OSF DCE Administration Guide--Core Components

How CDS Looks Up Names
Node A
?
?
Success!
Client
Clerk
1
Server
Server
Node B
Node C
/.:/Bristol_CH
/.:
/.:/London_CH
/.:/Sales
/.:/Sales
/.:/Sales is in
/.:/London_CH
5
3
4
2
6
/.:/Sales/Spell
LEGEND
= Request path
= Response path
= Child pointer
= Replica
= Object entry
/.:/Sales/Spell ?
As shown in Figure 13-5, the clerk locates the desired object entry by performing the
following steps:
1. On Node A, a spell-checking application requests the network address of the
/.:/Sales/Spell server. The clerk does not have that name in its cache, and the only
clearinghouse it knows about so far is the /.:/Bristol_CH clearinghouse on Node
B.
2. The clerk contacts the server on Node B with the lookup request.
3. The /.:/Bristol_CH clearinghouse does not contain the target object entry, but it
does contain a replica of the root directory. From the /.:/Sales child pointer in the
root, the clerk can learn how to contact clearinghouses that have a replica of the
/.:/Sales directory. The server on Node B returns this data to the clerk, informing it
that a replica of /.:/Sales is in the /.:/London_CH clearinghouse on Node C.
4. The clerk contacts the server on Node C with the lookup request.
5. The /.:/Sales replica in the clearinghouse on Node C contains the /.:/Sales/Spell
object entry, so the server passes the address of the spell-checking server to the
clerk.
124243 Tandem Computers Incorporated 13−7