Hello, I write again the drawing, because'm wrong .. (sorry if mistook)
Now they are connected as (new sniffer are with this structure)
switch B (1GB) <-> switch A (100Mb) <-> switch C (100Mb) ---> Desktop Z
| |
| -----> Printer
| |
| | -> Some desktop Y
| -> Some desktop and server
Kok-Yong: The printer has no option to change the option "automatic
negotiation" (at least I did not see on the web interface)
Well, after you finish analyzing .. (Count one business day
packages), I can say that in DesktopZ the
"tcp.analysis.retransmission" low at 0.25% (with the original filter
was reached 1.021%), also sniffed two desktop like "DesktopY" and I
had similar values.
One good thing (I think) is that the lost packets generated not
"duplicate ACK" so far (in the first sniffer might be seen AckDup#15 =
40)
During a period of time the original filter showed 6.397% and
"tcp.analysis.retransmission" 0.279%.
I do not know if the improvement is due to a decrease in congestion
or changing wiring.
Something that is happening is the loss of packets between DesktopZ
and Printer, arrives to see 10 packets lost in 3hrs for no longer work
(congestion should have been zero) and 90 packets lost between
DesktopY and Printer in the same period.
For the switch, the switchC is unmanaged, and switchB switchA are
managed, but I have no access (I have to ask permission) to verify
what I recommend is associated with half / full-duplex ports.
I mention this when checking.
Thanks, and good weekend :)
