BenchMark Tests
We BenchMarked SurgeLDAP against OpenLDAP which is a common LDAP server that is available.
 
The version of SurgeLDAP we used was:
 
SurgeLDAP v1.0k (build 1)

The verison of OpenLDAP we used was:
 
OpenLDAP: slapd 2.2.18 (Nov 24 2004 11:09:58)

For any support questions or issues please Email:

support-surgeldap@netwinsite.com
BenchMark
We used a 3rd party BenchMark tool to ensure that we did not bias the tool towards our LDAP server. The tool which we use was:
 
DirectoryMark 1.3

Using this BenchMark we generated a 10,000 record ldif file and imported this correctly into both LDAP servers.
 
We then generated 3 scripts:
 
Number of Threads   Searches per thread
Single Thread   1000
2 Threads   1000 (2000 total searchs)
10 Threads   1000 (10000 total searchs)

These are the scripts which were then used to generate the results.
Server Information
We setup both LDAP servers on the same server to generate better comparable results.
 
The server was:
CPU:   200Mhz MMX (single)
Memory:   64Meg
O/S:   RedHat 6.0
Harddrive:   5400RPM ATA66 IDE
Network:   10Mb

The BenchMark program was running on a seperate more powerfull machine to ensure that the BenchMark was not the limiting factor.
 
CPU:   2000Mhz Athon (single)
Memory:   512Meg
O/S:   Windows 2000
Harddrive:   7200RPM IDE
Network:   100Mb
OpenLDAP slapd.conf Setup:
schemacheck   on
database   LDBM
suffix   "dc=netwin,dc=co,dc=nz"
allow   bind_v2
SurgeLDAP Setup:
The only change that was made to surgeldap after the default installation was to stop surgeldap from limiting the number of connections per second from 10/sec to unlimited.
Results
See table below for results.
 
The server was:
  Single Thread   2 Threads   10 Threads
  SurgeLDAP OpenLDAP   SurgeLDAP OpenLDAP   SurgeLDAP OpenLDAP
operations / second 58.8 31.3   60.6 38.5   60.6 33.8
Average Time (ms) 16 31   32 49   162 291

Below are some extra performance results with just SurgeLDAP, as OpenLDAP wasn't able to be setup on these machines.