Rapidly Troubleshoot with Fabric Analytics

Deep network-centric visibility has been a major customer ask that has been mostly unmet by legacy networking designs -- until now. Big Cloud Fabric™, the industry’s only white box SDN fabric for building leaf/spine (Clos) fabrics – has now integrated the Fabric Analytics module.

Switches within Big Cloud Fabric (up to 16 racks or 38 leaf/spine switches) generate logs and statistics that are collected by the controller.  The Fabric Analytics module processes the data to generate a variety of analytics and trends.  It also provides extensive search and filtering capabilities – powered by open-source ElasticSearch engine. Through a convenient web-based interface of the controller GUI, Fabric Analytics provides centralized visibility for the entire fabric that is extremely easy to use by network operations team.  Also, Fabric Analytics module is architected to consume controller hardware resources opportunistically such that analytics processing does not impact normal Big Cloud Fabric controller operations. Fabric Analytics provides a set of preconfigured dashboards that are divided into two categories: Physical and Logical.

Big Cloud Fabric – industry’s only bare metal SDN fabric for building leaf/spine (Clos) fabrics – has now integrated the Fabric Analytics module in its Controller GUI.  An important operational benefit of software-defined networking is the increased network visibility that is possible with the adoption of an abstracted and centralized control plane. Big Cloud Fabric delivers on this promise with Fabric Analytics.

Figure 1: Big Cloud Fabric - Fabric Analytics Dashboard

With Fabric Analytics, network operations teams can now rapidly troubleshoot network issues and failure conditions throughout the fabric. Fabric-wide configuration changes (via CLI, GUI, or REST API) are visible directly in the controller to identify inappropriate access and/or root cause issues occurring due to network misconfiguration.  Fabric Analytics serves as a platform for advanced analytics and correlations. It dramatically reduces time to issue identification and resolution compared to the traditional physical box-by-box management approach, without the need for expensive third-party analytics tools or repositories.

For more details, check out the Fabric Analytics - At a Glance document.