This IBM Redbooks document is a comprehensive guide to threadsafe concepts and implementation in the context of CICS. In addition to providing detailed instructions for implementing threadsafe in your environment, it describes the real world experiences of users migrating applications to be threadsafe, along with our own experiences. It also presents a discussion of the two most critical aspects of threadsafe, system performance and integrity.
Originally, CICS employed a single TCB to process everything (such as application code, task dispatching, terminal control, file control, and so on) executed on what today is known as the application or Quasi-reentrant (QR) TCB. Over time, CICS added specialized TCBs to help offload management tasks from the overcrowded QR TCB. VSAM subtasking, the VTAM High Performance Option, and asynchronous journaling were all implemented on separate TCBs. Of course, the DB2 and MQ Series attachment facilities also employ TCBs apart from the application TCB. Distributing processing among multiple TCBs in a single CICS address space is not new, but customers and ISVs had little control over which TCB CICS is selected to dispatch a given function.
Beginning with CICS Version 2, all of that has changed. Applications can execute on TCBs apart from the QR TCB. This has positive implications for improving system throughput and for implementing new technologies inside of CICS. Use of the MVS JVM inside CICS and enabling listener tasks written for other platforms to be imported to run under CICS are examples of implementing new technologies.
CICS Transaction Server for z/OS Version 3 Release 2 provides additional functions and enhancements. This updated book covers the latest features, including local and RLS File Control threadsafe commands, threadsafe CICS journaling commands, threadsafe definition for system autoinstalled global user exits (GLUE), and threadsafe WMQ commands. |
