[291] in Tooltime
Scopus benchmarking
daemon@ATHENA.MIT.EDU (Steven Wade Neiterman)
Sun Oct 13 17:53:31 1996
Date: Sun, 13 Oct 1996 17:55:05 +0100
To: tooltime@MIT.EDU
From: Steven Wade Neiterman <wade@MIT.EDU>
Cc: lang@ISIS.MIT.EDU
At last Thursday's Help Desk staff meeting, it was agreed that we would
evaluate the database performance with greater data volumn that currently
used.
Peter Lee created "dummy' data by creating a four-fold duplication of the
data previously loaded.
The current environment:
Test database (help-them) Production database (help-it)
mit_logs = 4,249 records mit_logs = 860 records
netinstall = 322 records netinstall = 137 records
My testing was done using the Windows client, and I presume with no one
else using the database.
Results:
QBE search was just slightly slower with the higher record count, took
about five seconds on test and a two-three seconds on production.
History search was noticeably slower, about 65 seconds to search on
"eudora". Strangely enough, searching on multiple words, e.g.
"eudora%macintosh" took only eight seconds. Go figure.
Let me know if you get different results. And thanks to Peter for
generating the data and to Bob Lang for suggesting the test.
..Steve
