[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 james at ximian.com.

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

--- shadow/78075	2006-04-12 16:12:13.000000000 -0400
+++ shadow/78075.tmp.28888	2006-04-12 16:28:38.000000000 -0400
@@ -1,12 +1,12 @@
 Bug#: 78075
 Product: Mono: Class Libraries
 Version: 1.1
 OS: unknown
 OS Details: 
-Status: NEEDINFO   
+Status: REOPENED   
 Resolution: 
 Severity: Unknown
 Priority: Normal
 Component: Mono.Security
 AssignedTo: miguel at ximian.com                            
 ReportedBy: naresh at ximian.com               
@@ -205,6 +205,50 @@
 similar load...
 
 size	cipher		mono		wget
 950Mb	RC4/128		1m58s		1m29s
 40Mb	RC4/128		6.023s		4.061s
 2.5Mb	RC4/128		1.393s		0.469s
+
+------- Additional Comments From james at ximian.com  2006-04-12 16:28 -------
+I tried to duplicate the test just now, using stock apache from FC5. 
+I got the following results:
+
+-------------------------------------------------------------------
+time mono mget.exe https://itchy/foobar
+GET https://itchy/foobar
+saving to: foobar
+completed in 00:26:01.5953670.
+
+1449.53s user
+52.67s system
+26:02.22 total
+
+-------------------------------------------------------------------
+
+time wget --no-check-certificate https://itchy/foobar
+--15:01:30--  https://itchy/foobar
+           => `foobar'
+Resolving itchy... 192.168.1.102
+Connecting to itchy|192.168.1.102|:443... connected.
+WARNING: Certificate verification error for itchy: self signed certificate
+WARNING: certificate common name `itchy.snorp.net' doesn't match
+requested host name `itchy'.
+HTTP request sent, awaiting response... 200 OK
+Length: 1,500,000,000 (1.4G) [text/plain]
+
+100%[==================================>] 1,500,000,000   10.94M/s   
+ETA 00:00
+
+15:03:42 (10.88 MB/s) - `foobar' saved [1500000000/1500000000]
+
+43.47s user
+21.69s system
+2:11.68 total
+
+-------------------------------------------------------------------
+
+mget.exe was 33 times slower, so I am highly unconvinced that there is
+no bug here.  Given that I was running against apache, it's pretty
+likely that I was using a 256 bit key (I don't know how to check).  We
+can try to knock down our encryption to 128 bit, but I still think
+this needs fixed.