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-July/000483.html | 63 +++++++++++++++++++++++++++++++++++++ 1 file changed, 63 insertions(+) create mode 100644 pipermail/nel/2001-July/000483.html (limited to 'pipermail/nel/2001-July/000483.html') diff --git a/pipermail/nel/2001-July/000483.html b/pipermail/nel/2001-July/000483.html new file mode 100644 index 00000000..5189b33f --- /dev/null +++ b/pipermail/nel/2001-July/000483.html @@ -0,0 +1,63 @@ + + + + Re[2]: [Nel] TCP vs. UDP + + + + + + +

Re[2]: [Nel] TCP vs. UDP

+ Dim Segebart + zager@teleaction.com
+ Thu, 5 Jul 2001 17:45:39 +0200 +

+
+ +
Just facts:
+We  run  relatively large chat server with our own server software. We
+use  TCP  for all communications between Java clients and server-side.
+At  peak-load  time we have 2000 opened TCP-sockets at once. Profiling
+shows  up  what most of CPU time server performs write().
+And finally SW and HW - we use SunOS 5.7 aka Solaris7
+Our homemade software written in C++ utilize 2,5 CPU out of 4 in our
+Sun Enterprise Server 450 with 4 CPU and 1G RAM
+May  be  someone  have  the  similar  service on UDP and can share his
+statistic ?
+
+--
+Dim Segebart
+
+
+
+
+ + + + + + +
+

+ -- cgit v1.2.1