Open Source Databases "50% Cheaper"

pete314 writes, "Open source databases can cut the total cost of ownership of a database by up to 60% compared to the cost of running proprietary databases from Oracle, Microsoft or IBM. According to data collected by Forrester Research, the savings average about 50%. Open source databases however still struggle to reach mission-critical enterprise applications because enterprises perceive them to be less secure and stable."

Comments

enterprises also want

[rubycodez]

enterprises also want paralleling clusters and failover clusters. The open source databases are getting there, give it few more years.

Re:enterprises also want

[wiggles]

The one place that the Open Source products have a long way to go is support. Companies don't think mysql and postgres are unreliable, they're just not backed enough. The company I work for could give a rat's hindquarters about TCO -- they just want to outsource their risk so that if something breaks, the CIO/CEO/Chairman has someone to argue with. The Chairman can play golf with Larry Ellison as he tries to get more concessions out of Oracle, but he can't play WOW with the 19 year old kid who added some bit of code to mysql.

And before you say it, MySQL AG is still small potatoes compared to Oracle, Microsoft, or IBM.

Re:enterprises also want

[dingDaShan]

It's true. Any person who is in an IT job does not want to risk going to an open source database (even if it saves the company money) because:
1. He does not want to risk his job if there are problems... if there is an Oracle database, he can just use the Oracle support and tell the boss that they are using the best 2. There is no monetary incentive for the IT professional to switch. If the IT professional would see a benefit to himself that would outweigh the possible problems, there might be a switch, but until then what is the benefit? 3. Some companies just want it to work because they want to do business. Most managers don't want to get involved in database technical issues. They just want results and are willing to pay for it.

Re:enterprises also want

[MouseR]

Disclaimer: I work for Oracle, though not in any DB department (but rather native Mac applications).

One issue with comparing Oracle with Open Source solutions is the lack of feature parity. And By that, I'm not saying OSS isn't good. PostGres and MySQL are pretty good but they currently lack load-balancing and data replication across multiple data centers. Oracle also has PL/SQLwhich is currently (AFAIK) has no (or short) equivalents in OSS DBs. (PostGRE hasPL/pgSQL but apparently is not feature parity.

When a big corporation already relied on these features, it's hard for them to justify making a move to another DB system. Even another that might have similar features.

Re:enterprises also want

[ThreeE]

People that berate managers for making decisions based on financial statements are naive. Let's see how fast you make a decision when it means a zero vs. non-zero on your paystub...

Re:enterprises also want

[plumby]

The one place that the Open Source products have a long way to go is support. Companies don't think mysql and postgres are unreliable, they're just not backed enough. The company I work for could give a rat's hindquarters about TCO -- they just want to outsource their risk so that if something breaks, the CIO/CEO/Chairman has someone to argue with.

That's less about support and more about image.

The quality of the support (such as likelihood of getting someone to be able to fix your DB when it's fallen over at 3am etc) and therefore the level/cost of risk that this incurs is something that should already be right there in your TCO.

What you're talking about is how it will look to your boss (and who you can blame) if things do go wrong. Oracle's support could (theoretically) be 10 times as costly and half as good as some local MySQL support company but if Oracle screw up, it's seen as Oracle's fault not the person that chose them.

Deciding against a supplier because the level of support/liability is not acceptable to you is a good reason to make that choice. Deciding against it simply because you'd look like an idiot if it did go wrong is not a good reason (but it's the one most people seem to take).

0% savings for me

[Crudely_Indecent]

For those of us who can't afford to run a commercial database package, and have been running open source databases from the beginning, this isn't news. MySQL and Postgres are your friends.

Re:0% savings for me

[leamanc]

Which brings up a good point. How many "enterprises" need an Oracle or DB2? MySQL and Postgre, despite their obscure limitations that really only matter to ubergeeks, can work just fine for non-Fortune 500 companies. Heck, they would (and do) work fine for some of the big companies. Most of the small businesses have limited in-house IT, but usually have a guy or two that can learn PHP and tie into an open source SQL with that.

The big boys are the only ones who need the big DB vendors, and even in that case, it's more so they can write off the cost of purchasing the licensing and paying the implementation team. Better to give the money to other big businesses than to Uncle Sam. (Obviously written from a USA-based perspective.)

Re:0% savings for me

[GuidoW]

I don't think this holds true for database systems.

As far as I can see, the big databases like Oracle or Sybase are targeted at big corporations that don't have any second thoughts about hiring a full-time DBA - and as a consequence you will need one to use these products. Okay, maybe not a fulltime DBA, but at least a trained professional (trained for that particular application) to invest quite some time to even get stuff going.

For Postgres, OTOH, you'll just need someone smart with general knowledge of SQL and of the platform it's going to be running on - he needn't even have used Postgres before.

Re:0% savings for me

[glwtta]

Open Source Databases are only free if you don't value your time.

Dude, have you actually used both Postgres and Oracle? There is not an entity on this earth less respectful of my time than Oracle (well, maybe ClearCase) - the thing is an absolute nightmare to administer. Sure, it needs the complexity because of its advanced scaling capabilities; but most of us are not amazon.com, and never will be.

On the other hand, the administrative overhead of running Postgres is damn near 0% (MySQL is a different story entirely of course).

Sure we are a small company, and only have under a TB of data in our databases, but there are a lot of companies in the same position who shell out ridiculous amounts of money for Oracle (only for the name-brand, nothing else), and then someone ends up stabbing themselves in the eye in frustration (might be a slight exaggeration). Or else pay for a full time DBA; I've worked for a company with 3 developers and 1 full time Oracle DBA - that's just nuts.

