Re: Press Release review comments

On Fri, 14 Dec 2007, Simon Riggs wrote:

> I measured ~900% performance gain on a pure INSERT workload. Any
> transaction whose major component is the commit time will benefit from
> increased performance, i.e. short transactions get even faster.

I saw something in the multi-hundred percent range as well for some simple
benchmarks. The huge win is cases where right now the limiter is the rate
of commits to disk, and in other situations I wouldn't be surprised to
find little or no improvement.

I'm not sure how to make that distinction clear to regular users though.
I've been thinking about it since your first message in this thread, and
so far I'm at a loss as to how to fairly and accurately characterize the
improvement in a way that's short enough to put into the press release.
It's hard to say any variant on "you can trade-off reliability for speed"
without sounding like you don't care about reliability; you need the
context that some competitors have been doing that already without making
it so obvious to appreciate the feature.

> Many web apps fall into this category, especially those currently
> running on MySQL, Solid etc.

When I was working on the big MySQL comparison piece over the summer, the
only upcoming 8.3 feature I mentioned by name was async commit because it
will shift the landscape more than anything else. You certainly have my
vote that it's more noteworthy than the background writer changes, but
writing a description blurb saying that may take some work. Maybe you
should hire the PR people those guys who worked on the BGW change used,
they managed to get a small tuning feature listed right up there in the
big press release.

--
* Greg Smith gsmith(at)gregsmith(dot)com http://www.gregsmith.com Baltimore, MD