Threading model

A multithreaded model provides threads that are used for handling network connections. Threads are also assigned to the requests made by remote clients and the replies received from CICS.

The threading model uses the following objects:

You can set both the initial and maximum sizes of the resource pools for these objects; see Configuring Gateway daemon settings for information on setting configuration parameters.

The following table shows thread limits that must be considered when setting the number of connection manager and worker threads on the various operating systems:
Table 1. Thread limits
Operating system System-wide limit of the maximum number of threads Process limit of the number of threads
AIX® 262,143 32,768
HP-UX No limit (30,000 kernel threads) 30,000 (refer to the max_thread_proc parameter under Configurable Kernel Parameters in the SAM utility)
Linux Equal to the maximum number of processes 1024 (see your Linux Threads documentation for more information)
Solaris No limit No limit
Windows No limit Limited by the amount of virtual memory available for the process (by default a thread has 1 MB of stack meaning that 2028 threads can be created per process)
The threading model is illustrated in the following figure:
Figure 1. CICS Transaction Gateway Threading model for TCP/IP and SSL protocols using a persistent socket
This figure shows the threading model used for TCP and SSL. Two paths are shown from the application to the gateway and back. The application and gateway are on two different computers. The first path starts with an open call from the application which then passes to the gateways protocol handlers, then to the gateway connection managers, from where it returns to the application. The part of the communication path which is between the two machines is labeled handshake. The second path starts with a 'JG.flow(ECIRequest)' from the application, which passes to the gateway's connection managers where 'ready for new work' is labeled as a requirement for continuing. The path then continues to the workers section of the gateway software. Here ECI communication is used to forward data to the CICS server, via the client component of the gateway. The path then is marked wait, while the reply returns over ECI. After the reply is received the path is shown as going straight back to the application. while the path is traveling between the machines on the network it is marked as TCP and SSL.

Information Information

Feedback


Timestamp icon Last updated: Tuesday, 19 November 2013


https://ut-ilnx-r4.hursley.ibm.com/tg_latest/help/topic/com.ibm.cics.tg.doc//ctgunx/ctgthred.html