Network go-slow

THE INTERNET is starting to slow down and clog up as a result of the explosive growth in traffic created by new devices and applications in recent years, but a solution is at hand thanks to a Science Foundation Ireland (SFI)-funded research project, led by a team at Waterford Institute of Technology.

“The internet is no more than 40 years old, yet it is still primarily based on the technology of the 1970s,” explains project leader Dr William Donnelly, director of the Telecommunications Software Systems Group at Waterford Institute of Technology. “You could say that it is now operating well beyond its initial design.”

With the rapid expansion of communications this reliance on relatively old technology has become a serious problem. “The telecommunications solutions that have worked well up to now are just not flexible enough to handle an explosive growth in traffic,” he adds.

“Convergence is adding to that problem because users now expect a high level of service, no matter where they are or how they connect to a network. Telecommunication engineers have helped to create these high expectations because existing telephone services usually have a performance level of 99.99 per cent. Our computers might occasionally crash, but no one expects the phone line to fail.”

According to Donnelly the problem is potentially very serious. “Unless networks can be managed more efficiently, congestion is going to bring everything to a halt. We now have users communicating with users, users communicating with machines, and machines communicating with machines,” says Donnelly. So it’s not just one network, but many. “It’s a heterogeneous environment spanning many different network technologies.”

And the answer is to make the networks themselves deliver the solutions and become self-managing, as they deal with increased congestion. This is where the SFI project comes in. The Federated, Autonomic Management of End-to-end communication services (Fame) Strategic Research Cluster (SRC) will develop solutions that can be applied to build network and service management systems that understand changes in the environment and co-ordinate their actions to reconfigure network resources and services to effectively deliver on an end-to-end basis.

Fame brings together academics and industry specialists and includes collaboration with industrial partners in Ericsson, Cisco, IBM, Teléfonica ID, Alcatel-Lucent and HP and academic partners in UCD, NUI Maynooth, TCD, and UCC.

The project is pushing the barriers of what is technically possible in terms of allowing forms of self-management, allowing some parts of a network, and some services that run on these networks, to “work out for themselves” what is needed to operate efficiently.

Donnelly explains the difference between the solutions the Fame group is working on and the systems in operation at the moment. “In the past, engineers made predictions, and planned accordingly. But this linear approach no longer works because the operating environment has become a lot more complex, and demand is completely unpredictable. It is no longer possible to predict the behaviour of network users.”

He cites the introduction of the iPhone as an example. Several mobile service providers have had problems coping with the increased use of multimedia services generated by these devices with quality of service suffering as a result.

Similarly, when a major event occurs in a venue like the O2. “You can have a huge increase in the number of people in that one area with many of them using their camera phones. This causes a huge increase in traffic on the network and you need to be able to decide how you can deliver a proper quality of service in those circumstances. This might be achieved by pushing some of the traffic onto local wi-fi networks while keeping priority users on the 3G network.”

This smart use and sharing of resources is what the Fame project is all about. “The network not only has be able to decide how to configure its resources, it has to be able to discover new resources when necessary,” Donnelly points out. “What we need now is a more dynamic communications framework. If the next generation of networks is to work, it will have to be self-governing, self-healing, self-optimising and self-protecting.”

He compares this to what happens in nature. “During exercise, we convert glucose into energy. If the exercise becomes more intense and the glucose supply becomes inadequate, the body keeps going by burning up fat. The fat is like the spare capacity in the telecommunication system, and instead of being localised, that capacity is general. If we look at current management we usually find that there is just one fall-back position, and this has been fixed by predictions.”

The Fame project is working on replacing this simplistic predictive approach with a more sophisticated system with the automatic capacity to balance real-time needs against real-time capacity. This capacity can then be fine-tuned to distinguish between different levels of demand. It can therefore allocate extra resources to applications like video-conferencing which uses up a lot of bandwidth while offering less to voice or data at times of the day when there is less demand for these.

Achieving this level of response, according to Donnelly, is not just a matter of getting more efficiency. Autonomic management will also keep costs down and this has become a huge issue as service providers struggle to deliver more and more with the same resources. At present the service providers buy enough bandwidth to satisfy expected demand. Having to buy in extra capacity can eat into profits and hurt the bottom line. However, if they can better balance demand, they can squeeze more out of the existing system and cut costs significantly in the process. But this can’t be done simply by making existing systems more efficient or by creating new bandwidth resources.

“Traffic is being driven by user behaviour, and we’re getting patterns we have never had to deal with before, such as virtual communities, geographically separate, but linked by telecommunications.”

And these issues are being tackled by Fame. It is hoped that the solutions created will enable the creation and growth of an Irish-based international communications service management cluster that will grow the considerable investment already made in this sector by multinational firms such as the Fame partners as well as indigenous communications management companies.

Ultimately, this cluster will have the potential to help position Ireland as a global leader in communications management related research, product development and professional services.