Cross-Architecture Access: Overview

Cross-architecture access is a feature of SAS/SHARE software that enables a SAS/SHARE server session and its client sessions to execute on machines that have different architectures. For example, a server and its clients can execute on machines that have different internal representations of data such as the IBM System/390 and Intel Pentium, or Digital Equipment Alpha VMS and Hewlett-Packard Precision Architecture.
Cross-architecture access enables you to move data or applications from one type of operating environment to another. For example, a UNIX application that uses a SAS library on z/OS issues a LIBNAME statement in the same way that a z/OS application does—that is, by specifying the z/OS physical name for the SAS library and the name of the z/OS server. Using cross-architecture access, you can do the following:
  • move a SAS application from z/OS to UNIX, leave the data on z/OS, and continue to access the data without changing the application
  • move an application's data from UNIX to z/OS, leave the application on UNIX, and change only the LIBNAME statement that accesses the data
  • duplicate the application on both operating environments and simultaneously access the data on z/OS
Cross-architecture access enables users to read and write SAS data across architecture boundaries. It enables applications that run in one type of operating environment to read DBMS data that is accessed through server-managed SAS/ACCESS views when that DBMS is available only under another operating environment. For example, a SAS session on a Sun workstation can use a SAS/ACCESS view to read the contents of a DB2 table on a machine that runs the z/OS operating environment. For more information about using views under SAS/SHARE, see SAS Data View Programming Considerations.
SAS/SHARE software is especially well-suited to the following types of applications:
  • those that require access to a single record at a time
  • those that use a WHERE clause to subset large data sets
  • those that execute procedures against small data sets
An application that processes large quantities of data, especially through multiple passes, might benefit from moving a copy of the data to the computer on which it executes, or from using SAS/CONNECT software to remotely execute SAS on the computer on which the data is stored.
SAS/SHARE 9.1 supports access to some other types of SAS files, such as SAS catalogs, when the architecture of the server machine differs from the architecture of the client machine. For details, see Cross-Architectural Differences and Cross-Architecture Restrictions and Limitations.