Thinksum

EnterpriseDB

My company, Trusted Network Technologies, embeds the Postgres (sorry, I just can't get used to calling it PostgreSQL) database in our system.  We use Postgres to store our configuration data, but we mainly use it to store security audit data.  Since our system collects security audit data for connections we see on the network, certain tables in our database can obviously get real large real fast (one of our customers captures about 100gig every week).  So we need fast insert, reliability, and then we need fast query on very large data sets.

A few months ago, we decided to use the new commercial Postgres company, EnterpriseDB.  They've assembled many of the Postgres technical leaders, and they offer a fully tested and supported version of Postgres with some improved features and utilities, plus consulting and support.  We decided to use them because we had struggled ourselves with a few support issues related to Postgres, and we wanted a more direct way to tap into Postgres experts to make sure our system was fully tuned.

Our first few months with EnterpriseDB have been great.  We've received excellent advice on tuning our system, and we expect it to show huge dividends in upcoming releases.  For anyone shipping a product with an open source database, take a good look at EnterpriseDB.  They're also trying to compete with Oracle inside the enterprise, and frankly I don't see why they shouldn't be strongly considered there too.  If you're already using Postgres for any business, then picking up support from EnterpriseDB seems like a no-brainer to me.

What's the Value of Open Source?

A recent post on Slashdot on the availability of the Sunbird calendar from Mozilla called it "good news for all of us waiting for decent free calendaring software."

And I thought to myself, yup, this is what Richard Stallman's free software movement has brought us (and let's not kid ourselves, all of this never would have happened at least not yet without Richard and FSF and GNU).  A lot of folks could care less about the free software (free as in freedom of speech) as long as they get the free software (free as in free beer).

And now IBM is backing open source and putting considerable investment behind Linux and Eclipse and PHP, and Google is backing Firefox.  Because both of them hate Microsoft, and these projects further their own interests.  A lot of open source developers are now directly or indirectly working for IBM or Google to offer free software (free as in beer) in a war against Microsoft.

And now Richard Stallman is calling for a free BIOS, and suggesting that developers should boycott Intel.

Do we really need or want a free BIOS?  Someone will probably do it, it might be fun, but is there an ethical or scientific or protective reason we really need it?  Is it morally an important issue, so I should boycott Intel?  Hmmm.

I've written in the past about some of the things that absolutely work in the open source model (see Economics and Availability or A New Software Business Model).  But I've never really been on board with the ethical argument (freedom of speech).  I'm with Bill Joy ... "I don't need to see the source code."  I use a lot of open source software myself, and I never have and probably never will look at the source of most of those products.  I use them because a) they're high quality and meet my requirements which include portability; and b) they're free (as in beer).  As IBM, Google, RedHat and others become driving forces behind open source, the argument that open source is somehow more ethical seems specious at best.

Open source succeeds, in my mind, for a few key reasons:

• The people who work on it are passionate about what they are doing, and are subject to public scrutiny, which results in better products.
• By it's nature open source avoids "vendor lock-in," so products are available and supported across many platforms and configurations.
• It's the greatest marketing scheme in the world -- free as in free beer.  If you give away free beer, you are guaranteed a crowd ... if it's good beer, people will be knocking down the doors.  It's a great way to eliminate barriers, generate word-of-mouth, and get a new product out there fast.  It worked for Tim Berners-Lee and the world wide web.  The economics of production (cheap labor) and distribution (the internet) make it possible.

So open source is obviously succeeding and will obviously continue to succeed, but it's more about value (high quality at low cost) than values (moral superiority).

Economics and Availability

On the 20-year anniversary of his start of the GNU project, Richard Stallman has written an analysis on the state of open source. While his analysis is thought-provoking and well worth reading, and his altruism is passionate, his position that open source is driven by an ethical imperative is misguided.

The popularity of open source is driven by economics and availability. Economics says that if a product is needed and costs less than others, and the product is of equal or superior quality than the competition, consumers are going to prefer that product. Of course, consumers also need a way to get the product, and that's where availability comes in.

In the case of open source, the software is cheaper (often free as in beer) because programmers have generally donated their time and released the result into the public domain. And the results are available and easily accessible thanks to the internet.

A product that is cheap and available, and meets a need, can of course be wildly popular even if it isn't open source. Napster is a great example. But, as discussed in a previous post, the open source model does offer potential advantages for improved quality, and it may garner increased adoption because it attracts programmers as well as end users. In fact, open source is a great way to market to the developer community. But even with developers, economics and availability are often more important than actually providing code. BEA, and others, have proved this with successful programs that involve giving away software to developers even though the software isn't open source. As Bill Joy recently noted, many people don't need to see the source.

If Richard Stallman wasn't a programmer, and GNU never began, open source and other free software would still exist. That's because 1) there are lots of programmers with time on their hands, 2) programming is fun and educational (for programmers), and 3) the internet provides the means to easily distribute results. As programming tools evolve and programming gets easier, the amount of open source will only increase. Someday, most programs run in businesses and homes may in fact be open source, as Richard Stallman hopes, but not because a group of programmers sought to change society, but because of supply and demand.

A New Software Business Model

While Microsoft tries to discredit it, Linus Torvalds succinctly points out why open source works. It's about transparency and accountability -- if everyone and anyone can see what you're doing, and nothing is hidden, then what you're doing damn well better be good or it won't survive. And it's also about cooperation -- allowing anyone with an interest to contribute, albeit through a controlled process.

The success of open source is unquestionable. I'm not sure people working inside the software industry realize exactly how successful or revolutionary all this really is. Linux and the Apache Web server are two of the most successful software applications of the last decade. Is there any doubt that other new, successful open source applications will appear in the next decade? And the most important point for all of us working in software to get is that customers aren't just using this because it's low cost, but also because it's just as good or better.

The question facing Microsoft and every software company is what to learn from this? Obviously, software companies are in business to make money, so they can't give away their applications for free. But what software companies can learn from the open source model is what makes software better -- transparency, accountability, and cooperation. There is a new software business model to be found in all this, one where the software company still charges for its intellectual property, but it uses an open model to the benefit of its customers.