[Mono-bugs] [Bug 78075][Nor] Changed - Mono SSL stack performance/tuning issues

[bugzilla-daemon at bugzilla.ximian.com]

Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.

Changed by flucifredi at ximian.com.

http://bugzilla.ximian.com/show_bug.cgi?id=78075

--- shadow/78075	2006-05-01 20:12:09.000000000 -0400
+++ shadow/78075.tmp.12920	2006-05-02 16:30:35.000000000 -0400
@@ -419,6 +419,50 @@
 negotiated during the SSL/TLS handshake. That way you'll be sure if
 it's AES256 or (I really doubt) something else.
 
 Side note: we do not share the same definition of "order of magnitude"
 because in mine: 1 order of magnitude == 10x, 2 order of magnitude ==
 100x ... ;-)
+
+------- Additional Comments From flucifredi at ximian.com  2006-05-02 16:30 -------
+In order: 
+
+>I'm not sure how you can compare (or interpret the results from) ZLM
+>times versus wget times.
+
+installing the testing rpm from the cache takes 7 seconds. hence,
+"time rug bin -y <rpm>" and "time wget" (assuming James confirms that
+we are not reading the download byte by byte or doing something
+equally silly) is rather comparable. 
+
+ -> James, can you comment on mget vs zlm ? I also still need to know
+that we are not doing weird things in the download code.
+
+>You can also use ethereal (on a smaller RPM) to see which cipher 
+> gets negotiated during the SSL/TLS handshake. That way you'll be 
+> sure if it's AES256 or (I really doubt) something else.
+
+I use openssl in client mode to check what cypher is being used, and
+did so for all the results posted - when I say one cypher was used, it
+was.  As far as the default is concerned, I did explicitly test with
+AES 128 and it had the same low performance as did the default. 
+
+The default choice turns out to be "Cipher is EDH-RSA-DES-CBC3-SHA".
+The 37 second figure I posted above is, as I stated already, with RC4-SHA.
+
+-> Sebastien, can you comment on the why I would not be able to
+compare download via rug/zlm vs wget ? As I stated, I am accounting
+for the fixed "estra" part of the process (install, etc), and all the
+other variables are kept the same. 
+
+> Side note: we do not share the same definition of "order of 
+> magnitude" because in mine: 1 order of magnitude == 10x, 2 order of 
+> magnitude == 100x ... ;-)
+
+Go figure, I also thought the same. And 7 seconds, 37 seconds, 2m7sec
+(127 seconds) are *indeed* in three different orders, the powers of
+ten 0, 1 and 2.
+
+
+
+
+