Yes, You Can Run Your Business On PostgreSQL

Ohio LinuxFest 2012

Josh Williams

End Point Corporation

Basically, because PostgreSQL rocks.

• We'll discuss some fallacies we tend to hear
• ... and why they don't make sense.
• A few alternative projects
• Case studies

But wait, who am I?

• Postgres user and contributor ~8 years
• End Point Corporation, Postgres consultant 2+ years.
• But...
• Don't worry, I'm on your side now!

Non-free Software Logical Fallacies

The "Can't Sue" Problem

• "With Open Source software, there's nobody to sue if it breaks."
• True... BSD license:
  • THIS SOFTWARE IS PROVIDED ... "AS IS" ... IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE...
• THE SOFTWARE IS NOT FAULT TOLERANT ... NO LIABILITY FOR ANY INDIRECT, SPECIAL, CONSEQUENTIAL OR INCIDENTAL DAMAGES... (- License from big $$$ Database)
• It's not really an option either way.
• Real translation: Management doesn't want full responsibility

"Enterprise level support isn't available"

• Sure there is... A few Postgres Support Companies

  (hi)

• And that's just a few.
• Not to mention systems level: Red Hat, Canonical, ...

"My Enterprise-Grade system can't fail"

• Sure, a big $$$ system will be quite solid, but...
• All risk is strictly non-zero.
• You may have to do some of the configuration and customization work anyway!
• If you build it yourself, you can calculate the risks.

"Open Source experts are hard to hire"

• 
  Source: indeed.com/jobtrends
• pgsql-jobs postings are up, as well as the usual hiring sites
• Do you need to?
  • Train who you already have.
  • Professional Service companies...

"You get what you pay for!"

• "Expensive software has higher quality code than free software."
• The PostgreSQL patch acceptance process...
• ... results in a very high quality code base.
• I can't say the same for some closed source products I've used.
• That famous Coverity report (February, 2012)...
  • Average defect density for proprietary codebases: .64
  • Average defect density for open source projects: .45
  • PostgreSQL 9.1: .21

"We bought this system for $1M"

• "(... in 1988.) We can't throw it away. Even if it is costing us $100k per year in licensing/maintenance."
• Sunk Cost Fallacy ... or "throwing good money after bad"
• To some extent, management afraid to go back on an old decision.
• It's not sunk cost... It's opportunity cost of not going with FOSS.

A Concession — There Are Legitimate Concerns

• My specialized software package requires $BigNameData.
  • Put pressure on the vendor to look at other options.
• Re-training my staff on {Postgres,Linux,etc} is costly.
• Other side of the coin... Migrations take considerable resources.
  • Figure out where your return on investment would be.
• Building a solution yourself takes time.
  • A pre-built solution won't be dropped into place, there will be testing and training time anyway.

Open Source Opens Doors Not Present with Nonfree Software

• Got talent? You have the source, you can customize the source.
• There's no such thing as a dead project. Fork the code and keep it going!
• Contributions back → Company publicity!

Open Source Logical Fallacies (Gasp!)

• One Size Fits All - There's other options...
  • MySQL (and this from a Postgres guy)
  • NoSQL (CouchDB, MongoDB, Cassandra)
• Look beyond core - Associated projects are good, too...
  • Replication alternatives: Bucardo, Londiste, Slony
  • Foreign Data Wrappers
  • check_postgres for monitoring ... etc...

"Okay, fine, we'll adopt."

• You're in good company...
• French Government Outlines Plans For Free Software Adoption
  • http://www.cio.com/article/717033/French_Government_Outlines_Plans_for_Free_Software_Adoption
• FAA maps all their airfields on PostGIS (PG East 2010)
• Oh yeah, and ...
  

"Okay, fine, we'll adopt."

• Some that End Point has supported...
• State Child Welfare agency.
• State department of health, data warehouse.
• Providence Plan, "the largest data warehouse in Rhode Island."

"Okay, fine, we'll adopt."

• Some that End Point has supported...
• outside.in
• Backcountry.com
• CityPass
• FrozenCPU
• FragranceNet

"Okay, fine, we'll adopt."

• Other famous ones...
• Afilias
• Instagram
• Skype
• Sony
• etc
• etc

In Closing

• You have little to worry about.
• Don't focus on one project and just on the core project.
• Figure out your ROI.
• You're in good company.