TCAP Overload Logic

Purpose

The Transactional Capabilities Application Part (TCAP) Overload Logic feature provides overload condition detection of SS7 card resources in the CSP. This feature will prevent the SS7 card from failing under overload conditions by reducing TCAP traffic on the card.

Resources Monitored for Overload Conditions

To provide TCAP overload detection according to traffic busy levels, the following resources will be monitored for overload conditions:

CPU usage

Memory usage

Message Control Blocks (MCBs) usage

TCAP Overload Condition Detection

The TCAP Overload Logic feature provides for four conditions to be detected:

Approaching Busy

Real Busy

Real Busy Clear

Approaching Busy Clear

Each condition listed above will be detected for the CPU usage, memory usage and MCB usage resources. As long as one of the resources is going into a BUSY state, for example APPROACHING BUSY, the system will be considered to be in that corresponding state. When a resource is going into a CLEAR state, for example APPROACHING BUSY CLEAR, the system will be considered to be in that corresponding state.

Call Processing Filter Facility

In order for the TCAP overload logic to maintain a stable state for a period of time and avoid frequent changes among the four conditions to be detected, a filter facility is provided.

The filter facility uses five APPROACHING BUSY levels to indicate the busy conditions. The higher level, the busier the condition. Busy level 0 indicates a not busy condition and busy level 4 indicates the busiest condition.

Every time an APPROCHING BUSY condition is detected, the busy level is increased by one. TCAP will reject any new TCAP Transaction Requests by sending an ABORT message to the network to abort a portion of new TCAP transactions (Begin [ITU/ETSI] or Query [ANSI]) and send a NACK to the host for new transaction requests (TC-Begin/TC-Query) according to the busy level. Based on the increase of the busy level, more traffic will be discarded and NACKED. For example, when the busy level is 1, one of every four calls will be discarded and NACKED, and if the busy level increases to 2, then two of every four calls will be discarded and NACKED. Conversely, a decrease in the busy level will result in more calls being processed by TCAP. When the busy level is zero, all of the TCAP calls will be processed.

TCAP provides a timer interrupt to detect and find the BUSY or CLEAR conditions by monitoring CPU usage, memory usage and MCB usage.

Busy and Clear Conditions

The following describes the process of handling BUSY or CLEAR conditions.

Approaching Busy

When an APPROACHING BUSY condition is detected for incoming new transaction requests, TCAP will send an ABORT message to the remote SS7 node that is running the TCAP on the network. TCAP will abort a portion of the new TCAP transactions (Begin/Query).

For the outgoing new transaction requests, TCAP will send a NACK to the host, also based on the APPROACHING BUSY level and the filter facility busy level. TCAP will also send an ALARM message to the host.

Approaching Busy Clear

When an APPROACHING BUSY CLEAR condition is detected, it is sent to SCAP/TCAP. At this point TCAP will resume call processing for the incoming direction, and ACKnowledge the host’s request for the outgoing direction based on the filter facility busy level. TCAP will also send an ALARM CLEAR message to the host to indicate a busy clear condition.

Real Busy

When TCAP detects a REAL BUSY condition, TCAP will instruct the Message Transfer Port level 3 (MTP3) to discard incoming calls and and free the memory. TCAP will also send a NACK message and an ALARM message to the host. TCAP will resume processing TCAP traffic when the resource busy condition goes away.

Real Busy Clear

When TCAP detects a REAL BUSY CLEAR condition, it will send an ALARM message to the host, and resume the TCAP call processing. TCAP will still send a NACK message to the host for new transaction requests, until the APPROCHING BUSY CLEAR condition is received.

Busy Condition Alarm Levels to Host

There are two different levels of busy alarms that will be sent to the host when a BUSY condition is present or clears.

Card Level Alarm

The card level alarm is sent to the host by TCAP and indicates the SS7 card is busy. The following alarm types are used:

ALARMbrdCARD_APPROACHING_BUSY and ALARMbrdCARD_REAL_BUSY,

Stack and Application Level Alarm

The stack and application level alarm is sent to the host by the SCAP and indicates the SCAP/TCAP stack is busy. The following alarm type is used:

ALARMgenSS7_SIGNALING_STACK_BUSY.

TCAP Active and Standby SS7 Card Redundancy - CSP

The TCAP call processing logic on redundant SS7 cards provides that when a busy condition exists on the active SS7 card, there also exists a busy condition on the standby SS7 card. TCAP calls are first processed in the TCAP of the active SS7 card, then the message will be forwarded to the TCAP of the standby SS7 card for processing. While the TCAP in the active card cuts down the traffic, the TCAP in the standby will also cut down the traffic too.