High availability ensures that a single point of failure does not cause failure of the total solution. High availability also allows increased capacity to be provided by the addition of more components.
A high availability scenario can be implemented in remote mode using one or more of these ways:
CICS® Transaction Gateway can exploit the high availability, load balancing, and workload management characteristics of CICS Transaction Server for z/OS®. CICS Transaction Gateway also provides integration with the Internet Protocol (IP) workload management functions of z/OS, including IBM z/OS Workload Manager (WLM) and IBM Sysplex Distributor.
TCP/IP load balancing can be achieved using TCP/IP port sharing or the Sysplex Distributor. If TCP/IP load balancing is used, connections are shared between several Gateway daemons listening on a single TCP/IP port and creating a highly available Gateway group. This provides a single end point for client requests, regardless of the type of transaction being handled.
Health reporting is an additional mechanism that can be used to influence the TCP/IP load balancing algorithm so that unavailable CICS servers are not used.
Dynamic server selection enables the Gateway daemon to dynamically select the CICS server at run time for requests from client applications. This provides the ability to avoid a single point of failure. Client applications can be written without prior knowledge of the CICS server names at run time. Two-phase commit XA transactions are eligible for dynamic server selection.