Overview of single source IP addresses for software applications

A single source IP address can be configured for the following software applications:

  • RADIUS

  • SFlow

  • SNTP

  • System Logging applications

  • TACACS

  • Telnet

  • TFTP

The above IP-based software applications use a client-server communication model, that is, the client’s source IP address is used for unique client identification. The source IP address is determined by the system and is usually the IP address of the outgoing interface in the routing table. However, routing switches may have multiple routing interfaces due to load balancing or routing redundancy, and outgoing packets can potentially be sent by different paths at different times. This results in different source IP addresses, which creates a client identification problem on the server site. For example, there is no way to designate a fixed IP address for outgoing packets for RADIUS or TACACS, so it is necessary to configure in the RADIUS or TACACS database all possible IP addresses that are configured on the switch as valid clients. When using system logging, it can be difficult to interpret the logging and accounting data on the server site as the same client can be logged with different IP addresses.

To decrease the amount of administrative work involved, a configuration model is provided that allows the selection of an IP address to use as the source address for all outgoing traffic generated by a specified software application on the switch. This allows unique identification of the software application on the server site regardless of which local interface has been used to reach the destination server.