Universal software, universal appeal

The vision of global applications is a step closer with the development of tools to create software that can work on any device.

The DEGAS project initially defined the key elements of a software program that are common to all devices, like security, and then separated those device-specific functions.

“Essentially, we wanted to address the problem caused by heterogeneous networks, because currently content and software cannot be used on any device or operating system,” says project coordinator Mr Corrado Priami of the computer and telecommunications department of Italy’s University of Trento.

Heterogeneous networks – ones with a wide variety of different devices like PCs, PDAs and even mobile phones – are already a major part of modern life.

But designing software for such networks is a nightmare. Each device uses a different operating system, and uses different applications, which can’t easily talk to each other. It’s a costly problem creating enormous inefficiencies and DEGAS has gone a long way to solving it.

The project developed a theory to handle heterogeneous networks and produced a set of tools to write software for use on a broad range of devices. Specifically, they developed a mobile adventure game and a mobile commerce solution.

In fact, the project was so successful that industrial partner Motorola immediately began commercial deployment of the mobile e-commerce software it developed, which works on a wide variety of handsets.

The team used a standard protocol, called Universal Modelling Language (UML) to design the key elements of the game and e-commerce software. “The main idea is that when you have to compile an application towards a specific device you can decide at which level of the process you start differentiating the application for a particular architecture,” says Priami.

So, the team developed a core programme that’s the same for all devices. With this approach they simply required compilers to adapt the very low-level intermediate language to individual devices.

One of the team’s coups was to develop most analysis tasks – like performance predictions for instance, or security checks – at the logical or universal level. This simplifies application development enormously. Another coup was the development of formal analysis and validation tools for application development. “We were able to push these tools in the standard development process of applications. That is now for instance used inside Motorola,” says Priami.

This ensures that the software works correctly, making the program more secure and easier to develop. The team also got a very practical demonstration of the power of formal analysis and validation. They were able to find faults with real protocols that are currently in use.

“Project partner IMM found a flaw in version 1.1 of the Single Sign-On Protocol through analysis.” Says Priami. “It’s another guarantee for producing high-quality software. The team developed tools that ensure that there cannot be a breach in security when the software is running in a network if certain conditions are satisfied. What’s more, performance in the final application will not suffer when the customer places constraints on the system.” These constraints are customer personalisation, like allowing his or her device to receive incoming messages, for example.

The team did build one demonstrator, a mobile massive multiplayer online role-playing game – or MMMORG for short. Essentially it’s a mobile adventure game, which many people can play at once. DEGAS’ trick here was to keep the server at the centre of the game very small, with most of the game controls run on the specific device. That way the latency issues that dog wireless games – where slow connections can ruin good games – are overcome, because the games need to send very little information to the server.

“We also developed a peer-to-peer protocol that allowed the devices to talk to each other without having a central control, and this is very good application. In particular, our formal analysis and validation tool were able to refine the communication protocol for security and performance. We made a demonstration of this application. I felt it was very impressive,” says Priami proudly.

Theoretically this approach could be applied to any architecture, or device, running any operating system. “Yes, theoretically, because we make no underlying assumptions about the physical hardware,” says Priami. “This is another of our breakthroughs.”

The team will now continue their work to develop their innovations. Currently the tools are in prototype, they are suitable for academic research but they are not ready to be used in the market.