Distributed Name Service (DNS) Management Operations Manual

Distribution Strategy of Network Control Nodes
How DNS Exports the Name Database
5–10 31258 Tandem Computers Incorporated
In this example, although each NCN has a complete copy of the DNS database for all
nodes under its control, an NCN cannot change the information in its database that
was defined on the remote nodes. Only a definition node can change its names.
Operators on an NCN can still control the objects and their names on remote definition
nodes. However, the name manager at the remote node performs the update.
The second best and somewhat less expensive name distribution method is one in
which all names are defined on the NCN with the database completely replicated on
another NCN and partly replicated on the other nodes. In this case, if the NCN
becomes unavailable, operators or applications on the remaining nodes must exist
without updating any definitions, or must use commands to make temporary copies of
names and update those. (This strategy probably requires some synchronization,
outside of DNS, between the remaining NCN and the other nodes.)
The least robust name distribution method is to centralize the database without
replicating any of the names. In this NCN configuration, operators and applications
on other nodes can use the features of DNS only by making remote requests. If the
link is severed, the nodes cannot be controlled autonomously.