“Today, computer networks are a life nerve in our society, but they are still managed in the same way they were decades ago when the internet was first established: manually. In case of for instance too much traffic or link failures due to broken cables, changes are required in the network routing to ensure that data packages can take a different route to their destinations. These changes are made manually by technicians, but often this leads to errors such as data packages being lost because of lost connections, or security breaches because data is unintentionally routed along cables that take it outside a company’s firewall, potentially resulting in data leaks,” Jiri Srba explains.
“If we instead automatize these changes on the basis of for instance network analysis, synthesis, modelling and model checking, we can be 100% sure that these issues do not arise, giving us a higher level of security and reliability in the networks,” he adds.
Collaboration of international experts
The project is a collaboration between Professor Jiri Srba, AAU, Professor Kim Guldstrand Larsen, AAU, Professor Stefan Schmid, University of Vienna, and Associate Professor Nate Foster, Cornell University. In addition, the network service provider Nordunet, represented by Josva Kleist and Henrik Thostrup Jensen, is use case provider in the project, offering real-life contexts in which the researchers can perform their analyses.
“We will use the latest techniques within for instance model checking to improve algorithms for automatic network analysis and synthesis. For instance, one case will involve capacity planning, where Nordunet aims to expand a certain network and its capacity, and we will model and analyze the network up front and determine both whether it will have the required capacity and also the potential effects of network errors and how to avoid them by automatic rerouting if they should occur,” Jiri Srba says.
The researchers’ aim is to develop a prototype of a tool that can provide faster, better and more reliable, robust and secure networks that are capable of handling the increasingly complex networks that our digitized society requires.
“The technologies that are used for transmitting data packages are still those that have been used from the very beginning of the internet, because it has been a well-functioning and crucial infrastructure that people have been reluctant to change. However, with the algorithms, model-checking tools, predictive techniques etc. that are available to us today, we believe it is possible to automatize both processes such as capacity planning and verification of network capacity and functionality, and the definition of network rules to ensure optimal capacity and minimal errors. We hope that in the future, our work in this project will lead to a tool that can be commercially used on an industrial scale to ensure robust, reliable and secure networks everywhere,” Jiri Srba finishes.