Adelaide mathematicians model internet traffic

Network designers could benefit from a new Government-backed project that aimed to model the volume, movement and type of data traversing the internet.

Chief investigator Matt Roughan from the University of Adelaide said there was a dearth of real internet traffic data because of network operators' commercial and privacy concerns.

He planned to study traffic data from US telco AT&T and research networks Abilene and GÉANT to discover the mathematical underpinnings of internet traffic.

The team would then use those mathematical theories to produce synthetic traffic matrices that network providers could use to test their designs.

Roughan planned to provide the synthetic matrices to other network researchers who would use the information to develop better design algorithms for optimising network routing protocols.

He likened internet traffic to vehicular traffic on freeway networks where a new road, or a block in an existing road, may affect speed and accessibility.

Network operators used routing protocols and redundant links to remain operational during outages, and were challenged to balance redundancy and efficiency, he said.

The University of Adelaide had worked with AT&T for ten years, Roughan said, claiming to have saved the telco US$12 million in a previous network upgrade.

Commercial network operators tended to be too rushed to develop new design algorithms, he said, while research scientists simply did not have access to enough data.

"I must have seen 1,000 research papers optimising designs of communication networks but this large area of research has been hampered by lack of data," he said.

"The people working on a particular network may have access to their own data. However, even then, when you are designing a network you are designing it for the future, so you need to be able to predict how the traffic will change or grow.

"You need to be able to generate different scenarios and understand the risks in a design, as well as its advantages."

The synthetic traffic matrix project won $425,000 over three years in the latest round of Australian Research Council discovery project funding.

Roughan had no plans to directly commercialise research outcomes, but said findings could improve the university's consulting services to operators like AT&T.

Although the Australian legal framework was "restrictive in terms of what [traffic data] telecommunications providers can provide", he hoped to gain access to local data through research network AARNet.

Customer privacy could be protected by technical safeguards and legal agreements like those between the university and AT&T, he said.