The Monitor program's job is collecting requested information from client programs, storing it, and returning it to any interested party. There are several steps to this process...
The client gives the server three pieces of information:
- The client's type (SBC, Node, RM, Supervisor ...)
- The client's computer's name (d0sbc001b, d0lxmast, ...)
- It opens a TCP/IP connection (using ACE)
When the server gets a query, it does a few things:
- It decides what information it already knows and which of that info is still "up-to-date" as defined by a timeout for each data element. This information is filled into the response immediately, without ever talking to the client.
- It takes all the questions that it does need to ask each client and groups them together into one big question. (Thus a maximum of one TCP/IP call is made to each client per display query.)
- It goes out and queries each client that it needs to in order to retrieve all requested information.
- It caches these responses in case some other display asks for the same information in the near future
- It packages up all the information into one response and sends it back to the display.
Go to monitor displays page to get tools for viewing monitor data and instructions for using them.