Though I'm not sure, I think I had heard about this deal earlier, but I had no idea of the significance this has:
BusinessWeek: A Baby Database's Chance to Grow Up?
To date, though, MySQL has been viewed mainly as a cheap database for running Web sites and as relatively unsophisticated compared to the whiz-bang wares of the database Big Three, Oracle, IBM and Microsoft. Further, MySQL was never seen as an apple-cart tipper on the order of Linux. But Mickos and his minions served notice to the database sector on May 27 when MySQL announced an alliance with German software giant SAP. Then on June 3, MySQL announced a $19.5 million venture-capital financing round including marquee Silicon Valley VC firm Benchmark Capital.
Combined, the two developments could give MySQL much needed momentum. SAP, which racked up $7.4 billion in 2002 revenues and net income of $597 million, is the top dog in complex business software. It claims that as of April, 2003, it controlled about 54% of that market, up from 50% at the end of 2002 and well ahead of rivals Oracle, Siebel Systems, PeopleSoft, and J.D. Edwards (the latter two have announced a merger deal in which PeopleSoft will pay $1.7 billion in stock to acquire Edwards). So the SAP deal gives Mickos the blessing of one of the most dominant companies in enterprise software, much the way IBM gave Linux a seal of approval that more than anything else helped the Penguin people crack the corporate operating system market.
I don't quite understand what SAP's software does. I knew about SAP DB, their underlying database, but thought that was their main product. I didn't realize that their real product was super-enterprise business software to do something. But what's happening is that SAP is basically giving their database code to MySQL, MySQL will improve at a faster pace by integrating the SAP DB code, and MySQL will become the software that is intended to underly SAP installations.
The shadow play here is SAP's long-held desire to commoditize databases. Most of SAP's customers now install Oracle, IBM, or Microsoft databases as the underpinnings necessary to run SAP's business software. That means SAP customers have to pay extra for a database license, often thousands of dollars per server. Compare that to annual support costs ranging from $1,500 to $48,000 for MySQL and free downloads of the software if customers want it.
The current setup of separate database and enterprise software is an additional barrier to SAP's sales. It also means it has less control over its customers. That's precisely why SAP has long offered its own SAP DB product, which is more sophisticated than MySQL but still lags behind Oracle, IBM, and Microsoft in terms of features. Three years ago, in hopes of catching the open-source wave, SAP even tried revealing its database's source code. But that didn't spark many new sales nor did it attract a community of interested open-source developers.
In this light, SAP's deal with MySQL is a natural. Under its terms, the German company will pass development of SAP DB to MySQL. The Swedish company then picks up commercial rights to SAP DB. Mickos plans to merge SAP DB's code with that of MySQL. And he hopes to incorporate heavy-duty computing features into MySQL's next release.
If he's successful, that could save MySQL years of development time and immediately put it on competitive footing with the big guys for lower- and middle-tier types of commercial database installations. "We're typically used in Web sites and departmental applications, but we're not the database for a business application in the enterprise. And we want to change that," says Mickos.
So, this deal is extremely significant. It won't fully play out for years, but MySQL is going to get a lot better and become a lot more important in the business market in the years to come.
Update: Read what Jeremy has to say, and check out the press release for the MySQL and SAP deal as well as the press release announcing the VC funding MySQL recently got.
Saturday, June 14, 2003
Permalink
SAP is enterprise resource planning (ERP) software, like Oracle 11i Applications' ERP product (no, this is not a new version of the database -- that's still at 9i), PeopleSoft, J. D. Edwards and Microsoft Great Plains.
Essentially, it involves recordkeeping of transactions, sales, purchases and stocks, and online analytical processing (OLAP) to figure out trends and patterns for marketing and management folks. It can also include human resources management (HRMS) and financial planning for the corporation. The better ones also allow for customer relationship management (CRM) whereby the best customers (highest volume, highest frequency, loyalty, etc.) can be identified so that marketing can target deserving customers in different ways than others.
If I remember correctly, SAP's the #1 ERP provider. The business pages these days are reporting on the high-profile hostile takeover attempt by Oracle Applications of PeopleSoft -- this is all about getting Oracle 11i ERP applications to dominate over PeopleSoft, so Oracle can get a larger bite out of the ERP outsourcing market.
Every ERP application makes heavy use of databases as the data store, as well as for effective OLAP -- it takes quite a bit of associative querying to be able to get useful statistics -- so this is pretty important for MySQL.
One does have to wonder, though -- MySQL is a pretty awful database as far as concurrent access goes, since its support for things like transactions, fine locking, triggers, stored procedures, etc. leaves a lot to be desired, and ERP applications are designed to do lots of concurrently-running database queries, insertions and updates. SAP DB's a lot better than MySQL (and no, it isn't SAP's primary product -- in fact, SAP R/3 etc. usually use one of Oracle, DB2 or Sybase as the backend, and SAP isn't threatened by the fact that they're giving SAP DB away for free), and there are more deserving databases -- especially PostgreSQL, even if it only runs on UNIX right now.
SAP DB's main issue right now is that its setup and administration tools leave a lot (a lot) to be desired. The documentation is very unpolished, and one doesn't even have an idea of what to do with it 'out of the box'. That's one thing I would expect to see change.