--- In ib-support@y..., "David K. Trudgett" <dkt@r...> wrote:

> BTW, did you know that Oracle forbids database comparisons in its
> EULA? Unbelievable. Makes it look like they have something to hide,
> doesn't it?

I believe it's the same for MS SQL server. I spent some time
recently on the net trying to gather server comparison specs. All I
could turn up was...

http://mlmsoftware.com/sql.htm

"According to KeyLabs of Provo, Utah [the country's largest
independent testing company for this kind of product], InterBase is
6.5 times faster than SQL Server on average. Tests were done using
WinNT 4.0 on databases of up to 600,000 records, and up to 100 users."

This is backed up by the makers of Advantage database server. I had
to laugh: trying to trumpet their product but the Interbase
benchmarks weren't too bad in comparison and in fact surpassed theirs
on several occasions. I found a pdf doc they compiled based on
commissioned tests with Keylabs.

To summarise...

-----
Extended Systems contracted with KeyLabs,a full-service third party
test facility located in Provo,Utah,to conduct a series of benchmark
tests comparing Advantage Database Server with Borland InterBase and
Microsoft SQL Server.The focus of the testing was to establish
relative times for clients to perform a series of tests against each
of the above mentioned servers.

The tests were run using an Advantage native driver against Microsoft
and Borland native drivers and the Advantage ODBC Driver against
Microsoft and Borland ODBC drivers.

The benchmark tests were combinations of SELECT,UPDATE,and
movement operations run on 10,50 and 100 user Microsoft Windows NT
networks. The native driver test results show Advantage performing,on
average,124 times more operations per hour than the Microsoft SQL
Server solution and 19 times more operations per hour than Borland
InterBase. The ODBC driver test results show Advantage performing 30
times more operations per hour than Microsoft SQL Server and 6 times
more operations per hour than Borland InterBase.
-----

The important thing to note here is the comparison of Interbase to
MSSQL -- the version of interbase was 4.2 and the version of MSSQL
server was 6.5.

Continuing...

-----
Test Results
The following graphs depict the benchmark results for the 100 user
tests in terms of average number of operations per hour and the time
to complete each of the seven benchmark tests.

Native Driver Results

On average,Advantage native drivers performed UPDATEs 218 times
faster than Microsoft SQL Server and 25 times faster than Borland
InterBase. SELECTs were performed 29 times faster than Microsoft SQL
Server and 43 times faster than Borland InterBase. The performance
difference when positioning to the next record was negligible.

ODBC Driver Results

On average,Advantage Database Server via the Advantage ODBC Driver
performed UPDATEs 85 times faster than Microsoft SQL Server and 8
times faster than Borland InterBase. SELECTs were performed 14 times
faster than Microsoft SQL Server and 43 times faster than Borland
InterBase. The performance difference when positioning to the next
record was negligible.
-----

Interbase -- in these tests -- performed poorly with SELECTS. This
was a concern because I had to do a FB vs MS SQL server comparison to
convince a client to go with FB. It's not easy explaining all the
variables that affect performance to managers who simply want a
couple of bar charts showing a clear winner.

full pdf is here...

http://www.advantagedatabase.com/NR/rdonlyres/3f2c25d36m62btxyeq2mhpy5
hmemczbpsl3utepq3kxb2wf6pipf2yter6khxasllqwxxeousnlgh5syq346bgndpg/per
formance_benchmarks.pdf

That's one hell of a url. Are they trying to hide it from Microsoft,
perhaps?

Regards,
Andrew Ferguson