You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I don't think it is really an issue, but I think it will make the application better than before.
I was working with an old device (I think it was created around nineties) that has an VXI-11 TCP server on it. I was trying to communicate with this device using both a Linux(Ubuntu 14,04) and a Windows machine. In the Windows machine, the communication was successful but on the Linux side it was not. In the Linux client, there was a lot of re-transmission and the calls started to become even more delayed because the way Linux deal with TCP protocol. I figured out it using Wireshark and analyzing the packets. After a while breaking my head on it, I figured out a solution.
In Linux we have a TCP feature called TCP_QUICKACK that is used to send out acknowledgements as early as possible than delayed under some protocol level exchanging, and it's not stable/permanent, subsequent TCP transactions. This feature was causing a lot of double ACK in the TCP communication because the device I was using can not deal with this quick ack packets. What I did to temporally fix the problem was disabling the socket TCP_QUICKACK in the socket properties in my application, but I think it would be more elegant and correct if we could add this option in the python-vxi11 library. I am running out of time so I forked the project and I will make a commit with this new option as soon as I get some time. If you guys have some better ideas than mine, I would be very grateful to listen from you. Thank you a lot.
Following is what I did to make my application work without changing anything on the python-vxi11 library. It is important to say that TCP_QUICKACK is restarted every socket send and recv command, so it has to be changed every time we send a command:
I don't think it is really an issue, but I think it will make the application better than before.
I was working with an old device (I think it was created around nineties) that has an VXI-11 TCP server on it. I was trying to communicate with this device using both a Linux(Ubuntu 14,04) and a Windows machine. In the Windows machine, the communication was successful but on the Linux side it was not. In the Linux client, there was a lot of re-transmission and the calls started to become even more delayed because the way Linux deal with TCP protocol. I figured out it using Wireshark and analyzing the packets. After a while breaking my head on it, I figured out a solution.
In Linux we have a TCP feature called TCP_QUICKACK that is used to send out acknowledgements as early as possible than delayed under some protocol level exchanging, and it's not stable/permanent, subsequent TCP transactions. This feature was causing a lot of double ACK in the TCP communication because the device I was using can not deal with this quick ack packets. What I did to temporally fix the problem was disabling the socket TCP_QUICKACK in the socket properties in my application, but I think it would be more elegant and correct if we could add this option in the python-vxi11 library. I am running out of time so I forked the project and I will make a commit with this new option as soon as I get some time. If you guys have some better ideas than mine, I would be very grateful to listen from you. Thank you a lot.
Following is what I did to make my application work without changing anything on the python-vxi11 library. It is important to say that TCP_QUICKACK is restarted every socket send and recv command, so it has to be changed every time we send a command:
The text was updated successfully, but these errors were encountered: