Chapter EIGHTEEN. ESA Life Cycle Management and Operations

At this stage in the ESA evolution, it's understandable that we spent the bulk of this book focused on the construction and deployment of enterprise services rather than on life in a mature ESA environment still a decade away. But the challenges are known, and the answers are already being formulated.

It's important to understand that both life cycle management and operations are already experiencing a wrenching evolution independent of ESA's appearance on the scene. The overwhelming majority of the tools and practices developed for monitoring operationsi.e., whether the application is running normally, how busy it is, whether it's load balanced, etc.were designed for an IT environment filled with homogenous, standalone entities...the world of SAP R/3 and its fellow monolithic applications. In that environment, the tools for installing, configuring, maintaining, and monitoring applications were built with the assumption that the only applications and systems that would be affected by further configuration, applying patches, or full-fledged upgrades were the ones receiving the changes. Not much thought was given to interdependencies because there weren't many.

But that began to change in recent years, as first application-to-application integration and then the arrival of integration platforms such as SAP NetWeaver, as well as the wild proliferation of hardware stemming from mergers and acquisitions, best-of-breed investments, and systems orphaned by industry consolidation combined to spread application functionality and data across increasingly heterogeneous IT landscapes. Processes began to run across multiple systems, leading to the complex environment we know today, which is filled with unintended interdependencies and is serviced with operations and management tools ill equipped for understanding and working around them. Often the only option under the circumstances is the quick-and-dirty one: flood the zone with engineers to test as many systems in the environment as possible during downtimes and then keep those downtimes to an absolute minimuma state of affairs that can be expensive and doesn't encourage consolidation aimed toward lowering total cost of ownership (TCO).

Now this state of affairs is about to become further complicated by the unprecedented flexibility of enterprise services, which promise to simultaneously decouple applications from systems, decouple the business logic from application functionality, and then granularize it into services and business objects, each one of which may have multiple instantiations supporting composite applications built on top of them. The landscape is poised to explode in potential complexity with these developments as newly independent layers of functionality and business logic are added to the platform, each of which may evolve at a faster rate than the underlying functionality called by the services. Another twist is the boundary-crossing nature of ESA, which will link automated processes within the company to partners, suppliers, and customers, without creating an entirely new set of interdependencies that will require an entirely new approach to monitoring and maintaining them.