Connectivity for high availability
An IBM® z/OS® Connect high availability (HA) environment can take advantage of TCP/IP load balancing technologies.
Both the requests to invoke APIs and services in z/OS Connect Servers, or RESTful API requests from z/OS applications, are made by using the HTTP protocol, which means that an HA environment can use known TCP/IP load balancing technologies to distribute connections from clients across multiple servers.
- Started task WLM classification can be used to classify HTTP traffic in a z/OS Connect Server so that it can be managed by workload management (WLM) along with other work in the system.
- Started task Sysplex Distributor can be used to distribute requests across multiple z/OS Connect Servers in a Parallel Sysplex®, with TCP/IP port sharing within each LPAR.
- Started task TCP/IP port sharing can be used to load balance requests across multiple z/OS Connect Servers in the same z/OS LPAR.
httpEndpoint in the server configuration file of multiple
z/OS Connect Servers to listen on the same shared TCP/IP
port, clients can send requests to a single hostname and port and allow the TCP/IP load balancing
technologies to distribute the request to one of the available servers. z/OS Connect APIs, or API requesters have no
persistent state and do not use HTTP sessions, simplifying both workload distribution and the retry
of failed requests.