NetSoft 2018 http://netsoft2018.ieee-netsoft.org/program/
The separation of control and data planes promulgated in Software-Defined Networking (SDN) architectures provides new opportunities for sophisticated control over the paths packets take while traversing the network. At the same time the complexity of data representing the network state continues to grow, reflecting complex topological and policy relationships between network elements and their domains. It is critical to the success of SDN to design an abstraction layer capable of storing network topology information and allowing to query it using standardized tools, in a way similar to how traditional databases enabled the abstraction of relational data. This abstraction would cleanly separate policy implementations in SDN controllers from the nuts-and-bolts of handling the topology data. In this paper we investigate multiple approaches to path queries in semantically-rich network descriptions using a variety of mechanisms. We describe a method for storing and querying semantic network graph data using Tarjan path algebras – to our knowledge used for the first time in the context of path-finding in computer networks. We compare the expressiveness and performance of queries that are representative of multi-domain SDN environments.