During periods of relatively low network activity, configurations with many EE connections can consume a substantial amount of CPU resources because of the EE LDLC keep-alive function. In addition to the overhead of the timer maintenance itself, the I/O cost of sending keepalive messages (TEST request) and the processing of the subsequent responses has a significant impact on CPU utilization when large numbers of EE connections are active.

The Enterprise Extender LDLC keep-alive reduction function assists in reducing the total CPU impact of Enterprise Extender LDLC keep-alive processing overhead. You can specify an initial and a maximum value (through the max_value parameter) for the logical data link control (LDLC) liveness timer interval, using the LIVTIME operand (see LIVTIME operand in z/OS Communications Server: SNA Resource Definition Reference for more information). In this way, each Enterprise Extender connection can maintain its own liveness timer value depending on network conditions. The liveness timer specifies how long the LDLC layer will wait without receipt of a network layer packet (NLP) from the connection partner before sending a liveness message (TEST frame) to verify that the connection is still operational. During periods of inactivity, an EE connection can expand its current liveness timer up to the limit specified by the max_value parameter. During each liveness interval, if no HPR data is received from the EE partner, the current liveness interval doubles in size. This process is repeated until the maximum liveness timer value is reached, or HPR traffic resumes. If the maximum liveness timer value is reached, the liveness timer will remain at that value. When HPR traffic resumes, the current liveness window is reset to the initial value and the EE connection is immediately tested to verify partner reachability.

The amount of CPU savings realized by LDLC keep-alive reduction is system-dependent. When HPR over EE configurations specify DISCNT=NO, significant CPU savings are possible.