Communication networks form the backbone of our digital society. They are increasingly used for critical services where correct and continuous operation is essential: home health monitoring, active management of power sources on the electrical grid, emergency phone service, and disaster response are just a few examples.
However, there are more and more concerns that today’s networks cannot meet the increasingly stringent dependability requirements. In the last years, network problems have led to major Internet outages in Asia, affected thousands of airline passengers and even disrupted the 911 emergency network. Such outages can lead to considerable losses in revenues and threaten lives.
Interestingly, most of these network problems were caused by human errors, rather than, e.g., physical failures. Arguably, the underlying reason is complexity: operators are typically responsible for a large number of devices whose behaviour is based on many low-level configurations. Translating high-level security policies and performance goals into low-level configurations requires operators to become “masters of complexity”. With the increasing scale of communication networks, the complexity is likely to grow further in the future.
The project Fast and Quantitative What-if Analysis for Dependable Communication Networks (WHATIF) aims to tackle this by designing more automated approaches to manage and operate networks.
The project runs four years from 2020 and comprises members from the Communication Technologies Group at the Faculty of Computer Science, University of Vienna and the Distributed, Embedded and Intelligent Systems group (DEIS) at the Department of Computer Science, Aalborg University.
The main objective of the project is to develop the theoretical foundations for a powerful what-if tool which automatically analyzes whether the current network fulfils its requirements, even under failures, and if not, reconfigures the network accordingly. The tool should provide provable guarantees and at the same time, be as fast as possible.
Also, while the main focus in this project is to develop the theoretical foundations, the partners plan to implement all algorithms in a prototype tool, which will also allow for demonstration to companies.