Overview

In general, the TCPResponse interface measures the response time of a particular service by

• sending a connection request to the appropriate TCP port of the machine on which the service resides and

• waiting for the appropriate response message from the server machine.

For example, the interface measures the response time of the HTTP server www.somecompany.com by sending a connection request to TCP port number 80 of the machine named www.somecompany.com. TCPResponse then waits for a connection confirmation message. The time interval between the sending of the connection request and the receipt of the connection confirmation is the response time.

If the interface does not receive a response within a specified time limit, it writes the digital state I/O Timeout. However, TCPResponse does not wait for a service to respond or to time out before it performs the next measurement. For example, the user configures three points so that the interface measures FTP response times to three FTP servers whose IP addresses are, respectively,

• 192.168.100.11

• 192.168.100.12

• 192.168.100.13

PI TCPResponse sends three FTP connection requests, one right after the other. That is, the interface does not wait until the response time measurement for 192.168.100.11 has completed before it sends a connection request to the other two machines. A value of 0 for a point means that the measured response time is less than 1 millisecond.