The GETBULK operation merely requests a number of GETNEXT responses to be returned in a single packet rather than having to issue multiple GETNEXTs to retrieve all the data that is needed. This is generally more efficient with network bandwidth and also allows an agent to optimize how it retrieves the data from the MIB instrumentation. However, there is also the possibility of an overrun, which means to get back more data than was needed because parts of the MIB tree that wasn't required were returned as well.
The GETBULK operation takes two parameters:
- the number of objects that are only expected to return a single GETNEXT instance, not multiple instances. Managers frequently request the value of sysUpTime and only want that instance plus a list of other objects.
- the number of objects that should be returned for all the repeating OIDs. Agent's must truncate the list to something shorter if it won't fit within the max-message size supported by the command generator or the agent.