| Description | This object allows managers to avoid downloading application
directory information when the directory is set to a known
(usually fixed) configuration.
If the value of this object isn't 0.0, it signifies
that the entire contents of the apmAppDirTable,
apmHttpFilterTable, apmUserDefinedAppTable and
protocolDirTable are equal to a known state identified
by the value of this object. If a manager recognizes this
value as identifying a directory configuration it has a local
copy of, it may use this local copy rather than downloading
these tables. Note that it may have downloaded this local copy
(and the ID) from another agent and used this copy for all
other agents that advertised the same ID.
If an agent recognizes that the entire contents of the
apmAppDirTable, apmHttpFilterTable,
apmUserDefinedAppTable and protocolDirTable are equal to
a known state to which an ID has been assigned, it should set
this object to that ID.
In many cases when this feature is used, the application
directory information will be in read-only memory and thus the
tables may not be modified via SNMP requests. In the event
that the tables are writable and a modification is made, the
agent is responsible for setting this object to 0.0 if it
cannot determine that the state is equal to a known state.
An agent is not obligated to recognize and advertise all such
registered states as it may not have knowledge of all states.
Thus, a manager may encounter agents whose DirectoryID value
is 0.0 even though the contents of the directory were equal to
a registered state.
Note that the contents of those tables includes the
protocolDirLocalIndex and appLocalIndex values. In other
words, these values can't be assigned randomly on each agent,
but must be equal to values that are part of the known
state. While it is possible for a manager to download
application directory details using SNMP and to set the
appropriate directoryID, the manager would need to have some
scheme to ensure consistent values of LocalIndex variables
from agent to agent. Such schemes are outside the scope of
this specification.
Application directory registrations are unique within an
administrative domain.
Typically these registrations will be made by an agent
software developer who will set the application directory
tables to a read-only state and assign a DirectoryID to that
state. Thus, all agents running this software would share the
same DirectoryID. As the application directory might change
from one software release to the next, the developer may
register different DirectoryID's for each software release.
A customer could also create a site-wide application directory
configuration and assign a DirectoryID to that configuration
as long as consistent values of LocalIndex variables can be
ensured.
The value of this object must persist across reboots. |