In the area of network visibility, having high-resolution metrics is useful to unveil patterns and behaviors that would otherwise be "averaged out" in smoother, lower-resolution signals. Visibility tools that have access to packets are, in theory, able to produce metrics up to the packet-by-packet resolution, that is, the best one could hope for. Nevertheless, high-resolution metrics are particularly demanding in terms of storage and this has actually posed a practical upper-limit on the metrics resolution, that rarely offer sub-minute samplings. Fortunately, thanks to the tremendous evolution of big-data stores, it is now possible to rethink network visibility solutions for the generation and storage of high-resolution network metrics. This talk discuss the challenges behind the generation and storage of high-resolution metrics and demonstrates how opensource software ntopng, InfluxDB, and Grafana can be used together to build an effective high-resolution network visibility solution. Intended audience is technical and managerial individuals who are familiar with network visibility.
Speakers: Simone Mainardi