Using global resource serialization effectively requires an established installation-wide convention for naming data sets, especially those on shared DASD volumes. If your installation does not have a standard convention for naming data sets, you should probably define and implement a standard before trying to use a global resource serialization complex to serialize access to global resources.

Restriction: Some data set naming conventions, such as one that relies heavily on system-dependent information like virtual storage addresses, do not work well with global resource serialization. If your installation has such a standard, you might want to redefine it as part of your long-term planning. A data set naming convention that works well with global resource serialization can avoid both very long RNLs and frequent changes to the RNLs.

In general, a data set naming convention that works well is one where the high-level qualifier or qualifiers have a meaning that proceeds from the general to the specific. For example, assume that an application named ACCOUNTS has three data sets: MASTER, TRANS, and ERRORS. If the data set names are ACCOUNTS.MASTER, ACCOUNTS.TRANS, and ACCOUNTS.ERRORS, you can use the high-level qualifier (ACCOUNTS) to cause all of the ACCOUNTS data sets to be either local resources or global resources.

Data set names that proceed from the general to the specific tend to work well with global resource serialization. Such names make it easy to identify and, if necessary, subdivide large groups of resources with a minimum number of entries in the RNLs. If your installation uses TSO/E, the format of the TSO/E userid is also important because it affects the names of user data sets.