Last year, a group of researchers built a different programming framework, known as Haggle, for pocket-switched networks. The Haggle library adds collections of code to support manipulating data on pocket-switched networks using a variety of platforms, including Windows and Windows Mobile, Mac OS X and iPhone, Google’s Android, and Linux.
The difference between Haggle and D3N is whether the intelligence–the knowledge of how to interact with pocket-switched networks–is inherent to the language or in a separate code library. D3N builds knowledge about the way pocket-switched networks work into the programming language. This makes programming for pocket-switched networks simpler. Programs written in D3N can, for example, grab data from the network with a simple command. Developers working with Haggle can still grab that data, but the programming is more complicated.
D3N code can also be tested more easily, an important feature if service providers are to be convinced to allow the programs on their phones. “[Mobile providers] want to make sure that the software that gets put on their phones is very, very reliable,” Crowcroft says.
“So you need some alternative” to ease the creation of applications for pocket-switched networks, says Intel’s Fall. “We have done it with [the Haggle] programming interface. They have taken it further and have written a language for development.”
The embryonic D3N language, and pocket-switched networks in general, still have unaddressed issues. For example, one ill-studied topic is security on pocket-switched networks. Passing data from phone to phone could ease the spread of malicious software. In addition, attackers could flood the network with digital pollution such as spam.
“Pollution in a peer-to-peer system is a really hard problem [to solve],” Crowcroft says. “I hope it is not a deal breaker.”
Getting people to actually use pocket-switched networks could also be a problem, as both wireless connectivity and Bluetooth drain power. Passing along data for other users could dramatically shorten battery life. The research group is working on a hybrid communications model that uses low-power Bluetooth to search for interesting data among the network nodes and then uses the more power-intensive wireless networking system built into the phone to transfer the data.
“We don’t care what type of radio forwards the data,” Crowcroft says. “Battery life is going to be a far more important aspect of this.”
If the problems are solved, and cell-phone makers convinced that such additional functionality is useful and secure, such networks could give users an alternative high-bandwidth way to communicate.