LogSnap: Creating snapshots of OpenFlow Data Centre Networks for offline querying
No Thumbnail Available
Sreenan, Cormac J.
Institute of Electrical and Electronics Engineers, IEEE
Software-Defined Networking (SDN) has enabled automated modification of the behavior of network devices to match changes in network policy. This facility has driven adoption of SDN in Data Centre Networks (DCNs), particularly multi-tenant DCNs, where network policies are used extensively and can change rapidly as tenants arrive, leave, and modify their resource usage. It is useful for a DCN operator to have a way to query the past state of a network, e.g. for debugging or verification. In a multi-tenant DCN whose behaviour changes frequently under the programmatic control of SDN, this is an important but complex function to provide. While SDN makes the problem more challenging, it also helps to provide the solution - changes in network policy are communicated in packets sent from an SDN controller to the network devices, and those packets are amenable to capture and analysis to reveal the state of the network. Our solution, LogSnap, records messages exchanged over time between an SDN controller and switches in a network, and can quickly recreate the network in an emulated environment for any point in the recorded history. We have evaluated the system for its accuracy, the speed with which it can recreate the network, and quantified the storage implications of speeding up network reproduction.
Software-Defined Networking , Data Centre Networks , Network Management , OpenFlow , Control systems , Protocols , Data centers , History , Network topology , Topology , Tools
Sherwin, J. and Sreenan, C. J. (2019) 'LogSnap: Creating Snapshots of OpenFlow Data Centre Networks for Offline Querying', 10th International Conference on Networks of the Future (NoF), Rome, Italy 1-3 Oct., pp. 66-73. doi: 10.1109/NoF47743.2019.9015187
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.