A Metric Definition, Computation, and Reporting Model for Business Operation Analysis

摘要

This paper presents a platform, called Business Cockpit, that allows users to define, compute, monitor, and analyze business and IT metrics on business activities. The problem with existing approaches to metric definition and computation is that they require a very significant development and maintenance effort. The cockpit overcomes this problem by providing users with a set of abstractions used to model the problem space, as well as development and runtime environments that support these abstractions. The cockpit is based on three conceptual models: the business domain model defines the business data to be analyzed, the metric model defines the business metrics of interest for the user, and the reporting model defines how metrics should be aggregated and presented in the reports. The proposed approach provides the following key benefits: ⅰ) it allows the definition of many different reports without writing code; ⅱ) it reduces metric computation times; ⅲ) it enables the definition of different ways of computing a metric based on the characteristic of the object being measured; ⅳ) all the code of the cockpit is independent of the business domain to be managed. As such, it can be applied to many scenarios. Domain independence, however, is not achieved at the expense of complexity in the configuration: to apply the cockpit to a given domain, users are simply required to provide an abstract description of the part of their data model that is useful for business operation analysis purposes. The cockpit, and the features described above, have been developed and refined over the past few years. Our research started in the context of business processes, and we have then applied the same concepts to other domains, such as inter-bank transactions.