Back to Archived News
Integrien Files Four Patents for Real-Time Performance Analytics
Tuesday, October 21, 2008
New Breakthroughs Allow Prediction of IT Performance Problems Three Hours Before Traditional Systems Management Tools
Irvine, CA -- Integrien Corporation, the leader in real-time performance analytics, today announced it has filed four technology patents designed to drive the prediction and isolation of performance problems in mission critical applications and business services. These new breakthroughs extend Integrien’s four existing patent filings in the development of a new mathematical foundation for analyzing IT performance data to enable the prediction of performance and availability problems in complex IT systems.
The first filing, "Methods for the Cyclical Pattern Determination of Time Series Data Using a Clustering Approach," presents a new formulation that allows any cyclical pattern in a dataset to be identified. These methods allow events such as month closing, payroll, holidays, etc. to be identified in time series data so they can be treated differently in analysis.
The second patent filing, "Dynamic Thresholding via Weighted Quantile Analysis," presents a new formulation for dynamic threshold determination, allowing extremely accurate "bands of normalcy" to be created for time series data, enhancing signal to noise ratio and thus significantly reducing false positive rates. With reduced false positive rates, IT performance problems can be predicted with much better accuracy.
The third patent filing, "System and Method for Correlating Fingerprints for Automated Intelligence," builds on existing patents related to Integrien Alive’s Problem Fingerprint™ technology, which predicts re-occurrences of previous performance problems. This new algorithm allows the determination and isolation of different reasons for the occurrence of a problem, including the identification of root cause for each.
The fourth patent filing, "Dynamic Problem Determination Using Aggregate Anomaly Analysis," represents a major breakthrough in the prediction of IT performance problems. This new algorithm provides a mechanism for alerting to cascading failures that result in performance and availability problems before they are detectable by traditional monitoring solutions. The algorithm allows a typical prediction time of three hours before existing monitoring solutions could alert to the same performance problem.
"Recently, we’ve begun to see claims from traditional systems management vendors of ‘predictive analytics.’ However, the underlying foundation of these technologies is based on standard statistical methods and techniques," said Dr. Mazda Marvasti, CTO of Integrien. "Through our experience with live customer environments, we have learned that these standard methods are inapplicable to IT data and that a new foundation is required to make sense of such data. These patent filings represent the latest work in the development of this foundation, without which, real-time, predictive analytics for IT performance data is not possible."
Integrien Corporation is the real-time performance analytics company that enables the predictable operation of mission critical applications and services. Integrien helps IT operations to manage complexity by applying real-time analytics to monitoring data to provide the intelligence that will prevent problems from negatively impacting business. Integrien Alive™ allows IT to take a proactive approach to performance management, telling them when an application problem is brewing with a single Smart Alert™ that isolates the impact and likely root cause. As a result, Integrien customers achieve higher quality business operations that are more efficient and continuously available at required service levels. The company has a host of strategic partnerships with leading vendors such as IBM, HP, BMC, VMware and Tata Consultancy Services (TCS), ensuring quick time to value.
Back to Archived News