Improved Routing for the In-Band Network Telemetry Problem

The continuous growth of networks imposes on managers to find new technologies for better handling of traffic measurement for network wide-visibility (i.e., the ability to access most desired flow-level information per hop for all flows), performance monitoring requirements and network robustness (i.e., the ability to survive partial network failures). In-band network telemetry has emerged as one of the latest network monitoring paradigms for the mega-scale network. Previous research has focused on orchestrating the processes of collecting in-band network telemetry data from the network and addressed the scalability. Yet little has been done to route network flows efficiently to address the low network coverage and monitoring applications performance in this new network monitoring paradigm. This paper analyses two mathematical models for in-band network telemetry. The first model, proposed by [1], exploits shortest path flow routes as inputs. The second model proposed in this work, extends the first to compute flow route paths efficiently, providing a comprehensive solution to the problem.