Re:0% savings for me

[Anonymous Coward]

Some clarifications about Oracle.

[1] Has been fairly easy to install since version 9.
[2] There is a free version of Oracle which is extremely simple to install on linux or windows and works out of the box w/out any DBA. (Limitations - 1 CPU, 1 Gig memory, 4 Gig user data ... fine for many small apps.) Includes backup/restore scripts.
[3] "Standard One" edition (which is fine for many applications) costs 5K per processor, not >100K.
[4] Comes with a great 2-tier web development framework installed by default. (Formerly called htmldb, now Oracle application express.)
[5] Self-tuning and smart defaults has diminished need for DBA on smaller apps.
[6] PL/SQL is a very strong procedural language for logic close to the data and for data processing in general (e.g., exception handling, pipelined table functions, external tables, merge cmd, exceptions into clause, sql xml functions, sql analytic (e.g., rank(), cust_dist() ) functions, etc.) You can implement complete systems with only the database platform.
[7] Oh, should I bother mentioning multi-versioning concurrency control (non-blocking reads) or flashback query/table/database?
[8] Strong pl/sql api's (e.g., job scheduler, compression, http, email, encryption, debugging, ldap, xmltype, file, etc.)

I don't work for Oracle. I am just a developer whose used it for a number of years. It is an incredible piece of software engineering with over 30 years of maturity. The SQL optimizer is amazing piece of computer science. Open source databases are real and may catch up some day. But I prefer to work with the best.

And the truth is: it isn't necessarily expensive or hard to use.

How is this a surprise?

[jimicus]

Unlike a desktop PC, any serious database installation demands a serious database and at least some professional expertise, even if it's just "sysadmin of many hats, one of which happens to be dabbling in the database".

Therefore personnel costs probably don't vary that greatly. This only leaves two costs: the application and the database itself. Generally speaking, the business will choose the application first and the database second (or they certainly should do), which means the cost (if any) of the application falls under the heading of "we've got to have it so it really doesn't much matter how much it costs, within reason".

This leaves the backend database, assuming there's a choice in the matter (not all applications support all databases, despite SQL being nominally independent). In such a project, licensing that is about the only really variable item in the list.

Re:SQL Server = Almost Free

[sstern]

An experienced DBA can set up a new installation in a couple of minutes.

An experienced DBA can set up MySQL with many useful tools in a matter of minutes, too. And you can pay him more because you're not paying Microsoft.

The important question is whether you've bought an application that requires a specific database. As I look at various enterprise apps, they don't come stand-alone, but come in versions tailored for specific databases. If more people said "Do you support MySQL?" we'd see greater use. Chickens and eggs, but you have to start somewhere.

Re:SQL Server = Almost Free

[profplump]

First, if you're afraid of command-line work, you aren't running Oracle, or anything else in that class.

Second, unless you're doing something out of the ordinary, simply installing mysql or postgres in the same way you usually install programs (be that apt-get, rpm, MS Installer, etc). is all you need to get the database up and running. The same is true of the GUI tools to manage the database -- the Windows installer for postgres includes PgAdminIII in the same package as the database itself.

I'm not bashing MS SQL Server, but let's not pretend it has some magical ease-of-use that doesn't exist in other packages.

Re:SQL Server = Almost Free

[kpharmer]

I actually found sql server to be quite expensive - from licensing (which was running > $80k for a 4-way on enterprise edition) to labor.

The lack of command-line features meant that many operational activities that could be automated required a dba to manually do the job via the gui. And lets not even talk about how you had to completely recreate DTS packages when promoting them from dev to test to prod...

So, there are labor savings that you can get on sql server vs oracle, db2, postgresql, etc - but the lack of a command line interface wasn't a driver in my experience.

Cheap Does Not Equal Reliable

[Anonymous Coward]

A cheap database is not necessarily a reliable database. Since an open-source (OS) database is not owned by any company, there is no final person who takes responsibility for any error in the database. By contrast, DB2 from IBM is a commercial database, and IBM guarantees error-free operation.

An OS database is appropriate for a non-commerical or minimally commercial environment: e.g., Slashdot. A commercial database like DB2 is appropriate for a heavily commercial environment like a bank.

Which database would you want to power your checking account? 'Nuff said.

product support is over-rated

[kpharmer]

> People are far more dependable when they're working for money than for charity.

not when they suck - which they frequently do when working on product support teams.

yes, I'm glad that I'm working with supported products - but I also avoid calling them like the plague. It is very much a worst-case scenario.

Re:your mileage may vary

[Anonymous Coward]

So, a $100k oracle lisense running on a $100-200k 4-way (i know, assumes discount) will out-perform postgresql (free) on a 16-way ($1m) in many ways.

200k for a 4 way ?
1m for a 16 way ?

A 4 way or a 16 way what ?

Check out X4600's from Sun - 16 Opteron cores and 32G RAM for less than 55k USD

Where are the test results?

[msobkow]

Do you seriously think any CIO with a functioning brain cell is going to go with free unsupported software when they can't even find a single reference to such databases from any certified performance evaluation companies or organizations?

The downtime cost of one single failure in a five year period for a mission critical system can easily run 100 times the cost of a commercial product with support. Only bean counting fools risk their entire business without properly assessed risks and disaster recovery plans.

Not having someone to source the recovery of the smouldering crater that was your data center is a huge issue.

Re:your mileage may vary

[quantum bit]

If you're paying a million for a 16-way you're an idiot. Partition your data and cluster cheaper machines. That's what Google does.

Re:SQL Server = Almost Free

[Pfhreakaz0id]

Umm, you do know that sql server has it's own scheduler for running scripts and stuff? SQLAgent? Obv. I guess not.