Measuring round-trip time with nping
Published: March 21, 2017
Recently, I was debugging a performance issue where a site was spending an above average amount of time running HGETs against a Redis instance. I came upon this snippet of text from Redis’ benchmarking documentation.
Network bandwidth and latency usually have a direct impact on the performance. It is a good practice to use the ping program to quickly check the latency between the client and server hosts is normal before launching the benchmark
https://redis.io/topics/benchmarks#factors-impacting-redis-performance
However when I went to ping the server running Redis I didn’t have much luck…
$ ping -c 10 -W 1 172.24.16.119
PING 172.24.16.119 (172.24.16.119) 56(84) bytes of data.
--- 172.24.16.119 ping statistics ---
10 packets transmitted, 0 received, 100% packet loss, time 9999ms
ping operates by sending ICMP echo packets to a host and measuring the time it takes to get back the response. Unfortunately, many firewalls block ICMP packets, which was the case here.
Fortunately, there are several alternatives which you can use to send TCP packets to your port of chosing. I recommend nping, which is part of Nmap, a security scanning tool.
nping does not output in exactly the same form as ping, but it’s very similar.
NOTE: In order to use nping in TCP mode you need to sudo.
$ sudo nping --tcp -p 6379 -c 10 172.24.16.119
Starting Nping 0.5.51 ( http://nmap.org/nping ) at 2017-03-21 21:13 EDT
SENT (0.0191s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (0.0198s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=885804817 win=0
SENT (1.0193s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (1.0197s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=1708485529 win=0
SENT (2.0208s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (2.0213s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=1795782001 win=0
SENT (3.0225s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (3.0230s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=958728373 win=0
SENT (4.0241s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (4.0247s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=1716784081 win=0
SENT (5.0259s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (5.0264s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=242344193 win=0
SENT (6.0275s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (6.0280s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=915900691 win=0
SENT (7.0292s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (7.0298s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=1326783693 win=0
SENT (8.0309s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (8.0313s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=1502254247 win=0
SENT (9.0325s) TCP 172.24.32.118:13971 > 172.24.16.119:6379 S ttl=64 id=47276 iplen=40 seq=2983128163 win=1480
RCVD (9.0329s) TCP 172.24.16.119:6379 > 172.24.32.118:13971 RA ttl=63 id=0 iplen=40 seq=2102463273 win=0
Max rtt: 0.619ms | Min rtt: 0.330ms | Avg rtt: 0.459ms
Raw packets sent: 10 (400B) | Rcvd: 10 (460B) | Lost: 0 (0.00%)
Tx time: 9.01355s | Tx bytes/s: 44.38 | Tx pkts/s: 1.11
Rx time: 10.01494s | Rx bytes/s: 45.93 | Rx pkts/s: 1.00
Nping done: 1 IP address pinged in 10.05 seconds
As you can see it includes the fundamental piece we’re looking for, the “Avg rtt”.
Conclusion
If you have any questions or comments, feel free to drop a note below, or, as always, you can reach me on Twitter as well.
Hi, I'm Max!