

Google’s new WiFi hotspots ($129/each or 3 for $299), features Expandable mesh Wi-Fi using 2×2 Wave 2 specs with simultaneous dual-band Wi-Fi (2.4GHz / 5GHz) for 802.11a/b/g/n/ac with TX Beamforming. Their mesh network links the different hotspots together and determines the best path for your data.

Open Mesh access points like the MR 1750 ($250, below) support 50-100 simultaneous users and includes a free license for CloudTrax, Open Mesh’s cloud-based network controller. Open Mesh’s cloud controller, called CloudTrax, manages your network. You can check the status of your access points, see connected users and their usage, and update network settings. No on-site server required. Google’s new WiFi hotspots ($129/each), have similar specs but don’t have the cloud control.

Every Open Mesh access point is automatically mesh enabled. You can add additional units wherever you need them. All access points work together to form a self-organizing, self-healing wireless network with seamless roaming between devices. Their third party software includes Facebook WiFi and MediaShift ads for additional, on-line revenue.

In an emergency, individuals could download videos and other information stored on the hotspot. YouTube Go can save videos locally, on the hotspot SD card. Dozens of hours of news and videos could be ready to go, as well as thousands of downloads from the royalty-free Internet Archive including Old Time Radio and e-books and texts.

Solar Operation
Complete independence from grid power is a challenge. The local access points currently use about 15 watts of power while the satellite terminal uses about 80 watts. A 200 amp/hr deep cycle boat battery ($400) could power a hotspot with backhaul and a large screen for about 24 hours (8 amps x 24 hrs = 192 amp/hrs). A timer could automatically shut the system down from 1 am to 6 am.

Much more solar and battery capacity would be needed for the satellite terminal, which may require four, 300 watt panels to keep the batteries charged. This proposal does NOT provide solar or batteries for the hub. Instead, a small generator provides emergency power.

Hotspots use less power. If our power draw is one 27″ monitor powered by USB Type-C (~30 watts) and one MIMO hotspot (~15 watts), then it totals 45 watts, equivalent to ~ 12 volts x 4 amps (P=IE).

A 12 Volt, 200Ah battery has a capacity of 200amps x 12v = 2.4kWh of electrical energy (P=IE). To get this battery charged in 5 hours you need power output of 2.4kWh/5 = 240 Watts of solar. One 240 watt panel delivers ~ 12 amps. But we’re using only 4 amps most of the time. We can keep the battery topped off all day. Genasun charge controllers have low RF interference, important for this application.

If our 40 watt load runs 14 hours (8am – 10pm), that = 840Wh of the 2.4kWh battery. Our daily load, therefore, should consume only about 1/3 the battery capacity and our solar panels should quickly charge the battery. One, 12V, 200a/h battery should work, but we’ll use two, 12v, 100ah Trojans or gell cells (the power equivalent).

Qualcomm’s Wave 2 MU-MIMO chips are expected to lower power requirements. A newer one-piece Wave2 hotspot may also eliminate the need for a dedicated wireless backhaul device, lowering power and equipment costs.

Manufacturers of the Connected Table (specs) use a 530 Watt Solar array feeding into a 12 volt, 225 Amp-Hour Sealed Gel-Cell battery. They provide four AC electrical receptacles and 8 USB charging ports with a Daily Energy Production of 2.1 kWhrs. It provides power for up to three days during inclement weather and provides shade in the daylight and LED lighting at night. However, the Connected Table costs $17,000, more than the entire budget of this proposal.

Perhaps local universities could design an alternative solution, with an incentive provided by mall operators or through grants (with in-kind matching by local home supply stores or marine businesses). After all, a picnic table with umbrella, the solar panel, charge controller and battery may cost $1000-$2000. You can buy logs cheap on Craig’s List and they could make a very attractive, organic kiosk.

Advertising Revenue
Could “free” WiFi be self-sustaining? Who knows. The goal is to pay for operational costs ($200-$600/month). With 100+ businesses in the Jantzen Beach Mall, it seems like a good bet.

Digital signage may generate the most ad revenue. Google’s Bluetooth Beacons could pull in passersby and make it interactive and useful.

Selling ads on a Hayden Island website could supplement revenue from the big screen. The web site would feature topics like eating, events, destinations and social media, with a goal of 5000 page views/month.

