Disclaimer
The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.
Liability Disclaimer
In no event shall Author be liable for any damages wha2tsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.
Posting
i'm sure that the machines are running Full Duplex.
If they are not, duplex mismatch generally really, really slows transfer rate.
I thought that the switch could separate completely the data flows.
In theory they can, but in practice much depends on the nature of the data flows and the architecture of the switch. Unfortunately, the 2960 is designed more as a user edge switch than a high performance switch. For example, the 2960, I believe, only has 2 MB of buffer RAM shared by all ports.
But what is not clear to me is how it's possible that heavy traffic between machines A and B can slow down communication between machines C, D and E.
Or, for example, when you write "between machines C, D and E", if C and D are sending to E, concurrently, you can often easily generate egress congestion on the port to E.
Or, if the heavy traffic between A and B is not all ASIC handled, other non-ASIC traffic will be impacted.
To further analyze the situation, you need to look at the switch stats for resource issues.
