Re: [Fwd: PostGreSQL information]

Поиск
Список
Период
Сортировка
От Paul Ramsey
Тема Re: [Fwd: PostGreSQL information]
Дата
Msg-id 3EF1D905.5070601@refractions.net
обсуждение исходный текст
Ответ на [Fwd: PostGreSQL information]  (Justin Clift <justin@postgresql.org>)
Список pgsql-general
Hi Brad,
Here's some answers to your questions and some references. Firstly, a
very good thumbnail page of information is here:

http://www.postgresql.org/users-lounge/features.html

> Can you please provide me information on licensing and maintenance
> costs, as well as details on platforms PostgreSQL is available on?

PostgreSQL is an open source product, so you can download and use the
code for free, without restriction.

Support packages are available from a number of different companies,
with different pricing options.

http://techdocs.postgresql.org/companies.php

> For example is  it available on Tru64 Unix and can it take advantage of the
> Large Memory Model, of a 64 bit operating system?

It is available for Tru64.
It can take advantage of 64 bit operating systems.
I don't know about memory, my inclination would be to say to would take
advantage of whatever shared memory you were willing to expose for it,
but will have to get verification from elsewhere.

 > Also, what are the
> database connectivity options, ODBC, JDBC etc.

ODBC, JDBC, of course. Also bindings for C, C++, Perl, Python, TCL, PHP,
and almost every other language under the sun. There is a .Net provider too.

> How does it's
> performance compare to Oracle and DB2 etc.

My personal experience is that it blows Oracle out of the water
speedwise, but my personal experience tends to be with transactionally
light applications. Some benchmarks published a couple years ago showed
PgSQL at that time was comparable to the "leading commercial database"
under some standard transactional testing scenario (a TPC-C style test),
and PgSQL has improved alot since then.

Performance benchmarks are hard, because they are expensive to do, and
licencing from the proprietary vendors generally prevents you from
publishing the results without their say so. So one could spend alot of
time and money creating an unpublishable benchmark. :/ If performance is
a deal breaker, the best thing is to put all the candidates on the bench
yourself and stress them the way you think they are going to be
operationally stressed.

Paul

--
       __
      /
      | Paul Ramsey
      | Refractions Research
      | Email: pramsey@refractions.net
      | Phone: (250) 885-0632
      \_


В списке pgsql-general по дате отправления:

Предыдущее
От: Steve Crawford
Дата:
Сообщение: Re: Single Instance of Backend
Следующее
От: Tom Lane
Дата:
Сообщение: Re: A creepy story about dates. How to prevent it?