Luiza Aguiar didn't overthink where to store and manage data for BlindWays, an iPhone app that helps people who are blind or visually impaired find bus stops.
Aguiar is the director of products at Perkins Solutions, the technology division of Perkins School for the Blind, which released the app in September. She said cloud for mobile applications is a natural fit.
"We made the decision to go to the cloud basically for the same reasons most people do," Aguiar said -- "for ubiquitous data anywhere, anytime for mobile users."
That, and cloud can easily scale to handle data for more than just the nearly 8,000 bus stops of the Massachusetts Bay Transportation Authority, the public transit system that serves the Boston area, including Perkins' Watertown, Mass., campus. The school plans to roll out the app to a second U.S. city in the first half of next year.
The app is an add-on to GPS technology, which helps people get to the general vicinity of a bus stop, typically 30 or so feet away. When users who are blind or visually impaired get there, they open the app and, with the help of verbal directions from a screen-reading program, are guided past landmarks, or "clues," such as a fire hydrant and a park bench to the bus stop sign.
Managing cloud for mobile
Perkins chose Amazon's cloud computing platform Amazon Web Services because "it was one of the more mature and robust offerings for cloud services out there that would give us the pieces we need quickly and at the right price," Aguiar said.
AWS is the top-selling cloud infrastructure service, followed by Microsoft's Azure and Google Cloud Platform.
Raizlabs, the Boston mobile software developer that built the app, recently handed over management of the app to Perkins. One administrator in the school's IT department keeps track of logins and passwords. These are needed to manage the crowdsourcing aspect of the app -- volunteers, many of them sighted, walk the routes to the bus stops, note landmarks and use their mobile phones to get them into the app. "We want to know who has write privileges to our data," Aguiar said.
Perkins also manages two data sources that fuel the app. One is called General Transit Feed Specification, or GTFS, which shapes the way timetables for buses and trains in different cities are formatted and lists the names of the stops and the exact locations. The other is NextBus, which uses GPS information to track buses and predict when they will arrive at particular stops.
All that information, plus landmark clues submitted by contributors, is stored in the cloud, put in the right order and beamed to the app, which is available on the Apple App Store and is free for users and contributors. There have been 1,310 downloads since the launch, Aguiar said.
Getting clued in
Now that the architecture is in place, Aguiar said, adding in data for another city's bus system will be relatively simple. An expansion city hasn't been identified yet, but there are three basic criteria: It needs to use the GTFS format; it has a population of people who are blind or visually impaired; and "there's a motivation for civic engagement in the community for crowdsourcing."
That's important, Aguiar said, because the more clues there are, the more accurate the routes will be. Perkins is right now tapping volunteers to fill out the routes for the Boston area's bus stops. This month, the school has initiatives with corporations like Oracle, Marriott and Wells Fargo to get teams of people to download the app, check the map on Perkins' site for bus stops that don't have clues and then start visiting bus stops and registering clues.
"What they like about doing this volunteer work is you can do it anywhere, anytime on an iPhone," Aguiar said.
Today 3,300 bus stops in the app are outfitted with clues. Aguiar plans on having the thousands of remaining stops filled in by the end of January.
More cloud for mobile
Perkins is looking into how other technologies can be used to help people who are blind or visually impaired, and Aguiar said the public cloud is where she'd like to manage and store data, especially for mobile initiatives. Now the vendor is Amazon, but Perkins could go the popular multicloud route -- with some projects running in a second vendor's cloud and others in even a third's -- if it wanted to use a particular tool.
One example is Google's Cloud Vision API, which lets developers build software that can detect and classify objects and faces in images. Using that in an app could lead Perkins to the Google Cloud Platform. And use of IBM's supercomputer service Watson could warrant a subscription to Big Blue's Bluemix.
"I would love to pilot something like that where we get people close enough with the crowdsourced clues, but then we can add some sort of verification or confirmation with the iPhone camera," Aguiar said. "We're not necessarily there yet, but that could be something worth exploring further as it matures."
Ex-government CIO Bill Oates helps direct innovation at Perkins
Cloud for mobile: Benefits for businesses
Talking SMAC -- social, mobile analytics and cloud