Troubleshooting Admission Control

This topic discusses class of service (CoS) and call admission control (CAC) serviceability.

To troubleshoot call admission control, you should understand the classifier policy profiles configured on your system. A classifier policy is the configuration that maps QCI (4G) and TC/THP (3G) to internal forwarding queues and defines packet loss priority. You can have multiple classifier policy profiles on your system. Therefore, understanding how these multiple classifiers interact with your system and with each other is key to understanding what to look for when you have problems with admission control.

To understand CoS, you must understand the CoS policy. This policy is the configuration that manges quality of service (QoS) parameters. You can have multiple CoS policies on your system.

CoS and CAC serviceability also depends on two other configurations:

• Resource threshold policy which controls your system for CAC. You can have multiple resource threshold policies configured on your system.
• The bandwidth pool, which allocates bandwidth sharing among APNs and the gateway. You can have multiple bandwidth pools configured on your system.

Finally, you need to know about local policies. A local policy is a collection of a classifier profile, a CoS policy profile, a resource threshold policy profile, and a bandwidth pool. A local policy is so termed because it is attached to the gateway or to individual APNs.

You can troubleshoot class of service and call admission control by examining:

• Total system bandwidth and per APN bandwidth can be configured with percentage allocations to each QCI/Traffic-Class.
• System ensures each QCI gets allocated system bandwidth optimally
• Maximum-bearers configuration for the gateway
• High or low threshold percentages for CPU, memory, system load, or maximum bearers with local policy to apply when a threshold is reached
• Forwarding-class or loss-priority definition per QCI or traffic class
• Local policy to cap maximum GBR, MBR, and AMBR values per APN

Use the following commands to troubleshoot this environment:

• For subscribers, use the command:
  user@host > show unified-edge ggsn-pgw subscribers extensive
• For preemption lists (priority levels), use the command:
  user@host > show unified-edge ggsn-pgw status preemption-list detail

1. Check the session status to determine whether it is a visitor, roaming, or home session.
2. Look up the local policy being applied to the APN.
3. Match this local policy with its classifier profile, the CoS policy, and the bandwidth pool

1. Identify rejected calls by entering:
   user@host > show unified-edge ggsn-pgw qos statistics apn apn-namel

   Counters such as “No resources”, “Service denied”, “Authentication Fail”, “APN access denied” indicate rejected calls, but not necessarily by CAC.

2. To verify the cause for rejected calls, look in the Routing Engine stats section:
   Active BearersCPU Load (%)Memory Load (%)

   These counters can indicate that the system is running out of resources.

3. To verify that resource exhaustion is the source of the problem, enter these commands:
   user@host > show unified-edge rmps table gateway-bearersuser@host > show unified-edge rmps table apn-bearersuser@host > show unified-edge rmps table anchor-pfe-bandwidthuser@host > show unified-edge rmps table bandwidth-pools