Interview

Sun SPOTs Go Open Source

An interview with Arshan Poursohi

• By Mary Grush
• 02/13/08

Recently Sun Microsystems announced that it would open source both the hardware and operating system for Sun SPOTs, its sensor networking product based on Java technology. CT talked with Arshan Poursohi, a staff researcher in the Sun SPOTs lab on Sun's Menlo Park, CA campus to find out how higher education institutions can incorporate Sun SPOTs in research and instruction.

What is Sun's kit for Sun SPOTs?

It's a development kit for wireless sensor development based on Java. The premise is, there are lots of Java developers out there, and a lot of interesting applications in the wireless sensor network space. And bringing those two together will really ramp up both developers that are doing wireless and embedded research, as well as giving Java developers another avenue of attack for what they're coding.

What are some examples of how Sun SPOTs are already being used?

You can put them into buildings to monitor, for instance, temperature conditions so you can fine tune the air conditioning and heating. You can deploy them on volcanoes and see what the eruption conditions are in a flow from top to bottom -- so you can see what happens as it's beginning to erupt. Robotics is another example. The embedded computer is so small and powerful enough now, that it can do a lot of really diverse tasks. So, if you can dream it up, you can probably attack it with a Sun SPOT.

Sun made recent announcements about open sourcing both hardware and software for Sun SPOTs. Could you tell me a bit about those announcements and what the opportunities might be for research and instruction in higher education?

Here in the labs we've been working with higher education collaborators for ages. And when they can't open the code entirely there's always a little bit of a barrier because there's that bit of stuff that they can't change and must depend on the company for. The announcements were that we're open sourcing the entire stack -- the Sun SPOT libraries, as well as the Squawk JVM, and the hardware. So now everything is open to the developer. University professors and students who are considering basing either curriculum or research around this device now have access to everything.

If you want to make modifications to the Java VM, you are able to do that. If you want to modify the hardware a little bit so that it fits your application better, or if you are just writing an application and you want to see what's going on under the covers, all of that's available now.

So the operating system and hardware are both now open?

Operating system, hardware, everything-open. The whole thing.

Can you explain a little bit about Sun's initiative in higher education with the Sun SPOTs?

Ever since the Sun SPOTs project started, we've noticed, through interns, word of mouth, or professors that we work with regularly, that there's been a great amount of interest in Sun SPOTs. And in the past year and through the remaining part of this fiscal year, we want to build that community up, in the university, K12, and in the professional levels. But we're concentrating on the universities, because that's where a lot of research is done, and where a lot of existing wireless sensor network work is done, along with robotics and things like that. So, we want to build momentum in the community and make it such that it sustains itself even if we're not there. That's the goal.

How many higher education institutions within the US are already working with Sun SPOTs?

Approximately 30 inside the United States.

Could you give me a couple examples of the work they're doing?

One great example is Warren Wilson College, that's working on a typical wireless sensor network deployment measuring values of water quality, soil, and other features of the Panamanian jungle. They started by deploying this at their local college -- they have a river there-so they're doing a local version as well as working with the Cocobolo Nature Reserve in Panama, to deploy Sun SPOTs there. They're recording values back at their campus, so while the sensors are deployed in Panama, they are monitoring them and deploying code to them from the college in North Carolina. Another interesting school is right next door to Warren Wilson: UNC is doing a lot of autonomous robotics research and work with sensor networks as well.

How can other universities get involved?

There's a proposal program now, and you can find out more about that on sunspotworld.com. We're taking proposals from anyone associated with an accredited university, worldwide. We do have a list of places where we are authorized to ship the SPOTs. You submit your proposal by the 15th of the month, and we'll respond with a decision by the 1st of the next month. We are only granting the developer kits, so there are no cash or ancillary equipment grants -- you have to be able to provide the rest of your system yourself.

If you don't want to go through the grant system, we have a great discount for the education market. You can find out about that from your local education sales rep. A single development kit is $300, and you can purchase a classroom pack of 15 kits for $3,500.

Do you have any predictions about what kinds of things could be done in the future with Sun SPOTs?

Yes, we have lots of those! People talk about cell phones, and there are billions of them, with many people getting more than one. There's a limit to the market, but when you talk about embedded computers, it's trillions. And there are embedded computers everywhere -- parking meters, door knobs, light switches, and you've got hundreds of them in your car, if not thousands. So, if you can imagine all of these, plus the recent occurrence in the last decade or so that a lot of these devices are networked. They are not completely isolated any more. So, as you get this huge network of devices talking to each other, you can come up with some really interesting applications.

Fleet tracking, for example, is pretty regular. That's putting a GPS and a cellular modem inside of a truck that's driving around the United States. Now you can imagine, if you had sensors on those devices and they could talk to each other and pass information around, that would become sort of an instant infrastructure. You could see where the trucks are going and how fast they're traveling, but if you added an accelerometer, then you'd know about road conditions such as where all the potholes are. If you added a temperature and humidity sensor, you could get a pretty credible weather map of the US refreshed at a rate that's not even thought about now.

So, all these embedded computers are going to start talking to each other. You'll walk into rooms that already know how you like the temperature, you'll get into a brand new car that you've never been in before and it will know how you like your seat adjusted. We're talking about the Internet of things and machine-to-machine communications.

I imagine Sun is very interested in seeing what other kinds of ideas might come out of colleges and universities.

Yes. It's just amazing. You put some of these devices out there, and every person has different ideas, based on their own interests. And I'll say that Sun is also interested, from a commercial point of view, given all those trillions of devices, to see that all the developers are using Java. That's the case we're making -- that the wireless sensor networking environment is ready for a higher-level language that is easier to use, and that Java is the right choice.