commit 396b22a87bb23a18e0f4e2e3e659bf84442ed2ef
Author: Peter J. Holzer <hjp@hjp.at>
Date:   Thu May 6 22:34:17 2021 +0200

    Add README

diff --git a/README.md b/README.md
new file mode 100644
index 0000000..4c241d9
--- /dev/null
+++ b/README.md
@@ -0,0 +1,22 @@
+Database Connection Benchmark
+=============================
+
+
+What it the minimal time to get any result from a database?
+
+We open a connection, perform a trivial query (current_timestamp is
+probably fine) and disconnect.
+
+There are (at least) three times we are interested in:
+
+* The time from initiating the connection to the end.
+* Same plus the time to load any necessary libraries
+* Same plus the overhead of starting the script
+
+The first is the one important for long-running servers which for some
+reason don't want to keep a persistent connection to the database.
+
+The last is for command line tools, CGI scripts, etc. 
+
+The second doesn't have that much practical importance. That's just
+personal curiosity.