Robert Greene, Leon Guzenda and Rick Cattell on Sun Microsystems acquisition of MySQL.
On Wednesday Jan 16, 2008 Jonathan Schwartz, Chief Executive Officer and President, Sun Microsystems, Inc., announced in his blog that SUN is acquiring MySQL AB.
On 26 February 2008 Sun Microsystems, Inc. announced it has completed the acquisition of MySQL AB, for approximately $1 billion in total consideration.
Kevin Harvey, Chairman of the MySQL board of directors told InfoQ that there were two main drivers behind Sun’s purchase of MySQL ” it solidifies Sun’s role in the Web 2.0 datacenter, and it also confirms Sun’s position as a leading provider of open source software.”
I have asked three of our experts, Robert Green, Leon Guzenda, and Rick Cattell a few questions on this. Robert is responsible for defining Versant’s overall object database strategy, Leon is responsible for the Objectivity object database strategy. Rick worked for several years at Sun Microsystems, and now he is an independent consultant.
Q1. What does this announcement mean for the database market in general? and specifically will it have any impact on the object database market in your opinion? and if yes, how?
RCG> I think this announcement means that companies who were concerned about putting MySQL into their enterprise environments will now rethink things. If Oracle was not concerned before about the MySQL threat, it ought to be now. It is interesting to watch as people are beginning to pay for these products, previously perceived as “free”, in the form of services and value added capabilities(in MySQL’s case, better tooling). I don’t think there will be any direct impact to the OODB market other than the perpetuation of changing attitudes that what counts most is using the right tool for the job. It’s that change in attitude that’s having the greatest impact on the OODB market.
LG> MySQL is a conventional RDBMS built and sold using the open source model. Sun has traditionally been vendor neutral in its approach to DBMS sales, partnering with whichever DBMS company a joint customer expressed interest in. They have always had a strong partnership with Oracle, for instance. As Oracle also sells its product on IBM hardware, competing
directly with DB2, these partnerships are interesting in their complexity. Will Oracle shift more of its attention to sales on HP
equipment rather than Sun’s. I doubt it. However, there will undoubtedly be some pressure on Sun’s sales people to work new deals that can be 100% handled and supported by Sun.
I can’t remember a situation where Objectivity/DB has been in competition with MySQL as we tackle completely different kinds of application, so this won’t affect us directly. Likewise, our customers almost all find us without Sun’s help, so I don’t think it matters that much to us.
RC> Sun has had a good adoption rate on its open source offerings: their application server, Open Office, Java, and so on. I believe that the MySQL acquisition was exactly the right move for Sun at this point, and also will be a big benefit to open source users.
The acquisition will be good for open source users because Sun will push MySQL innovation in new directions, Sun will provide long-term stability for MySQL, which has been under attack from Oracle (who recently acquired both InnoDB and SleepyCat, the “engines” for MySQL), and there will be synergy and benefits between MySQL and Sun’s current open source
offerings, e.g. the application server and development tools.
The acquisition was exactly the right move for Sun because unlike Microsoft, IBM, and Oracle, Sun did not have a strong database component in its software stack. Sun’s software stack is open source (again, the right move I believe). Unlike Sun’s current database offerings with PostgreSQL and Java DB, which are only strong in narrow markets,
MySQL has a very large following in a wide variety of applications. MySQL thereby gives Sun a complete software stack with “best of breed” solutions pretty much across the board. It also allows Sun to tune that software stack for its platform (for example, optimizing MySQL for Sun Solaris, and utilizing innovative proprietary hardware features).
As for the object database market, I don’t see the acquisition having a big impact one way or the other. Object database systems are being used in different markets than relational database systems, for the most part. However, Sun’s obvious support of open source is a “shot in the arm” for open source databases. Also, Sun’s Java Persistence API and
the adoption of object/relational mappings is a boost for object databases, because these allow object databases to be more easily and naturally substituted for relational databases in application servers and web servers. Sun will likely do some tuning of MySQL with JavaPersistence.
By the way, I recently left Sun to do independent consulting as Cattell.Net, so the opinions and speculations I express here are purely my own. But as I mentioned, I believe the MySQL acquisition was a great move, so I remain positive on Sun’s future if they play their cards right with the MySQL technology and customers over the next coupleyears.
Q2. Schwartz in his blog says “….customers confirmed what we’ve known for years – that MySQL is by far the most popular platform on which modern developers are creating network services. From Facebook, Google and Sina.com to banks and telecommunications companies, architects looking for performance, productivity and innovation have turned to MySQL.”
