Re: BUG #6269: Anomaly detection

On Tue, Oct 25, 2011 at 10:51 AM, Paul Stapersma
<paul(dot)stapersma(at)gmail(dot)com> wrote:

> PostgreSQL version: 8.3.3

> For a project at my University, we compared PostgreSQL with MySQL's InnoDB.
> In this research, we found several cases in which anomalies where detected
> in Isolation levels that guaranteed not to have these anomalies.

It looks to me that you've made an honest attempt at this comparison.

Thank you for noting points such as
"PostgreSQL’s overall performance was for all isolation levels better
in comparison to InnoDB’s performance."

To allow this type of study to be widely accepted, you need to declare
who you are and where you are from. And to allow your results to be
confirmed. Accountability and verifiability are essential aspects of
good science. I would also like you to state your funding source, so
that it is clear whether or not marketing funds inspire or reward such
studies. Full disclosure is appreciated.

Using InnoDB 5.0 and PostgreSQL 8.3 isn't very useful and I'm
surprised you haven't used the latest production releases of both
products, especially when we have ground breaking features in exactly
this area in the latest release. Perhaps I misunderstand - you don't
actually say which release of InnoDB you use, though you have URLs
linking to both 5.0 and 5.1 versions.

As others have observed, some other points need more careful
examination and it would be good to provide exact details of your
tests to allow any such results to be reproduced. If problems do
exist, we will be very interested.

There are many people here that would be happy to help if you find
problems, so please don't be put off by these responses to your work.

--
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services