This is a problem confronting the majority of companies today as they move to massively distributed business services interfacing across diverse technologies. Although few vendors have a complete answer, this manager is in discussion with multiple vendors to evaluate products that address their unique requirements. In testing to date, they have found that, while multiple products can discover the component parts of an application, the “glue that pulls it all together is the hard part.” Currently, for many products that do application discovery and dependency mapping, that “glue” involves a great deal of manual modeling, particularly for custom applications.
Building a model of applications and their dependencies is becoming a focus for many IT organizations, with 33% of those responding to an earlier EMA 2008 survey indicating that “mapping of applications to the underlying infrastructure” and “monitoring of end-to-end application and service execution” are “very high priorities.” Although infrastructure “monitoring and automated detection” remains the priority for 2008 for the majority of respondents surveyed (63%), EMA expects a shift towards automated application dependency mapping over the next several years.
EMA is extensively covering the application management space in multiple practice areas, and recently published the cross-practice paper Managing Applications in 2008: What Does ‘End-to-End’ Really Mean? This paper details EMA’s new Application Management Semantic Model, which describes the types of automation necessary to build a complete, “big picture” view of application execution.