Hi Bhisham and welcome to Firewall.cx!
The question you raise is a good one - something most of us see in our everyday routine, but usually don't give too much attention to.
In theory, yes, you should get a constant steady 1ms reply when pinging a node in your network, however there are many factors which are not taking in consideration, which usually explains why rarely see that '1ms' every time:
1)
Network utilization. Is the network we are using overloaded? Is there perhaps heavy traffic between another workstation and the one you are trying to ping? Is there someone on the network causing major broadcasts which are replicated to all ports on your network switches, consuming valuable bandwidth?
2)
End device utilization. When you ping someone, your workstation (and therefore operating system), generates an ICMP packet and sends it off to its destination. When it arrives at the destination host, it must also process the ICMP Echo and create an ICMP Echo Reply. While this might not be that much of a deal these days with our superfast computers, if the remote workstation is busy processing other important things e.g the remote user is using a heavy application that is consuming a lot of CPU cycles, it can 'delay' the generation of the ICMP reply packet, therefore giving you a false indication that there are delays in your network. The same applies for a router or any other network device. If its busy - it most probably will delay to respond back.
As you can see, immediately, there are factors which we do not usually take into account when performing these simple tests.
Here on Firewall.cx, you'll find an article we wrote covering network performance, that can provide a valuable insight on how fast your network really is. The article is
Measuring Network Performance: Test Network Throughput, Delay-Latency, Jitter, Transfer Speeds, Packet loss & Reliability
and is very popular.
We are also considering expanding it to help answer questions like the one you've just asked.
I hope this helps answer your question.
Chris.
