Probes

State probes track the availability of a remote service by actively transmitting network packets to a specified remote host. Tracked Object Manager supports three probe protocol types:

• An ICMP probe that monitors a device, by sending an ICMP ping to the IP address the probe is assigned to.
• A UDP probe that is capable of port service verification, by sending the port a UDP packet and waiting for an ICMP “Port Unreachable” response if the port is down. A UDP probe can also be configured for Application Content Verification (ACV) if the remote server supports a protocol that responds to a UDP packet, such as the UDP Echo protocol.
• A TCP probe that is capable of port service verification, by monitoring the appropriate port for services such as HTTP, Telnet, SMTP, and FTP. A TCP probe can also be configured for ACV for the verification of a layer 7 (OSI model) application running on the server.

The rules and attributes defined by the probe dictate how and when to transmit a packet to a remote host or peer. Unlike tracked objects, a probe does not contain the configuration of the remote entity. Client applications do not register for probes — they create probe sessions. The application provides the IP address and port tuple when creating the session. A probe session, like a tracked object, informs client applications of state events. It is the responsibility of the application to take an action based on the event.

Timing probes gather packet timing measurement for protocol packets. This is a more specialized type of probe, and it does not provide state events. Instead, each request for a probe session provides the client application with the packet transmit and receive times. Currently, ICMP and UDP timing packets are supported. The ICMP timing probes utilize the ICMP echo/reply paradigm. UDP timing probes may be configured to use the UDP-echo or DNS protocol.

The ICMP and UDP probes have the ability to gather packet timing measurements instead of monitoring the state of remote entity. Thus, each request for a probe session provides the client application with the packet transmit and receive times.

Examples of client applications that use probes include:

• Policy Based Routing (PBR), which uses an ICMP probe to monitor a next hop IP address.
• On the S-Series, Server Load Balancing (SLB), which monitors an LSNAT real server IP address using an ICMP ping, or a port using TCP or UDP port verification. SLB also can verify an application running on the real server by configuring a TCP or UDP probe for ACV.
• On the S-Series, Transparent Web Cache Balancing (TWCB), which uses an ICMP probe to monitor a cache IP address, or TCP or UDP probes to perform port verification on the cache server.
• The Virtual Router Redundancy Protocol (VRRP), which uses an ICMP probe to monitor a critical IP interface.
• The IP service level agreement (SLA) application, which uses an ICMP or UDP timing probe session to collect packet timing information. This type of data collection provides the IP SLA application with the statistics it needs to perform its calculations.
  

  Note

  Prior to the S- K- and 7100-Series Firmware Release 7.21, the tracked objects functionality was performed in policy based routing by the route map pinger feature. On the S-Series, probe functionality was performed in SLB and TWCB by fail detection. Both route map pinger and the previous application based fail detection have been removed from the S- K- and 7100-Series firmware and have been replaced by the Tracked Object Manager feature.