2nd Completed Case Study - Monster Labs, Inc. - Mailing list pgsql-advocacy

Hi everyone,

Just received this from Thomas F.O'Connell of Monster Labs, Inc.  It's
the Case Study they just completed in relation to their PostgreSQL
usage, and we should probably get this onto the Advocacy site as soon as
time permits.

Ewald/Robert, are either of you interested in helping with this?

:-)

Regards and best wishes,

Justin Clift


-------- Original Message --------
Subject: Re: [HACKERS] Anyone want to assist with the translation of
the     Advocacy  site?
Date: Thu, 21 Nov 2002 13:28:33 -0600
From: Thomas F.O'Connell <tfo@monsterlabs.com>
To: Justin Clift <justin@postgresql.org>

Justin, here's my completed case study.

Waiver
******

Thank you very much for your responses about PostgreSQL.  We greatly
appreciate the time and thought you've put towards answering our
questions.

We are constantly updating our marketing information, and we would like
the opportunity to use your name and comments where appropriate in our
materials.

To grant the PostgreSQL team use of your name, title, company name
and/or comments, please fill out the following waiver.

*****

I, Thomas F. O'Connell(full name), grant the PostgreSQL.org Marketing
Team
the right to use my details and comments for the sole purpose of
producing
marketing materials and resources to attract more users, companies, and
developers to the PostgreSQL community.

Please fill out the information and mark with an "X" which pieces of
information you grant us permission to use:

X  Name:  Thomas F. O'Connell
X  Title: DBA | Systems Programmer
X  Company: Monster Labs, Inc. (http://www.monsterlabs.com/)
X  Industry: web technology

     Date:  21 November 2002


Case Study Questions
********************

1) How do you currently use PostgreSQL?

PostgreSQL is currently the primary O-RDBMS system we use as a backend
for our web-based content management system and application service
provider, SiteMason (http://www.sitemason.com/). We use it as an SQL
database against which we run complex queries in both development and
production web environments. We have several Apache web servers that use
mod_perl and DBI (with DBD::Pg) to connect to our PostgreSQL
installations.

2) Why did you choose PostgreSQL for your deployment?

PostgreSQL supported the primary enterprise requirements we had when
evaluating RDBMSes: transactions, full referential integrity,
subqueries. We needed an ACID-compliant database, and PostgreSQL met
this requirement. The fact that PostgreSQL has some object-relational
features was a bonus.

Additionally, our company supports the concept of Open Source, and our
IT staff is very comfortable communicating directly with developers and
receiving the kind of support that only the Open Source community is
able to provide.

We had experience with Oracle, and it was resource intensive in terms of
both hardware and people required to maintain it. After experimenting
with PostgreSQL, we determined that it was robust and reliable enough to
meet our needs in a live commercial environment. The price tag is
attractive but comes second to the fact that PostgreSQL has met all our
requirements for an RDBMS.

3) What's the nature of the data you store in PostgreSQL?

We store two types of data: 1) internal data used to control the web
applications we provide and 2) user data collected from the web, mostly
on behalf of customers using our products. The majority of the data is
numeric or ASCII, although we do have Unicode support. We have enabled
multibyte support in our PostgreSQL installations.

4) How much data (in MB) do you host in PostgreSQL?

In our local production environment, we host approximately 450 MB. This
figure has grown fairly consistently over the course of the past year.

We also offer an appliance version of our system that is usually
installed remotely. The largest of these installations is currently
hosting approximately 200 MB, but we expect this to more than double
over the course of 2003.

5) How many IT people currently work on your PostgreSQL installation(s)?

We have 1 DBA (myself) and 1 sysadmin who are responsible for
maintaining the PostgreSQL installations and 3 other developers whose
work often involves querying the database(s).

6) What's the biggest advantage to using PostgreSQL in your deployment?

Reliability. Even when we've had faulty hardware, PostgreSQL has been
surprisingly robust. When a system fails for external reasons,
PostgreSQL's recoveries have been surprisingly fast and complete.

7) What's the biggest disadvantage?

I have done a fair amount of performance tuning, but there are still a
few queries I run that seem like they could stand to be a bit faster.

Additionally, without any active PostgreSQL developers on staff, we are
reliant on the PostgreSQL community to provide us with new features. For
instance, we've hoped for replication and more significant development
of the object-relational model for years, now. This is one downside to
the Open Source community: inconsistent pace and direction of
development.

8) Can you briefly describe your network architecture?

Our local architecture consists of a layer of load-balanced front-end
web servers running Apache. These serve up cached or static pages. Any
processing requests are sent to a layer of Apache mod_perl servers that
have direct connections to our database layer. We currently have a
monolithic setup for PostgreSQL while we wait for replication to be
avialable. There is a failover setup that receives periodic snapshots of
our PostgreSQL installation and is ready to go live in the event of a
primary database server failure. All machines are GNU/Linux systems
connected by 100 Mb Ethernet.


***********

> Thanks heaps Thomas.

glad to help. sorry it took so long!

-tfo

Thomas F. O'Connell
DBA/Systems Programmer
Monster Labs, Inc.
http://www.monsterlabs.com/
110 30th Avenue North, Suite 1
Nashville, TN 37203
phone: 615-301-2600 x132
fax: 615-301-2602