Is there a reason why this support request was assigned to
davepitcher? Also, I don't see a .cap file attached (though
I'm not sure what a .cap file is).
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I figured I'd chime in here. This is pretty close to what I
would expect. I have, in general, found that high-throughput
requests are about an order of magnitude slower when going
to a remote system. This is (largely) due to the overhead of
establishing a tcp/ip connection for each request. In the
future we intend to look into caching established connections
so that we don't have this overhead for every request (much
as HTTP 1.1 added over HTTP 1.0). This won't remove all the
disparity between local and remote requests, but it should
bring it down from an order of magnitude to something closer
to a small factor.
I just did a few tests and I'm actually getting substantially
worse performance than you. I'm in the area of 2-3 orders of
magnitude, depending on just how many routers/switches I'm
going through.
We'll look and see if there is any short term fix we can put in
place. Otherwise, we'll have to wait until we can test out the
connection caching idea.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Logged In: YES
user_id=285070
Is there a reason why this support request was assigned to
davepitcher? Also, I don't see a .cap file attached (though
I'm not sure what a .cap file is).
Logged In: YES
user_id=279901
I figured I'd chime in here. This is pretty close to what I
would expect. I have, in general, found that high-throughput
requests are about an order of magnitude slower when going
to a remote system. This is (largely) due to the overhead of
establishing a tcp/ip connection for each request. In the
future we intend to look into caching established connections
so that we don't have this overhead for every request (much
as HTTP 1.1 added over HTTP 1.0). This won't remove all the
disparity between local and remote requests, but it should
bring it down from an order of magnitude to something closer
to a small factor.
I just did a few tests and I'm actually getting substantially
worse performance than you. I'm in the area of 2-3 orders of
magnitude, depending on just how many routers/switches I'm
going through.
We'll look and see if there is any short term fix we can put in
place. Otherwise, we'll have to wait until we can test out the
connection caching idea.
Logged In: YES
user_id=764552
Please find enclosed the .cap file (IP traffic monitored w/
Ethereal/TCPDump).