Insight architecture

A typical installation of OnCommand Insight includes data acquisition and data warehousing with reports, all easily accessible from a web-based UI. For more secure environments, acquisition can be done through a remote acquisition unit.

The major components of the Insight architecture are shown in the following diagram:

OnCommand Insight Server
The OnCommand Insight Server contains the main data repository and analysis components. The server is continuously building an end-to-end topology of the environment, analyzing the environment, and generating alerts when an incident or violation is detected.
Acquisition
The Insight collection engine is built on one or more acquisition units. Each Insight server contains a local acquisition unit and can support remote acquisition units. Each unit is a service running on the network that accesses (through modules called data sources) and collects data from devices in the data center. Information collected by the acquisition units is then sent to the server for analysis.

The collection engine is designed to be highly modular and easily patched.

Integration API
An API allows the collection of data from external agents. Integration data can be viewed in the web UI using queries and widgets. Dashboards can contain 'native' Insight data and integration data. You can apply filtering, roll-ups, and grouping to the data in these dashboards.
Ethernet monitor

Insight uses an Ethernet monitoring unit (EMU) to inspect Ethernet NAS traffic flow between the storage, network switches, and clients. This enables administrators to easily identify the heaviest NFSv3 workloads contributing to Internal Volume activities.

Web UI
The HTML5 web-based user interface for Insight enables you to set up data sources and your monitoring environment, including policies, thresholds and alerts. You then use the web UI Asset Dashboard and asset pages to identify and research potential problems. You can create custom dashboards with a variety of widgets, each of which provides extensive flexibility in displaying, analyzing, and charting your data.
Anomaly Detection
Anomalies are performance change events in application infrastructure that do not conform to previously observed and expected patterns. Anomaly detection targets the infrastructure servicing an application, identifying changes in processing patterns and behaviors. These cyclical processing patterns include historical "ebbs and flows" in a workloads performance during hours of business operation and non-business operation.
Data Warehouse

The OnCommand Insight Data Warehouse is a centralized repository that stores data from multiple Insight servers and transforms data into a common, multidimensional data model for querying and analysis.

The OnCommand Insight Data Warehouse enables access to an open database consisting of several data marts that let you generate custom capacity and performance reports such as chargeback reports, trending reports with historical data, consumption analyses, and forecasting reports.

The Data Warehouse consolidates and prepares data for reporting for one or multiple installations of Insight. The data includes history, trending, inventory, chargeback, show back and data presentations to support long-term planning of the data center's infrastructure.
Cognos
Cognos is the reporting engine for Insight, an IBM business intelligence tool that enables you to view pre-defined reports or create custom reports. Insight reporting generates reports from the Data Warehouse data.
Note: If you install Insight on a Linux server, you can only use Cognos if the Data Warehouse is installed on a Windows server. For information about installing the Data Warehouse on Windows, refer to the OnCommand Insight Installation Guide for Microsoft Windows.