International Business Machines Corporation

Annapureddy S Reddy

Kiet Q Nguyen

Shubha S Tangirala

Thomas Charles Burke

Arvind Srinivasan

Aravind Srinivasan

A method for managing key performance indicators (KPIs) in an enterprise business having a plurality of business units can be provided. The method can include receiving a target for a general KPI and at least one constraint for a general attribute, wherein the general KPI is an aggregation of KPIs and wherein the general attribute is an aggregation of attributes. The method can further include reading a formula for each of the business units, wherein a formula defines how changes in a KPI are correlated with changes in the unit's attribute. The method can further include calculating for each business unit a separate target for the KPI of the business unit such that: a) the general attribute remains within the at least one constraint and b) the target for the general KPI is attained and sending to each business unit a separate target that was calculated.

Management of business process key performance indicators

An anomaly detector compares a first event from a business activity monitoring system to a second number from an independent trusted source, wherein the first event includes a first number. The anomaly detector determines that the first number from the business activity monitoring system is inconsistent with the second number. The anomaly detector generates a calibration relationship between first number and second number based on time, responsive to a determination that the first number is inconsistent with the second number. The anomaly detector stores the calibration relationship. The anomaly detector receives a third number from the business activity monitoring system at an interval after comparing the first number. The anomaly detector displays the third number to a dashboard. The anomaly detector displays a qualifier to the third number.

Business activity monitoring anomaly detection

The present invention discloses a data structure which, given an identifier for a Representational State Transfer (REST) resource, can rapidly yield a configured target and simultaneously yield all configured pattern based rules and constraints for the target. The disclosed data structure is a tree structure including nodes for URL portions. Each node is associated with a hash tree specifically grown in a manner that ensures collision occurrences are non-existent. The tree structure is effectively two or more superimposed trees; one for URL pattern matching to determine a target, another for determining constraints. A single tree traversal, which can be based on a progressive hash, can be used to concurrently determine a target and a set of constraints, which represents improved performance over conventional implementations that require multiple, distinct query/response operations to produce equivalent results.

Growing community of inventors

Morrisville, NC, United States of America

Annapureddy S Reddy