Identifying and overcoming monitoring and management challenges in SOA

The very nature of SOA involves implementation of services that are distributed and loosely coupled, and thus monitoring these services is complex due to the involvement of disparate systems that may include external systems and external resources (for example, messaging queues, databases, and so on). Tracing transactions across a loosely coupled implementation, particularly if it involves invocations to external systems, is extremely complicated. The reusable nature of SOA increases the importance of managing availability and performance of these services and greatly increases the need for closed loop governance. In order to achieve the desired Quality of Service (QoS), each service endpoint must literally be managed like a resource. Managed services should have near zero downtime, performance metrics, and a defined service level agreement. In a composite service's infrastructure, it's necessary to monitor and manage the end-to-end view of the systems, as well as provide detailed information about the performance and availability metrics of individual services. Each part of the overall SOA system can appear healthy while individual service transactions can be suffering.

Another important aspect of SOA monitoring is logging. The distributed nature of SOA makes a standardized logging approach difficult to implement. In addition to monitoring services in real time, the administrator is also required to perform standard administrative duties such as backups, code deployments, performance tuning, purging of old data, and more. In general, SOA infrastructure administrators are swamped with the following tasks and activities:

• Managing multi-tier transaction flows

  • Spanning shared components/services

  • Deployed across several tiers in different containers

  • Across the enterprise

• Obtaining performance metrics and visibility into SOA services

  • Beyond generic Java classes and methods

  • Framework and metadata visibility

  • Specific knowledge of the Oracle platform

• Maintaining control over configuration changes

• Performance tuning the service infrastructure

• Performing time consuming administrative tasks

• Code deployments

• Cloning and scale up

• Backups and restores

• Purging and cleanup

• Troubleshooting faults and exceptions

• Policy and security administration

This book is intended to provide you, the Oracle SOA Suite 11g administrator, with a thorough understanding of how to perform each of these tasks and activities.