Operational policies

Intelligent Management uses service level management and policy-driven goals to achieve a healthy and robust environment with a high quality of service. Operational policies consist of service policies and health policies. With service policies, you can differentiate between applications according to their perceived level of importance and target values. With health policies, you can identify conditions to watch for and the product acts on these conditions to ensure a healthy environment.

Service policy

Service policies and, for most kinds of work requests, work classes are used to categorize and prioritize work requests. A service policy consists of a user-defined performance goal and (in some cases) an importance level. Service policies are related to work requests through transaction classes. Each work request belongs to exactly one transaction class, and each transaction class belongs to exactly one service policy. For most kinds of work requests, work classes are used to map incoming requests to transaction classes. Each work class is attached to a Java™ Platform, Enterprise Edition (Java EE) application and a basic request feature; Uniform Resource Identifier (URI) prefix for HTTP, method name for Internet Inter-ORB Protocol. (IIOP), and bus and destination for Java Message Service (JMS). Each work class specifies how the relevant requests are classified into transaction classes. For generic server clusters, and for Session Initiation Protocol (SIP), work classes are not used; instead, the rules for classifying requests to transaction classes are configured on the on demand routers.

A health policy, in contrast to a service policy, is the definition of specific health criteria that you want your product to protect itself against. The health management function uses the defined policy to search the environment for software malfunctions.

Different service policies can have different kinds of goals. The discretionary goal has no associated value or importance. An average response time goal has an associated response time threshold and importance, whereas a response time percentile goal has two associated values; percentage and time, and an importance.

The performance management done by the autonomic request flow manager, the dynamic workflow manager and the application placement controller achieves a defined balance of the performance results. The defined balance among nondiscretionary flows is achieved by either having the flows all under threshold by the same relative amount, expressed as a fraction of the threshold, or by exceeding the threshold by a relative amount that is inversely proportional, 100 - importance. The discretionary flows are given a minimal allocation.

A performance goal strategy requires a monitoring capability by the autonomic manager to determine whether specified performance goals are met, and a reporting capability to notify a provisioning module when changes are required. Furthermore, to account for the case when it is impossible to satisfy all performance goals, it is possible to assign a business value to each performance goal. Administrators must have an in-depth understanding of deployed applications so that they can create realistic performance goals.

Health policy

A health policy works much the same as the service policy, except that the health policy provides a health goal for the environment. A health policy consists of a health condition and a health action. A health condition specifies a problematic scenario in your environment. If this scenario occurs, the specified health action runs to make the condition better. You can specify a health policy to monitor different targets, such as a dynamic cluster or a server. Servers can be simultaneously monitored by multiple health policies. Armed with a set of conditions to look for, Intelligent Management monitors the environment until a problem is detected, and action is taken.