LinuxCon + ContainerCon + CloudOpen China 2018

The fast growing scale and complexity of data centers brings huge challenges to network analytics. To understand how network operates, huge amount of network metrics needs to be processed timely. Apart from that, new network topology could be built at any time, hence the analytic needs to be nimble to adapt these changes.

This tutorial will introduce a JStorm based solution to analyze network in real-time. An application based on JStorm is built to consume huge amount of network metrics data as stream. To handle the complexity of analysis, an analytic modeling system is built to work with the JStorm application so that dimension or topology could be recognized on the fly. The analyzed stream data is then sent to batch-data-processing system for further analysis such as high-level statistics or OLAP.