From DBSight Full-Text Search Engine/Platform Wiki

Table of contents

Benchmark Environment

This benchmark runs on DBSight 4.0.4, Jetty server, started by

java -jar start.jar
DBSight Version 4.0.4
DBSight Build 93
CPU Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz
OS Name Linux
OS Version
JVM Vendor Sun Microsystems Inc.
JVM Version 1.7.0-ea-b76

Benchmark Command and Result

This runs on a real index used in a high traffic web site, with 8 facet searches enabled. The index size is 2.8GB, used memory 320MB.

[chris@chris-server ~]$ ab -n 10000 -c 4 "http://clu-dt:8080/"
This is ApacheBench, Version 2.3 <$Revision: 655654 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd,
Licensed to The Apache Software Foundation,

Benchmarking clu-dt (be patient)
Completed 1000 requests
Completed 2000 requests
Completed 3000 requests
Completed 4000 requests
Completed 5000 requests
Completed 6000 requests
Completed 7000 requests
Completed 8000 requests
Completed 9000 requests
Completed 10000 requests
Finished 10000 requests

Server Software:        Jetty(6.1.6)
Server Hostname:        clu-dt
Server Port:            8080

Document Path:          /
Document Length:        37086 bytes

Concurrency Level:      4
Time taken for tests:   30.557 seconds
Complete requests:      10000
Failed requests:        0
Write errors:           0
Total transferred:      373484721 bytes
HTML transferred:       370860000 bytes
Requests per second:    327.26 [#/sec] (mean)
Time per request:       12.223 [ms] (mean)
Time per request:       3.056 [ms] (mean, across all concurrent requests)
Transfer rate:          11936.21 [Kbytes/sec] received

Connection Times (ms)
             min  mean[+/-sd] median   max
Connect:        0    0   0.1      0       1
Processing:     6   12  21.3     10     497
Waiting:        5    9  19.7      7     494
Total:          6   12  21.3     10     497

Benchmark Conclusion

This is not a thorough test, but a very real test. Don't be fooled by other tests that uses a small index, or just doing search with other features disabled. We could have done that. But those data would not be useful and are only marketing tricks.

But the numbers is not bad. With one single machine, you can see what you can get hundreds of searches per second. With sharded search enabled, you can handle thousands searches per second. You can build a small yahoo scale website.