Predefining CDRSCs (or model CDRSCs) with network specification
involves including NETWORK definition statements in the CDRSC definition
to identify the network in which the resource is located. When a CDRSC
definition follows a NETWORK definition statement, VTAM® recognizes the specified CDRM as the actual
owning SSCP and not as the next SSCP on the session setup path. A
session with that CDRM is required only if it is also the next SSCP
on the session setup path. The CDRSC definition statement can optionally
include the name of the CDRM that owns the logical unit represented
by the CDRSC.
In the following example, which is used to define TERM1 in HOST11,
TERM1 is defined to be in NET3, and its owning CDRM is HOST31.
VBUILD TYPE=CDRSC,…
NETWORK NETID=NET3
TERM1 CDRSC CDRM=HOST31,…
VTAM recognizes that HOST31
is the actual owning SSCP and not the next SSCP on the session setup
path.
Following are some advantages of using this method:
- In a configuration with no name conflicts, this method defines
the shortest session setup path. However, unless you are using NQNMODE=NQNAME,
a resource, even when defined with a NETID, cannot be predefined if
its name is not uniquely known in the host network. If you are using
NQNMODE=NQNAME, a resource, when defined with a NETID, can be predefined
even if the name is duplicated in another network.
This method
of predefining CDRSCs is designed for those systems that want to quickly
locate a resource and use its real name. Therefore, the name must
be uniquely known within the network in which it is defined. If the
name duplicates a name in an interconnected network, it must be using
its network-qualified name.
- By predefining CDRSCs in this way, you can define very specific
adjacent SSCP tables. Because the NETID and, optionally, the destination
SSCP are known, you can define adjacent SSCP tables for a specific
NETID or NETID and SSCP name. This can eliminate the VTAM need for a default SSCP list, which might
not provide the shortest session setup path.
- It adds security by restricting only authorized SSCPs to access
resources in this host, assuming no dynamic CDRSCs are permitted.
- You can use NQNMODE=NQNAME to enable CDRSCs in different networks
to have the same name.
Following are some disadvantages:
- If you specify the VFYOWNER=YES operand on the CDRSC definition
statement, changing CDRMs in a backup situation can be complex. That
is, if you have coded the CDRM operand on the CDRSC definition statement, VTAM assumes that it is the owning
SSCP name. The operator must issue a MODIFY CDRM operator command
in each network that contains a CDRSC definition, with the previous
CDRM listed as the owner. Then (as in Figure 1), if HOST31 fails
and TERM1 is acquired by HOST32 but the CDRM is not updated in HOST11,
the session setup attempt fails because HOST31 is still defined as
the destination of the CDRM owning the logical unit . By specifying
VFYOWNER=NO on the CDRSC definition statements along the route, VTAM does not verify if that CDRM
is the owner but assumes that the session request contains the backup
owner and thus routes the session request to that CDRM. Then, only
the operators at the session origin need to modify the CDRSC owner
to the backup CDRM owner.
- If you are not using NQNMODE=NQNAME, this method does not work
for interconnected networks with duplicate names. The method requires
that all the connected networks have unique resource names. In many
cases, renaming and predefining CDRSCs using NETID or using network-qualified
names is preferable to using the alias name translation facility.
This method is recommended for networks that contain unique names
and require fast session setup.
If the CDRSC moves randomly among networks, specifying VFYOWNER=NO
allows session setup to succeed, but added security is lost in that
the intermediate VTAMs on the session setup path do not verify the
owning host.