Abstract. In this project, the development of novel path provisioning schemes which will offer maximal reliability to networks affected by correlated failures is proposed. The goal of the proposed work is to develop novel path provisioning schemes, which exploit the principles of Software-defined Networking (SDN) and the multi-layer abstraction of a flow, to offer maximal reliability in networks facing correlated failures, while efficiently utilizing the network resources. Some specific goals of this proposal are: (i) To model correlated failures in SDNs that affect both the data plane (i.e., the network topology) and the control plane (i.e., the SDN controllers); (ii) To develop flow provisioning schemes with maximal reliability and efficient resource utilization in networks affected by correlated failures; (iii) To develop SDN architectures for provisioning flows with maximal reliability in networks affected by correlated failures; and (iv) To implement, on a laboratory prototype, reliable flow-provisioning schemes for SDNs affected by correlated failures.

The major advantage of using SDN to solve the reliable flow provisioning problem is its flexibility. For instance, if SDN flows are assumed to be equal to network paths, then it is expected that SDN-based solutions are equivalent to those solutions achieved by traditional path provisioning schemes. However, SDN flows are a much more powerful concept that offers native support for traffic engineering and quality-of-service (QoS). Thus, these SDN features will be explored to generate flow provisioning methods that simultaneously offer maximal reliability and resource utilization, when networks are affected by the presence of correlated failures. Furthermore, in the proposed work it is expected to implement flow-based provisioning schemes in a laboratory testbed because SDN-based solutions can be rapidly prototyped using real network devices.

Another open research area that will be explored in the proposed work is the mathematical modeling of correlated failures in SDN architectures. Models for correlated failures in telecommunication networks are limited to, and mostly motivated by, geographically correlated failures, while logically correlated failures have been briefly studied, specially, in the case of correlated failures occurring at the data plane of SDNs. This is motivated by the fact that the control plane of SDN-based architectures adds new potential sources of correlated failures to the data plane. Consider, for instance, that a failure or a malfunction affects an SDN controller. Such situation may also inflict correlated failures into one or more network-devices that are managed by the malfunctioning controller. To the best of the writer of this proposal, there are no studies or models about correlated failures inflicted on the data plane as a results of a failure in the SDN controller.

Publications.

• "A Novel Framework for SDN Teaching and Research: The Case of a Chilean University," C. Vega Caicedo, Y. Prieto, Jorge E. Pezoa, Sergio K. Sobarzo, and Nasir Ghani. Submitted to IEEE Communications Magazine, Special Issue on Telecommunication and Network Engineering Education, April 2019.
• "Optimal Multiculture Network Design for Maximizing Reliability in the Face of Massive Correlated Failures," Y. Prieto, N. Boettcher, S. E. Restrepo, and Jorge E. Pezoa. Under review in MDPI Applied Sciences, March 2019.
• "Shared-risk-aware Design for Survivable Migration in SDN Environments," Y. Prieto, C. Vega, J. E. Pezoa, and J. Crichigno. In Proc. IEEE Consumer Communications and Networking Conference (CCNC), January 11--14st, 2019, at the Flamingo Las Vegas Hotel, Las Vegas, NV, USA.
• "Micro failure region models inducing massive correlated failures on network topologies," N. Boettcher, Y. Prieto, and J. E. Pezoa. In Proc. 3rd IFIP Conference on Information Technology in Disaster Risk Reduction (ITDRR-2018), September 20--21st, 2018, at the Poznan University of Technology, Poznan, Poland.
• "Resilient Multiculture Network Design in the Presence of Exploit-Triggered Correlated Failures," N. Boettcher, Y. Prieto, S. E. Restrepo, and Jorge E. Pezoa. IEEE Latin America Transactions, 17 (9), pp. 2336-2344, 2018. Special Issue on New Trends in Electronics.
• "Increasing Network Reliability to Correlated Failures Through Optimal Multiculture Design," Y. Prieto, N. Boettcher, S. K. Sobarzo, and J. E. Pezoa. In Proc. IEEE Chilean Conference on Electrical, Electronics Engineering, and Informatics and Communication Technologies, to be held in Pucón, Chile, 19-20 October, 2017.
• “Reliable Path Provisioning in the Presence of Multiple Correlated Failures,” M. General and J. E. Pezoa. In Proc. of The 37th IEEE Sarnoff Symposium, New Jersey Institute of Technology (NJIT), September 19-21, 2016.
• “QoSApp: Dynamic Bandwith Management for QoS Applications Using OpenFlow,” D. Lártiga, N. Boettcher, Y. Prieto, and J. E. Pezoa. In Proc. of The 37th IEEE Sarnoff Symposium, New Jersey Institute of Technology (NJIT), September 19-21, 2016.

Funding agency: FONDECYT.

Program: FONDECYT Regular 2016.

Grant number: 1160559.

Funding period: April 2016 — March 2019.

PI: Jorge E. Pezoa.