I'm not a networking guru, (and this is sort of OT for Tomcat) but I'm not convinced by M.Schwartz's response that this is due solely to DNS name resolution. I don't think that would account for such a big difference in time--once the name was resolved to the IP address, behavior should be the same, no?
Saravanan, you say there's a 20x difference but what's the magnitude? Are we talking 10 ms vs. 200 ms. or 1 second vs 20 seconds? By "IP address" do you mean "127.0.0.1" or the IP address of the computer? My pet theory is that the TCP/IP stack is recognizing the IP address as being of the local box and is "shortcutting" packets from transmit queues to receive queues in memory, without actually getting the NIC involved or anything. Or maybe the NIC is doing this. Many years ago when I did hardware testing, we first ran into TCP/IP stacks that had this optimization. The other thought is that for some reason different maximum packet sizes are being used with these different connections? Of course, this is just a theory. Someone with a little more knowledge might be able to speak more authoritatively <grin>. dwh Saravanan Bellan wrote: > Windows 2000 > Tomcat 3.2.3 > > If I try to upload a file using the regular HTML/HTTP, there is a 20x > difference in performance > between using localhost vs the IP address in the Web Server URL. Ofcourse > I'm running the browser > on the same machine where tomcat is installed. Using localhost is 20 times > slower than using IP address. > > Doesnt make a difference if I use Apache/AJP/Tomcat or Direct to Tomcat web > server. > > Can somebody throw some light on this. > > Thanks, > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
