Dependability of Softwarized Networks

• Applicant: Professor Dr.-Ing. Wolfgang Kellerer, since 5/2023
• Subject Area: Electronic Semiconductors, Components and Circuits, Integrated Systems, Sensor Technology, Theoretical Electrical Engineering ; Security and Dependability, Operating-, Communication- and Distributed Systems
• Term: from 2020 to 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 446567116

The actual trend of networks towards softwarization (e.g., Software Defined Networking (SDN), Network Function Virtualization (NFV)) aims at reducing the complexity and specialization of hardware devices, by using commodity hardware combined with more complex software modules. These software modules are often released as open source and are developed by a large and diverse group of software developers. Hence, their quality largely varies which negatively impacts the softwarized network performance.Dependability deals with the trustworthiness of a system by addressing the threats (including hardware and software faults, intended and casual faults), the attributes to assess the dependability (e.g., availability, testability) and means on how to enhance the dependability of the system. This project studies the dependability of different modules of softwarized networks such as SDN controllers. This project targets to deliver a fundamental methodology and guidelines on the development and deployment of dependable i.e. more reliable and secure softwarized networks. In particular, this research project covers the different aspects of dependability applied to softwarized networks: i) threads: identification, characterization and modelling of software threats of different software modules: ii) attributes: implementation of existing attributes and metrics and propose new ones (e.g., software maturity or software aging rate) to evaluate and compare the dependability of different software modules; iii) means: recommend means to potentially increase the dependability of a software module. To obtain our results, we plan to use data mining solutions to characterize threats of different existing modules of softwarized networks publicly available in the repositories. Based on each threat characterization, we will apply existing models (e.g., software reliability growth models, stochastic reward nets, hierarchical models) to analyze the dependability. The dependability will be evaluated by assessing existing metrics and proposing new ones. The suitability of the different options (e.g., models, data mining solution) for each software module will be monitored during the project (according to the threat repositories updates) and finally evaluated with respect to the last available threat repository at the end of the project.

• DFG Programme: Research Grants
• Ehemalige Antragstellerin: Professorin Dr.-Ing. Carmen Mas Machuca, until 5/2023