optimize ADD/DELETE entry for each session/bearer to GTPUMgr on a need basis,
which reduces the signaling toward the GTPUMgr on standby chassis. During
audit, missing entries received from Sessmgr to GTPUMgr will be added to reduce
the audit failures and retain calls. These changes allow a substantial
reduction in standby CPU usage, performance improvements in GTPUMgr, and no
GTPUMgr audit failures.
GTPUMgr and Sessmgr are out of synch during session recovery/ICSR recovery,
which can happen for reasons such as GTP-U remote TEIDs mismatch, GTPUMgr audit
failures are triggered and calls are cleared.
audit failures during session recovery/ICSR recovery have been reduced to
retain the maximum number of calls. Instead of detecting GTPUMgr audit failure
and clearing call, Sessmgr entries are now reinstalled on GTPUMgr once audit
procedure is complete in session recovery/ICSR recovery.
Optimization on Standby Chassis
GTPUMgr was hitting high CPU utilization on standby chassis in
high VoLTE call connection/disconnection scenarios. For every VoLTE bearer
establishment and deletion, RAR is triggered from PCRF. Upon receiving a full
checkpoint on the standby chassis, GTPUMgr signaling for delete and then add
was sent for every bearer.
Call getting clear
by GTPUMgr when duplicate ADD session sent from Session Manager to GTPUMGR for
On standby chassis, signaling to GTPUMgr will now be optimized
during full checkpoint installation so that only the new bearer
addition/deletion related signaling is communicated to GTPUMgr. Signaling to
GTPUMgr on standby chassis for existing bearers, if there's no change in
bearers, shall not be generated on new full checkpoint install request.
reduces standby chassis GTPUMgr CPU usage in high VoLTE call