I am troubleshooting a display problem with netTcpBinding in a self-hosted WCF service and it is not clear by looking at the packet That's what's going on. On a new connection I sometimes see an overhead of 200-300 ms in a LAN connection (2 machines connected on the same switch). It is quick to join hand in TCP3 mode, but the next packet PSH ACK has taken 150+ MS. There is no load in the server, this is just a request.
Is this controlled by NetTapPoiding settings, or is it something on the server?
I am not sure that in addition to overheads of 200-300ms while creating a connection What's the problem with your performance?
If the request is for the first time requesting, the spinning of WCDF plumbing will take longer in those cases, while everything is built. I am thinking that this is what you have?
For example, the first request on a new process host, you will see a 'hit' on the first request. Another thing is that if IIS (if your host is) usually after 20 minutes of inactivity, the worker revolves around the process.
I do not think anything in the element will really affect this behavior (I know from the hand).
This application can do some work for that startup (generates cash) which may be an account for that interval if its caching, sometimes it needs to refresh its cache Maybe, or miss a cache and need to get additional data.
Otherwise, most TCP / IP stack tweaks are in the Windows Registry. See t for more information.
HH,
Z
Comments
Post a Comment