Cognitive Packet Network with Software Defined Networks using the Random Neural Network
0 likes400 views
The document discusses the SERCPN project funded by the EU's Horizon 2020 program, focusing on enhancing IoT network security, quality of service (QoS), energy efficiency, and privacy through AI-based routing within a software-defined network (SDN) infrastructure. It highlights the integration of cognitive packet networks and random neural networks to facilitate decision-making for optimal routing. Major advances include a security-aware routing protocol designed to protect critical traffic and devices from untrusted sources while minimizing energy use.
Cognitive Packet Network with Software Defined Networks using the Random Neural Network
- 1. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 SerCPN: QoS-Security-Energy Optimization using RNN-based AI routing with SDN Cognitive Packet Network with Software Defined Networks using the Random Neural Network SerIoT Web Site, 2 February 2021 Piotr Fröhlich, Erol Gelenbe, Mateusz Nowak (IITIS-PAN) ThisTechnology and the related Software Developed by IITIS-PAN is Demonstrated andTested in - SerIoT UseCase 2 by DT Group Horizon 2020, Project No. 780139
- 2. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 Software Product SerCPN Controller and IoT Network based on Smart Cognitive Packets and Random Neural Networks Within a Software Defined Network (SDN) Infrastructure, including a Fog for IoT substrate resulting in the SerCPN SDN-Controller with Online Cognitive QoS-Security-Energy Optimization and Reporting and Dynamic Path Modification (routing) to enhance QoS, Security, and Energy usage and ensuring Privacy policies 2 SerCPN: Aim of this Key SerIoT Innovation Enhance IoT Network Security QoS Energy Efficiency Privacy using AI Based Automatic Traffic Routing that exploits Machine Learning (ML) and Software Defined Networks
- 3. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 SerCPN SerIoT SDN-based CPN network SerIoT SDN-based CPN Software Defined Network • Centralized approach to flow control • Data plane & control plane separated • Convenient for IoT-aware infrastructure Cognitive Packet Network • Self Aware Network concept • Cognitive Packets for recognizing the state of the network • Random Neural Networks used as decision engine Security aware routing for IoT • Improving security level of IoT devices • Security information from • Anomaly detection • Offline analysis Service-based Fog Substrate • Pub/Sub-based ICN approach to Fog Management and Organisation • Flexible micro-service naming and representation • Intrinsic security features
- 4. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 Horizon 2020, Project No. 780139 SerCPN Innovation: Secure QoS Privacy & Energy-Aware Routing Minimize the Overall Routing Goal Function G(f,P) f=Flow, P = Path G(f,P) = α I(f,P) + βQ(f,P) + γE(f,P) + πR(f,P) I = Level of Insecurity, Q = QoS, E = Energy R = Privacy Policy match SDN Controller Analytics SFE
- 5. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 Horizon 2020, Project No. 780139 SerCPN Innovation: Secure QoS Privacy & Energy-Aware Routing SDN Controller Analytics SFE Major Advances of the SerCPN SerIoT Network • Combine SDN with Cognitive Packet Network for effective security aware routing – merging CPN with legacy SDN • Cognitive Packets used for recognition of current network conditions • Uses Random Neural Network as in- controller path decision (routing) engine • Combines Security, QoS, Energy & Privacy as Goals for Routing
- 6. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 Full trust– allow traffic with no limits No trust – mitigation decision (drop packets, send the flow to honeypot, disable node etc.) Limited trust (>0%, <100%) – introduce Security Aware Routing policies – gain time for final decision, network elements/flows are also protected in the meantime SerCPN Security aware routing Horizon 2020, Project No. 780139
- 7. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 Goals and objectives of Security Aware Routing: 1. Protect network devices against (potentially) harmful traffic - untrusted flows are not routed via sensitive nodes 2. Protect critical traffic against threats in untrusted nodes - critical flows are not routed via untrusted nodes Sources of security data: – Agent based cross-layer anomaly detection system (CERTH) – RNN-based DoS attack detector (ICCS) – Lightweight edge traffic anomaly detection (ATOS) – Autopolicy rules violation (IITIS) SerCPN Security Aware Routing Horizon 2020, Project No. 780139
- 8. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 Routing metrics: Security QoS – Cognitive packets – Delay as basic QoS metrics Energy – Traffic-Energy characteristics used – Goal: minimise total Energy usage Privacy – Strong (absolute) privacy rules – Weak (relative) privacy rules – SerIoT looks for balance between Security, QoS, Energy and Privacy SerCPN Multi-criteria routing Horizon 2020, Project No. 780139
- 9. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 SerCPN Network Management Plane – Routing-related components Cognitive Routing Module SerIoT SDN Controller Stats Collector Path Translation & Filtering SerIoT Forwarder SerIoT Forwarder Cognitive Packets Agent Path config (OpenFlow) Cognitive Packets Random Neural Network SerIoT Routing Engine Path config Data for analytics & anomaly detection Smart Packets Routing decisions Monitoring Network State & Configuration QoS Data Security Data Energy Data Privacy data Authentication Authorisation Management (Administration & visualisation)
- 10. This project has received funding from the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 780139 Horizon 2020, Project No. 780139 Major Advances of the SerCPN SerIoT Network • Combines SDN with the Cognitive Packet Network for AI-based effective security aware routing – merging CPN with legacy SDN • Cognitive Packets used for recognition of current network conditions • Uses Random Neural Network as in-controller path decision (routing) engine • Combines Security, QoS, Energy & Privacy in effective SDN based Goal Driven Routing Technical Background • E. Gelenbe, E. Seref, and Z. Xu. Simulation with learning agents. Proc. IEEE, 89 (2):148–157, 2001 • E. Gelenbe, Cognitive Packet Network. U.S. Patent No. US6804201B1, Awarded 2004 • E. Gelenbe. Steps toward self-aware networks. CACM 52 (7), 66-75, 2009 • O. Brun, L. Wang, E. Gelenbe. Big data for autonomic intercontinental overlays. • IEEE J. Sel. Areas on Comms 34 (3), 575-583, 2016 • F. Francois and E. Gelenbe. Towards a cognitive routing engine for software defined networks. 2016 ICC, IEEE doi: 10.1109/ICC.2016.7511138. • E. Gelenbe, J. Domanska, P. Frohlich, M. Nowak, S. Nowak. Self-Aware networks that optimize security, QoS, and Energy. Proc. IEEE, 108 (7), 1150-1167, 2020 • Piotr Fröhlich, Erol Gelenbe, Mateusz P. Nowak: Smart SDN Management of Fog Services GIoTS 2020: 1-6
- 11. Thank you Horizon 2020, Project No. 780139