Taking Mobile Applications into the Cloud
Cloud-enabled mobile computing is at the intersection of two of today's hottest areas in IT, coupling resource-starved mobile phones with the resource-rich cloud.
You want your mobility, in fact you expect it, but you'd still like easy access to critical data-intensive, high-performance computing resources. You'd love it if you could run your research programs from almost anywhere, just by using your mobile phone. You may need to link your phone securely to your institution's high-end computing facilities. Or maybe you'd just like to check on available parking spaces in your corner of campus.
If you've got your cell phone handy, and if someone has written an app that can link your phone to relevant cloud services, you might just get what you want.
Cloud-enabled mobile computing has such potential for higher education that it was highlighted prominently in sessions and discussions at Microsoft Research Faculty Summit 2011, a three-day mid-summer invitational event held each year in Redmond to bring academics, computer scientists, and government representatives together to explore computing trends. At the summit, thought leaders examine existing research projects in key technology areas and envision promising future research directions, from technical, policy, education, and outreach perspectives.
During the event, Victor Bahl, director of Microsoft Research's Mobile Computing Research Center, talked with Campus Technology about Project Hawaii, a featured research initiative that provides students at more than 20 universities with tools and resources to create cloud-enabled mobile applications using Windows Azure and Windows Phone 7. Project Hawaii is one of a series of Microsoft's research projects aimed at leveraging the cloud for mobile computing.
Mary Grush: Can you put the thinking behind Project Hawaii in context for us?
Victor Bahl: Sure. There is a fundamental problem in phones, which is, they are very resource-starved: They don't have enough battery power or CPU; the display is a problem.... And meanwhile, clouds are very resource-rich. If we could couple [the resource-starved phones with the resource-rich clouds] really tightly and take advantage of the ubiquity [of the technologies], we could do some great stuff! So Project Hawaii was created.
Grush: With Project Hawaii you are specifically helping developers--in this case, student developers--leverage cloud services as they write their mobile apps, is that correct?
Bahl: Yes. The thesis was this: There are lots of pieces of work in the computer science domain--natural language processing, object recognition, image recognition, speech synthesis.... How do we bring decades' worth of this research onto the phone?
Now, the kind of apps you see commonly on phones today are mostly the low-hanging fruit--like simple tabulations.
But if you were to create a framework so that people could take advantage [as they build cloud-enabled mobile apps] of all that's happening in computer science,... that would be great.
Grush: When did you start work on Project Hawaii, and what kind of success have you had?
Bahl: Project Hawaii started initially [by brainstorming with two professors] in spring 2010. We decided to create a cloud infrastructure for computer science students to unleash their creative energy on. The universities would create the courses, while we [at Microsoft Research] built the infrastructure, offered the support and the cloud services, and provided some devices. By Spring 2011 we had 22 universities teaching courses around this project.
And the original thesis that the cloud [coupled with mobile apps] can do remarkable things was correct. The students clearly demonstrated that they could build these kinds of apps. It was a massive success.
Grush: What are some examples of apps that the students have created?
Bahl: We list a few on the project page. One that I find interesting is a citizen science project done at Stanford. It is about collecting sensory data that the phones pick up [out in the field] and sending it to a central location in the cloud, which then provides visualization and analysis back to the users, who can see in real time what is going on around them. Another uses the camera of the phone to try to visualize and understand the scene to help the blind by 'talking' it out. There have been many--a variety of apps.
Grush: What types of cloud services are used most in these projects?
Bahl: Many of the apps depend either on lots of computation that needs to get done, or on crowd sourcing where a lot of data is coming from different sources and is being analyzed and sent back.
Grush: So does Project Hawaii aim to offer an array of cloud services--optical character recognition and speech-to-text are a couple you can see listed on the project Web pages--for developers to take advantage of?
Bahl: Yes, [the idea of Project Hawaii is to offer lots of interesting cloud services] that you can pick and choose from as you write your mobile application--you can put them together in a way that makes sense to you.
Grush: What other aspects of mobility might become more important for developers in higher education environments to think about as we start to leverage the cloud with mobile computing?
Bahl: Phones have a lot of sensors, so that the notion of location is very fundamental to the phone, and from that [it's at least possible to] know where the individual is located. You can imagine in a university environment, where there [is the potential for] a lot of openness around research and other things that we do, that location could become a fairly important aspect of mobility--one that [in general] IT people haven't really considered before. But if you are tasked to locate students, professors, and campus resources, precisely, as we might do with a cell phone and GPS, then the question becomes how to design a system that works reliably both indoors and outdoors.
Another thing to consider about smart phones and anything to do with mobility is spectrum. Here, the issue is that as more and more people use phones for their day-to-day tasks, they are going to be requiring more and bandwidth as the applications become richer. But there isn't enough spectrum in the world just to keep adding it--it is a limited resource. So you can already see a trend towards offloading to the campus WiFi network to keep the cell infrastructure from being over-taxed.
There are many additional issues that have to be dealt with, like security and privacy. How do you create authentication mechanisms, security and encryption mechanisms, and so forth?
Moving forward, we have to understand that as people move around, they are no longer limited to the classroom setting, and the whole campus is now a classroom for them. To be able to have that sort of connectivity, the bandwidth, the energy requirements, and importantly, the cloud,... all these issues have to be thought through really carefully.
Grush: How would you characterize the impact of mobility in general, going forward?
Bahl: Students are embracing it in a big way. Corporations are spending billions of dollars around it. It's going to become very pervasive. If you look at the developing nations of the world, they think of it as a way to leapfrog, by going completely mobile. It will infiltrate pretty much every aspect of our lives over the next decade or so.