Chapter 2. The Resolver 31
Advantages
When using a multiple stack environment, setting up a global environment with a minimum
set of common statements has the advantage of preserving the entire image from being
changed inadvertently. It has also the advantage of creating a default environment that could
be used for those who are not planning to create their own Resolver definitions.
Considerations
In a multiple stack environment, it is important to review how the applications will be
implemented. Some of these applications might be implemented to open their socket with a
specific bind and to do this they will need to code TCPIPJOBNAME to direct them to the
different stacks. Therefore, when designing the global definitions in the Resolver address
space, do not code a TCPIPJOBNAME in GLOBALTCPIPDATA, but allow it to be coded in
local TCPIP.DATA.
2.3.3 Recommendations
To implement the Resolver address space, it is important to first determine whether your
environment will require a single TCP/IP stack or multiple TCP/IP stacks.
In both cases the Resolver is an independent address space and has to be up and running
before the TCP/IP stack is started. However, when implementing for a multiple stack
environment, you must determine how the global files will be defined and which global
statements will be used. This will depend on how independent each stack will be within the
z/OS LPAR. The recommendation is to create a global TCPIP.DATA data set for a single
stack environment, and a default TCPIP.DATA data set if a multiple stack environment is
required.
In the next session we will configure the Resolver address space to create the global and
default definitions to provide the same environment for all applications, which do not have
specific needs. This will allow specific local statements to be defined at the stack level or
application level. To be able to do that, each stack (if more than one is desired) or each
application that needs to have its own resolver statements, has to define to its own
TCPIP.DATA data set for local statements.
It is also recommended to start the Resolver address space during the initialization of the
UNIX System Services via the RESOLVER_PROC() statement within BPXPRMxx, as we will
show in the next section.
2.4 How the Resolver address space is implemented
Based upon our recommendations, we will show the implementation details for the Resolver
address space.
2.4.1 Implementing the Resolver address space
In this section we are going to show the necessary tasks to implement and configure the
Resolver address space in our test environment. Figure 2-10 on page 32 shows the data sets
that are need to configure the Resolver address space.