(mySQL) benchmarks strike back
Andrew L. Gould
algould at datawok.com
Fri Feb 11 18:21:39 GMT 2005
On Friday 11 February 2005 10:46 am, Jorge Mario G. wrote:
> Hi there
> I just read
>tid=29%20result and as in any onther benchmark there is a lot stuff
> can be arguable. I would like to know why is that
> the problem is that "we" are pushing
> FreeBSD/postgreSQL as a database solution, and I am
> the guy to blame to, because I was the one who did
> advocacy for FreeBSD, so I'm sure my boss is going to
> ask me. And you told us to use FreeBSD instead of
> and I do not want to answer him "beastie is way more
> I'm doing my own research but some help from here
> would be nice!!!
> Jorge Mario Mazo
Interesting article; but how does it relate to the real world. I'm not
saying that the benchmarks aren't valid. Perhaps Linux has gained
advantages in performance over FreeBSD -- I don't know, I'm not
qualified to say. Regardless, benchmark test results should not be the
only criteria for selecting an operating system. Be wary of anyone who
tells you otherwise.
1. YMMV (Your mileage may vary.) I've seen benchmarks that favor Linux
before; but when I've tested Linux using complex queries with large
databases, the system slowed much more noticeably than with FreeBSD.
This is based upon my perceptions rather than benchmarks; but it
reflects the system's effect upon my productivity, which is very real.
Linux may have improved since then; but it demonstrates that benchmark
tests do not always reflect what happens outside of the laboratory.
You should run comparisons using activities that reflect your computing
2. Security -- See the link below. (beware of wordwrap)
3. Usability -- FreeBSD differs from Linux in many ways. For me, the
file system hierarchy and the way the operating system works makes more
sense to me. FreeBSD was easier for me to learn and to use.
4. The Linux distro used was Gentoo, which promotes system
optimizations. For example: Whereas most distributions install a
generic kernel, the Gentoo installation defaults to compiling a new
kernel that is customized to the hardware. Did the testers perform
similar steps with other operating systems? Also, opinions as to
whether Gentoo is suitable for a production server are polarized.
Linux gamers tend to say "yes". More conservative users often say,
"no". Running the benchmarks using SUSE Professional, Slackware or
Debian (or others) would have provided more comfort to business users.
5. Is MySQL different from PostgreSQL in ways that should affect the
relevance of the benchmark tests to your situation? (See item #1.)
Best of luck,
More information about the freebsd-questions