I was observing some ping between a WiFi Client (WC) and and Ethernet Client (EC) . EC is connected via ethernet to the Access Point that WC is associated to .
I ran Pings from EC to WC , the rtt was about 155ms
The ping from WC to EC , the rtt is about 2.3ms.
WC was at good RSSI (~-40) to the AP.
What could be the reason. Yes the ICMP payloads were the same size on the WC and EC.
The WC --- i changed the client from MAC book pro to Galaxy Note same result.
EC ---- Just a Dell Desktop.
Bharat C P
When you say the EC was "connected to" the AP, do you mean a wireless router/AP directly, or through a real network through routers and switches?
The EC is connected directly with the AP.No hops no switch.
Bharat C P
Just guessing that it might be a buffering issue.
Try 50 or so packets each with different lengths. 100 , then 200, 300, 700, 1000 and 1200 bytes.
Then redo the test starting with 1000 and reduce to 100 bytes.
See if what you get is what you expect - i.e. longer packets take proportionally longer time, etc.
I have seen buffering algorithms and non-overlapped I/O cause anomalies like this.
Have you tried using Iperf or jperf? slave on one side and server on the other and then switch directions ?
As a friend of my pointed out to me today, there may just be too many variables here to figure out.
Different processors and, just as importantly, different Software IP stacks are being used at each end.
My company makes several products where the same device over wireless, runs faster than over the Ethernet version. This despite the Ethernet connection being technically faster than 802.11b/g. It's all because of the different software stacks.
Packets sniffed over the air may give you some useful information though.
Comparing ICMP statistics from one platform to another now sounds silly. But based on my many years of programming experience, I would still be interested in the results of the test(s) I mentioned previously.