Blog post on EnterpriseDB...maybe off topic

Christopher Kings-Lynne wrote:

http://www.flamingspork.com/blog/2006/02/16/enterprisedb-where-is-the-source/

Any comments on this? Is he referring to EnterpriseDB extensions that
they don't make public?

I think so. Trying to "battle" the perception that EnterpriseDB is an
open source database. Seems though that little effort is made to
understand the actual relationship between EnterpriseDB and PostGreSQL.

Looks like an attempt at pitting "dual license GPL/closed source" vs.
"proprietary BSD based".

regards,
Lukas

Christoper,

On 2/15/06 11:14 PM, "Christopher Kings-Lynne" <chriskl@familyhealth.com.au>
wrote:

Any comments on this? Is he referring to EnterpriseDB extensions that
they don't make public?

I've noticed a lot of press lately is mentioning their name next to ingres
as an alternative to MySQL, so the MySQL folks might be feeling some
Postgres heat from their direction.

I also wonder where their project is too - they seem publicly opaque about
progress, etc. From the web site's statements it looks like they've written
a tool to tune the postgresql.conf file from which they claim a 50%
speed-up, but that's not new or unique "fork-level" functionality.

- Luke

Any comments on this? Is he referring to EnterpriseDB extensions
that
they don't make public?

I've noticed a lot of press lately is mentioning their name next to
ingres
as an alternative to MySQL, so the MySQL folks might be feeling some
Postgres heat from their direction.

I also wonder where their project is too - they seem publicly
opaque about
progress, etc. From the web site's statements it looks like
they've written
a tool to tune the postgresql.conf file from which they claim a 50%
speed-up, but that's not new or unique "fork-level" functionality.

What they don't say is whether that is a 50% speed up from the
default settings or a 50% increase from a carefully hand tunes file.

I also wonder where their project is too - they seem publicly opaque about
progress, etc. From the web site's statements it looks like they've written
a tool to tune the postgresql.conf file from which they claim a 50%
speed-up, but that's not new or unique "fork-level" functionality.

EnterpriseDB is a fork of PostgreSQL that contains a reasonable level of
pl/SQL (Oracle) compatibility.
My understanding (and I could be wrong) is that they support packages,
in, inout paramters etc.. in
the same syntactical way that Oracle does.

Joshua D. Drake

Josh,

On 2/18/06 7:15 AM, "Joshua D. Drake" <jd@commandprompt.com> wrote:

EnterpriseDB is a fork of PostgreSQL that contains a reasonable level of
pl/SQL (Oracle) compatibility.
My understanding (and I could be wrong) is that they support packages,
in, inout paramters etc.. in
the same syntactical way that Oracle does.

Thanks!

I figure they'll have to do quite a lot to make progress in their chosen
market, including:

- SQL*Net protocol compatibility
- Oracle Number datatype support
- ROWID unique row identifier
- Oracle Redo/Undo log format parsing and replay
- SQL Loader format support
- Oracle exp/imp format support

The broader Oracle enterprise market is used to a high level of integration
of Oracle instances across the enterprise, and their DBAs are highly trained
to use these features.

- Luke

Josh,

On 2/18/06 7:38 AM, "Luke Lonergan" <llonergan@greenplum.com> wrote:

I figure they'll have to do quite a lot to make progress in their chosen
market, including:

- SQL*Net protocol compatibility
- Oracle Number datatype support
- ROWID unique row identifier
- Oracle Redo/Undo log format parsing and replay
- SQL Loader format support
- Oracle exp/imp format support

I forgot one:
- Make sort ordering equivalent to Oracle (trailing blanks don't count, for
instance)

- Luke

Folks,

What they don't say is whether that is a 50% speed up from the
default settings or a 50% increase from a carefully hand tunes file.

AFAIT, most of their performance speed-up comes from two sources:
1) a carefully hand-tuned compile of Postgres using ICC, and
2) Improving on the default postgres.conf params.

BTW, they have set up 3 pgfoundry projects to contribute some-but-not-all
of their improvements to the community, and have actively sought feedback
from me, Bruce, Simon and others on how and what to contribute. They also
paid for Alvaro's work on shared locks.

So if that code has been slow in coming, that's due to their staff being
overcommitted (it's a start-up).

--
--Josh

Josh Berkus
Aglio Database Solutions
San Francisco