Self-healing and SDN: bridging the gap
Achieving high programmability has become an essential aim of network research due to the ever-increasing internet traffic. Software-Defined Network (SDN) is an emerging architecture aimed to address this need. However, maintaining accurate knowledge of the network after a failure is one of the larg...
| Autores: | , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2019 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/179243 |
| Acceso en línea: | https://hdl.handle.net/2117/179243 https://dx.doi.org/10.1016/j.dcan.2019.08.008 |
| Access Level: | acceso abierto |
| Palabra clave: | Software-defined networking (Computer network technology) Software-defined network Autonomic network management Protocol design Fault tolerance Network management Xarxes definides per programari (Tecnologia de xarxes d'ordinadors) Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Telemàtica i xarxes d'ordinadors |
| Sumario: | Achieving high programmability has become an essential aim of network research due to the ever-increasing internet traffic. Software-Defined Network (SDN) is an emerging architecture aimed to address this need. However, maintaining accurate knowledge of the network after a failure is one of the largest challenges in the SDN. Motivated by this reality, this paper focuses on the use of self-healing properties to boost the SDN robustness. This approach, unlike traditional schemes, is not based on proactively configuring multiple (and memory-intensive) backup paths in each switch or performing a reactive and time-consuming routing computation at the controller level. Instead, the control paths are quickly recovered by local switch actions and subsequently optimized by global controller knowledge. Obtained results show that the proposed approach recovers the control topology effectively in terms of time and message load over a wide range of generated networks. Consequently, scalability issues of traditional fault recovery strategies are avoided. |
|---|