TABLE OF CONTENTS Chapter 7 detailed the use of multiple types of management agents. This section covers how the management subsystem architecture needs to be modified for multi-board systems.
The underlying premise for the design is that management agents are always located at a central card. Some systems use a separate card for management function, but, for this discussion, we assume that the management function resides on the control card. The organization is similar to Figures 7.2 and 7.3, which show examples of the agents calling low-level routines. Since several of the managed entities could be on the line cards, the agents need to interact with an equivalent management function on the line card for configuration and control.
This management entity is often called a subagent. There is only one subagent for each line card. The management agents on the control card and the subagent on each line card communicate with each other for management operations. To implement this function, the management agent to lower layer interface of Figure 7.2 explodes into a multi-board messaging interface. The common agent on the control card will use the messaging capabilities of the ICCP and communicate with the subagent on the line card.
The subagent acts like an agent for the tasks on the line card. This is effectively a DAL extension since the line card tasks have no visibility into the distribution of the management function. Traps are sent to the subagent for the line card. The subagent, in...