Our goal is to generate $1000/month from 4 kiosks and $500/month from the web site, or $1,500/mo total. It should pay for itself in 12 months.

The system manager gets free advertising space. Fiber would provide the primary internet connection. The satellite would always be available as a backup or to supplement the (wireless) fiber feed.

Community Evaluation
Portland’s Community Centers, Parks and Fire Stations might be provided with free WiFi. Google Fiber could provide plenty of bandwidth to the wireless hub.

As a public agency, PBEM cannot support any activity that might be perceived as endorsing a business. However Neighborhood Associations can, and their endorsement may bring a 501(c)3 tax status, a vital incentive for donations.

A stand-alone 501(c)3 organization might assure continued support through donations and a complementary mix of public service and commercial interests.

City-wide, municipal wireless projects failed big-time 10 years ago. Today, however, Google fiber, MU-MIMO, and insatiable consumer demand are changing the landscape. Portland’s Google Fiber, at $70/mo for 1Gbps, provides enough bandwidth to feed dozens of smaller nodes, each with 50-100 Mbps wireless backhaul. Fiber and newer (5G) wireless technologies are increasingly cost/effective.

Portland’s Fiber-to-the-Premises Feasibility Study and Seattle’s Fiber-to-the-Premises Feasibility Study have more documentation. While MU-MIMO links and fiber hubs won’t enable a solution for everyone, other neighborhoods could learn from Hayden Island’s model.

Since “free”, unlicensed 5GHz spectrum will soon be taken over by Verizon, AT&T and T-Mobile for subscription cellular service, the time to act is now.

Proof of Concept
A proof of concept might utilize the same gear, testing a 5GHz Rocket5 ac Point to Multipoint radio ($120) plugged into a 90 degree sector antenna ($120), and paired with a one-piece 5 GHz Nanostation ($85), plugged into a Ubiquiti Amplifi hotspot ($299). A contractor like Invictus Networks might test the system at 8 different points around the island and produce a report for ~$1000 in the spring of 2017 after ViaSat 2 is operational and price points are established.

Project Summary
A new generation of consumer internet satellites can deliver wireless backhaul to four public hotspots. Solar-powered hotspots eliminate landlines, cellular service and even shore power electricity. Self-sustaining operation could be enabled through advertising. Monthly revenue from advertising is estimated at $200-$1000/month. Additional solar-powered hotspots could be easily added for $2000-$5000 each (depending mostly on Kiosk design) with the local owner and kiosk manager splitting revenue.

Grants or donations totaling approximately $12,000 would likely cover the front-end costs (less labor). Solar panels may be supplied via in-kind contributions by local companies like SolarWorld and SoloPower. Marine companies may help with electrical work. Realistically, another $3K probably should be added to the budget for labor and spare parts, but costs may also be reduced by $3K-$5K through in-kind contributions.

Potential grantors might include Meyer and Murdock foundations, Jantzen Beach Shopping Center, the City of Portland, private moorages and condominiums. It could be operational by summer, 2017.

Longer term, the Obama Administration has announced an Advanced Wireless Research Initiative, through the NSF, which would invest nearly $100 million in four city-scale testing platforms for 5G wireless. The design and building of four city-scale advanced wireless testing platforms should start in FY 2017.

This initiative could upgrade the WiFi Kiosks hubs described here with (5G) 10 Gigabit wireless backhaul (on 28GHz) and enable 100 Mbps end user speeds using the new 3.5GHz band. A 28GHz multi-point link can deliver 600Mb/s with a range up to 12 miles. That could link our satellite/fiber hub to Vancouver or Portland’s Emergency command center in the SE (for redundancy). In Kansas City, Alphabet is testing antennas on street lights. Huawei and ZTE can supply Citizen’s Band Radio Service (3.5GHz) in the US. Intel’s Cloud Radio Access Network, would move cell tower electronics to the data center, lowering cost.

Enabling wireless to compete with fixed wire-line providers like cable and fiber is the ultimate goal. It would likely require delivering 100 Mbps fixed wireless from the Kiosk to the end user for ~ $20/month. A cellular partnership using the Kiosk for delivering 5 GHz LTE might be rejected due to conflicting goals. A revenue model might be more akin to software as a service, similar to Bloomberg, and not based on GB/month or advertising.