Performance Intelligence

Performance Intelligence (PI) is a method of performance analysis that identifies the most important performance problems in IT systems and their roots causes. PI analyzes millions of granular data points, captured in the Performance Data Warehouse, and identifies which issues have most impact on the IT end user. By examining the data points along multiple dimensions, it identifies trends and correlates events to expose the true conditions impacting IT end users.

PI differs from traditional performance techniques in these ways:

• Wait time vs. System health
• Trends and changes vs. Point-in-time analysis
• Granular problem isolation vs. Broad system indicators
• Multi-dimensional business intelligence techniques vs. static analysis

Wait Time

Typical systems focus on counting server operations, which does not provide a good indicator of IT service. PI calculates wait time associated with each tiny operation, specific to individual users and programs, and uses it to expose delays in end user service.

Trend Analysis

PI captures historical data, presents trends and identifies changing conditions. Traditional analysis focuses on server state at a given point in time, without correlating how the data is changing. PI focuses attention on the changes so that IT staff can be aware of anomalies and react to dynamic conditions.

Granular Data

PI relies on minute, detailed measurements of each step taken by the database in processing each SQL statement. This can result in millions of data points, each representing a delay associated with an individual user request. Common monitoring systems capture broad statistics across the entire database, which is useful for determining “health” but not sufficiently detailed to find root cause.

Business Intelligence Analysis Techniques

PI applies Business Performance Analysis techniques to the database performance analysis domain. Capturing data points in a multi-dimensional “cube” structure, and then “spinning the cube” has become best practice for corporate performance systems. Ignite PI captures the detailed data points in the Performance Data Warehouse (PDW), then applies the same method to correlate and identify trends across SQLs, sessions, users, programs and more.