Technology -> Data Management
By: Robin Bloor, Chief Research Officer, Bloor Research
Published: 25th January 2005
Copyright Bloor Research © 2005
The recent announcement that Pervasive was throwing its weight behind PostgreSQL, as an open source offering, was interesting for two reasons. First of all, Pervasive already had a successful database product – Pervasive SQL - which has strong traction among ISVs. Pervasive regards its version of PostgreSQL (Pervasive Postgres) as complementary to this. Technically that's true, as Postgres is a true object-relational product whereas Pervasive SQL is a died-in-the-wool relational database product designed for speed and not object manipulation. Pervasive has, of course, enhanced PostgreSQL by adding some of its installation and administrative tools. Whether Pervasive will be able to make a business from Postgres is a different question – which brings me to my second point.
The Open Source database market appears to be getting crowded. This is a new phenomenon for Open Source. Usually there is only one product in play from the Open Source community. Admittedly there are exceptions – the most obvious being the existence of two desktop GUIs, Gnome and KDE, but in general you'd think that the nature of Open Source itself would naturally lead to products coalescing rather than competing.
This hasn't happened with database and there is a good reason for it. Databases are fundamentally data retrieval engines and the value of any given product lies to some degree in the architecture of its engine. You can't mix engines. They're built from the ground up. Also database functionality and the database engine are often strongly intertwined.
Anyway, there are now no less than 5 significant database products in the Open Source market. They are:
And if you like you can also get Sybase's Adaptive Server Enterprise (ASE) database for Linux, which is not Open Source but comes at no cost (if you limit yourself to a single-processor machine, a maximum of 5GB of data and a maximum of 2GB of RAM).
So is this confusing or is it confusing? There is no doubt that it is confusing, but there may be a good reason for choosing any of them depending on what you want to do. If you are looking for Enterprise capability then Ingres is a good bet. If you want object-relational then try Postgres. If you need a web site database then MySQL looks like the candidate. For Java try Cloudscape. And if you want hot transaction performance then FireBird is probably the candidate. However there is also a good deal of overlap in capability.
Currently the market leader is MySQL, but a recent survey of 400 database developers from Evans Data Corporation (Winter 2005 Database Development Survey) showed FireBird as the most used open source database for Enterprise applications and also the most used for single purpose applications. Indeed according to Evans Data, MySQL and FireBird are now roughly equal amongst Open Source users. However, that's the situation before CA (with Ingres), IBM (with Cloudscape) and Pervasive (with Postgres) put their marketing weight behind their respective products.
Can so many Open Source database products survive? I personally doubt it, but I may be wrong. They may all generate a healthy and sizable submarket that justifies their existence and, with Open Source products, it doesn't really matter how big the user base is, as long as it is big enough to generate the support network necessary to keep the product evolving and relevant. (CA is already reporting that it is making a profit from its open sourcing of Ingres – which probably means that Ingres has a brighter future as an Open Source product than it had as a proprietary solution).
Another question worth asking is how all of this is going to impact the proprietary database market. In my view, proprietary database products are going to see a significant decline in revenues in the coming years. Irrespective of which of the above products does well, or continues to do well, this now seems to be an inevitable outcome. With such a range of free database products available, many organizations are going to think twice about paying big database license fees – and having thought twice, choose not to pay them.
Posted: 25th January 2005 | By Cool_man :
From the list of database i think you have left out Daffodil DB whose version One$DB http://www.daffodildb.com/one-dollar-db.html was recently open sourced and is all ready to give derby a run for its money as it is a feature rich offering comparitively.
Posted: 25th January 2005 | By Leandro GFC DUTRA :
The Evans study has been subjected to various qualifications and critiques, it isn't probably a good idea to rely on it.
And Pervasive SQL isn't relational, because the SQL standard violates the relational model big time.
Posted: 26th January 2005 | By Dirk Heinrichs :
You forgot MaxDB, the database formerly known as SapDB.
Posted: 26th January 2005 | By mcrbids :
This is Drivel.
See, an OSS product, to "survive" doesn't need to be "top dog" - it only needs to attract enough developers so that it continues to advance and evolve for new O/S version and releases.
We can have MySQL, PostgreSQL (my favorite), firebird, and whatever else working simultaneously. It's a big world, with lots of fish. To survive, an OSS project doesn't need to catch most of the fish, just enough to continue to grow.
Posted: 26th January 2005 | By Michael :
You forgot SapDB/MaxDB
Posted: 26th January 2005 | By Luca Veronese :
It may also happen that Oracle, feeling the pressure from Open Source products, will start charging less dollars for their database licenses.
We will see if this will have the same market effect as Linux vs. Microsoft Windows (i.e. lowering the actual price of MS Windows).
This effect may also benefit proprietary RDBMS users.
Posted: 26th January 2005 | By halibut :
How this affects the proprietary database vendors depends on how many db application providers are willing to go to the trouble to port / certify their applications on these open source - platforms. Most clients can't afford to support more than one or two database platforms, so they are looking for a single database platform that can run the apps they have developed in-house and also bought off-the-shelf. My point is that the proprietary databases are not going to be hurt tremendously by this trend, and/or we will see one or two of these OSS databases take the lead in getting "certifications" from db app vendors.
Posted: 26th January 2005 | By Matt :
To determine which OSS database system will survive, wouldn't a 'easiest to migrate from proprietry DB' value have to be assigned? If PostgreSQL (or MySQL for that matter) use as a replacement for Oracle (or SQL Server) was virtually swap out / swap in then surely that OSS database would become more and more popular. This is what my management want to see, especially for the high cost multiple CPU licences that we are paying Oracle. The other factor to take note of is the use of DB2 on the mainframe. I know of two very large companies who have reduced cost with SAP by changing to DB2 on the mainframe for their DB with SAP.
Posted: 27th January 2005 | By Christof Wittig :
You also forgot to mention db4o (www.db4o.com), the open source object database, native to Java and .NET - with more than 100,000 downloads and commercial customers of the likes of BMW, Hertz, Bosch and INDRA.
You are right that there is usually only one player for each category - but an ODBMS is a different animal than an RDBMS, so is the target market: embedded DBMS instead of enterprise-class data warehouses.
Posted: 27th January 2005 | By cdmiller :
Don't forget Sleepycat's Berkley db or the up and coming sqlite.
How "crowded" is the open source database field, Freshmeat has 311 database server/engine projects listed:
The messages above were all contributed by IT-Director.com readers. Whilst we take care to remove any posts deemed inappropriate, we can take no responsibility for these comments. If you would like a comment removed please contact our editorial team.
We automatically stop accepting comments 180 days after a post is published. If you would like to know more about this subject, please contact us and we'll try to help.
Published by: IT Analysis Communications Ltd.
T: +44 (0)190 888 0760 | F: +44 (0)190 888 0761