Title

Technical Papers Parallel Session-II: Towards dynamic network policy composition and conflict resolution in software defined networking

Abstract/Description

Software defined networking (SDN) decouples traffic forwarding from network control logic, offering real-time network programmability and innovative service provisioning. Given the continued growth in available SDN applications, individual management of several network services poses scalability challenges and is no longer a viable option. Manually created policy constructs, especially in the presence of multiple application and services may overlap or conflict with each other leading to sub-optimal network operation. The present paper proposes an automated policy creation and validation framework employing user traffic profiling to dynamically capture a basic set of application “network intents”. A policy broker utility is designed to allow further customization of the derived network intents, translated to valid action statements that are dynamically implemented in the network. The profiling scheme and policy broker is tested in two different campus environments, resulting in automated policy creation and conflict resolution for each scenario with relatively low operational latency.

Location

Theatre 2, Aman Tower

Session Theme

Technical Papers Parallel Session-II: Networks & Computer Security

Session Type

Parallel Technical Session

Session Chair

Dr. Jawwad Shamsi

Start Date

30-12-2017 2:00 PM

End Date

30-12-2017 2:20 PM

Share

COinS
 
Dec 30th, 2:00 PM Dec 30th, 2:20 PM

Technical Papers Parallel Session-II: Towards dynamic network policy composition and conflict resolution in software defined networking

Theatre 2, Aman Tower

Software defined networking (SDN) decouples traffic forwarding from network control logic, offering real-time network programmability and innovative service provisioning. Given the continued growth in available SDN applications, individual management of several network services poses scalability challenges and is no longer a viable option. Manually created policy constructs, especially in the presence of multiple application and services may overlap or conflict with each other leading to sub-optimal network operation. The present paper proposes an automated policy creation and validation framework employing user traffic profiling to dynamically capture a basic set of application “network intents”. A policy broker utility is designed to allow further customization of the derived network intents, translated to valid action statements that are dynamically implemented in the network. The profiling scheme and policy broker is tested in two different campus environments, resulting in automated policy creation and conflict resolution for each scenario with relatively low operational latency.