Tweet
Hello,
For grins I did some network analysis on the new software and found some interesting results. Analysis done during market hours (this AM) connecting to several different servers with similar results. CPU load was never more than 6% and all other network traffic was stopped, especially my ridiculously chatty HP network printer. Capture and analysis by Wireshark.
Regardless of 5.1, 6.0 and workspace size, I received "TCP Previous Segment Lost" packets back from the eSignal server... which is usually followed by an exchange of two additional TCP control packets to recover.
With a 7 symbol workspace, 6.0.1.2 and 1 minute of capture, I saw 4 of these errors (even with essentially zero load on my end.) With my 203 symbol workspace, I saw 204 of these errors in the same 60 second period. Believe it or not, I saw similar frequency of these errors with 5.1.
I would expect a packet to get lost in the internet cloud every now and then, but this seems like a really high amount of them. Due to the relatively low load, I don't believe the packet loss is inside my local network.
Another interesting data point is the network load difference, which is staggering between 5.1 and 6.0. The 203 symbol workspace on 6.0 loaded about 40549 bytes/sec (.324 Mbit/sec) onto the network, while 5.1 with a slightly bigger workspace only loaded 9445 bytes/sec (.076 Mbit/sec).
(over 300% increase in network load)
Hope this helps,
lugz
For grins I did some network analysis on the new software and found some interesting results. Analysis done during market hours (this AM) connecting to several different servers with similar results. CPU load was never more than 6% and all other network traffic was stopped, especially my ridiculously chatty HP network printer. Capture and analysis by Wireshark.
Regardless of 5.1, 6.0 and workspace size, I received "TCP Previous Segment Lost" packets back from the eSignal server... which is usually followed by an exchange of two additional TCP control packets to recover.
With a 7 symbol workspace, 6.0.1.2 and 1 minute of capture, I saw 4 of these errors (even with essentially zero load on my end.) With my 203 symbol workspace, I saw 204 of these errors in the same 60 second period. Believe it or not, I saw similar frequency of these errors with 5.1.
I would expect a packet to get lost in the internet cloud every now and then, but this seems like a really high amount of them. Due to the relatively low load, I don't believe the packet loss is inside my local network.
Another interesting data point is the network load difference, which is staggering between 5.1 and 6.0. The 203 symbol workspace on 6.0 loaded about 40549 bytes/sec (.324 Mbit/sec) onto the network, while 5.1 with a slightly bigger workspace only loaded 9445 bytes/sec (.076 Mbit/sec).
(over 300% increase in network load)
Hope this helps,
lugz