Providing Open Architecture High Availability Solutions

Providing Open Architecture High Availability Solutions
43
• View or modify configuration
• Monitor applications
• Enable a system administrator to remotely access and control the system and its components
• Enable a network management system to interface to the system and its components
The key interface areas include the platform interface and the application interface.
The platform interface should provide other parts of the system (i.e., the operating system,
management software or applications) with chassis management functionality including power
supplies, fans and LEDs. The platform interface should also support hot-swap to allow the
insertion, detection, initialization, configuration and removal of system peripherals or cards.
The application interface should provide applications with the ability to allow detection,
registration and monitoring by other parts of the system. It should also enable the applications to
participate in the system’s fault management mechanism.
5.4.2 Objective
To provide a means by which management middleware and other system components can access
state and configuration attributes of the various components within the system, as well as control
these components via operational or configuration methods.
These attributes and methods may be utilized both within the system and external to the system.
5.4.3 Concepts
Health Status Information. Health status information indicates the healthiness of a system
component. It does not include fault information, which is covered in Section 6.0. Health status
information, for example, could be based on CIM [CIM] or the ITU X.731 recommendation
[X.731
] as a model.
State Information. The state information allows for the tracking of the system component’s role
within the redundancy model (i.e., active, standby, spare). State changes should be reported as an
autonomous message on one or more of the management interfaces.
Monitor Information. In general, system components with a monitoring capability should collect
information continuously, even if it is not being polled. The component may also attempt to notify
management interfaces if an exception event occurs.
Remote Monitoring. Health status and state information may be monitored by an external
management system.
Local Monitoring. The system can monitor itself and all of its system components for health status
and state.
Administrative Control. This operation allows a local or remote management entity to invoke
administrative actions on managed system and layer objects.
Fault Injection. Fault injection is the activity by which a failure condition is forced upon a system
component. It is used to verify that the fault detection, isolation, recovery and repair mechanisms
are functioning properly.