Measuring Latency

When mtping receives a response, it displays the time between when the request was sent and when that response was received.  It also includes a count of how many times the request was repeated.  For example:

host-16: mtping california.mtpip.net ACT_IDSTR california.mtpip.net (98.158.23.124:8080) Response 0 of 208 bytes after 0 repeats in 151ms (Version 2/0): servedat - 1.13.0 July 2012 - DEI, DOC-1.13.0 MTP-linux-3.15.3 455 00000001: ExpeDat-1.13A Data Expedition Inc. - DataExpedition.com, Internal Server Statistics (11300): Name: servedat Version: 1.13.0 Platform: linux Clock: Tue Sep 4 15:41:09 2012 Clock Skew: 0 seconds Delimiter: / Load: 0,2,2 Capacity: 10 Up Time: 5.74 weeks Config Time: 5.74 weeks Idle Time: 17 seconds

The above output indicates a latency of 63 milliseconds.

To continuously monitor latency, use the "-c" option to specify multiple requests:

host-17: mtping -c 3 DataExpedition.com 8080 ACT_PING dataexpedition.com (70.164.67.132:8080) 0 bytes from 70.164.67.132: count=0 repeats=0 time=60 ms 0 bytes from 70.164.67.132: count=1 repeats=0 time=17 ms 0 bytes from 70.164.67.132: count=2 repeats=0 time=15 ms 0 bytes from 70.164.67.132: count=3 repeats=0 time=17 ms --- dataexpedition.com mtping statistics --- 4 packets transmitted, 4 packets received, 0% packet loss round-trip min/avg/max = 15/27.250/60 ms

Latency versus Round Trip Time

The latency measured by mtping may be significantly higher than the network round-trip-time due to the following factors:

• Lost packets, indicated by non-zero "repeats".
• Operating system swap time, which may occur on the first ping after the server has been idle for a long time.
• Server load, indicated by a non-zero "rspn".  See Load Monitoring.
• Routing and packet priority may be different for UDP packets than for ICMP or TCP packets.
• Network delay and packet priority may vary depending on the size of the packets.

For a more accurate assessment of latency, continuously monitor over a significant period of time.