From 0ea5fc66924303d1bf73ba283a383e2aadee02f2 Mon Sep 17 00:00:00 2001 From: neodarz Date: Sat, 11 Aug 2018 20:21:34 +0200 Subject: Initial commit --- pipermail/nel/2001-February/000295.html | 68 +++++++++++++++++++++++++++++++++ 1 file changed, 68 insertions(+) create mode 100644 pipermail/nel/2001-February/000295.html (limited to 'pipermail/nel/2001-February/000295.html') diff --git a/pipermail/nel/2001-February/000295.html b/pipermail/nel/2001-February/000295.html new file mode 100644 index 00000000..e4127c7d --- /dev/null +++ b/pipermail/nel/2001-February/000295.html @@ -0,0 +1,68 @@ + + + + [Nel] NeL Network Engine + + + + + + +

[Nel] NeL Network Engine

+ Nicolas Hognon + cblt@cblt.org
+ Wed, 28 Feb 2001 15:10:53 +0100 +

+
+ +
> Looking at Apache or Samba projects, it seems that a good compromise is to
+> set a 'maximum client requests by thread' and spawns threads accordingly.
+> Apache uses process forking and memory sharing, but the design remains the
+> same. You then just tune this max_request_by_thread for each OS, say 1 for
+> Solaris which is said thread-efficient, 10 for Linux ? Just a hint ...
+
+ok this is a solution
+but as i said in my last mail the problem
+is how managing more than one socket per thread efficiently ?
+
+under windows i know other solution than select / poll but under
+unixes i don't know other solutions.
+
+some one knows other one ?
+
+-- 
+Nicolas Hognon
+nicolash@virtools.com
+
+Virtools - The Behavior Company
+Tel. (+33) 1 42 71 46 86 / Fax. (+33) 1 42 71 86 53
+http://www.virtools.com/
+
+
+
+ + + +
+

+ -- cgit v1.2.1