Will it change anything in this respect?
RCG> Well, I would hope the Schwartz believes in his message. The fact that Sun spent 1B for MySQL would suggest that he does not believe his perceptions will change for the worse, but I would hedge with a quote by Niels Bohr, “Prediction is very difficult, especially of the future”. The key is to provide value, so far MySQL has done this well, whether or not they have “peaked”, time will tell.
There are public declarations about where the technology succeeds and where it begins to break down, if they want to expand, there are known issues that must be addressed. The future is wrought with challenges due to unbounded data growth, coupled with concurrency and complexity. In the future, other drivers like “Green” abilities will outweigh the ability to simply get the job done. Almost anything can be forced to work, but when you have to decide between something that works on 1000 servers or 400, your decisions will be heavily influenced by these other factors.
I would predict, in the future, value will be driven by “Green” technologies, much like they have in the semi-conductor industry over the last decade.
LG> It might for ODBMSs that are targetting traditional IT applications.
Q3. Schwartz in his blog says: “The adoption of MySQL across the globe is nothing short of breathtaking. They are the root stock from which an enormous portion of the web economy springs”
Is this specific to MySQL or else?
RCG> I think the open source movement as a whole is the stock, MySQL is simply one of the branches, certainly big enough to hang a hammock without fear of breakage.
LG> Apart from the usual pressure to use the vendor’s technology, I can’t see Oracle, DB2 or SQL Server shops suddenly switching everything to MyQL because Sun now owns it. I think it more likely that MySQL users will be pressured to switch to Sun’s hardware offerings.
RC> I believe Schwartz is right: the adoption of MySQL has been incredible, particularly among the fast-growing web companies. This is another
reason that the MySQL acquisition was a smart move: it gives Sun an opening into these companies. Sun has suffered somewhat because these fast-growing companies have generally not bought Sun hardware,
support, or software. Sun has only done well with the traditional and more conservative “enterprise” companies. Now Sun has a complete open source software stack, gives customers a choice of operating systems, and offers competitive hardware with both Intel and SPARC architectures.
Sun is now well aligned with the fastest-growing sectors of the Internet market.
You might question how “adoption” translates into dollars for Sun, since open source is free. But I believe Sun is in a good position to monetize widespread adoption of its software stack, through support revenue, upgrade revenue, and synergy between software and hardware sales.
Q4. Schwartz in his blog says: ” So what are we announcing today? That in addition to acquiring MySQL, Sun will be unveiling new global support offerings into the MySQL marketplace. We’ll be investing in both the community, and the marketplace – to accelerate the industry’s phase change away from proprietary technology to the new world of open web platform”
What`s the meaning for the open source community in the database market?
RCG> Maybe this is supposed to be a trick question. I think the meaning is the same as it is to other software markets. The sum of the constituents that subscribe to it’s use and adoption. Software must increasingly provide value to
compete, even free and open software. If there is no appreciable value, then it will have no constituency. The key is to figure out where you are in that value curve and how best to drive adoption given your particular situation.
LG> Let’s not forget that Sun moved to the open source model as its own efforts started to lag the faster moving community. While this matters a lot in some highly dynamic and emerging markets it can even be a problem in enterprise applications. Red Hat was changing so rapidly at one point that equipment manufacturers and rigorous IT shops were having problems
achieving a stable base, so red Hat introduced the more pricey Enterprise Edition. Sun is probably aiming to make its money from services and bundled sales. Not all open source offerings have become commercially viable, but MySQL is a notable exception
Q5. Schwartz in his blog says: “The good news is Sun is already committed to the business model at the heart of MySQL’s success -”
Is MySQL business model usable/adaptable also for ODBMS? How?
RCG> As stated above, it’s identifiable value that is important. The MySQL business model for the sake of the MySQL business model is a non-starter. If you have a technology that has non-commoditized value, there are other equally viable business models. The ODBMS company I can think of that most closely matches the MySQL model is db4o, and they have a database value which is highly commoditized, so I guess that business model makes sense for them. Some of the other ODBMS companies have highly differentiated value, so they do not depend on a MySQL like business model. So, it appears that many business models work.
Which one works best is another question altogether. Ultimately, a business model has the goal of returning profits to it’s owners and shareholders in a competitive landscape. So, what is the best company/business model, one that has 50% of a 10B software market and only earns 60M/yr for shareholders ( at a loss ) or one that has .005% of the market and earns 25M/yr for their shareholders at a profit? Again, I think the important point is understanding where you are in the value curve to help establish the business model that makes the most sense. I think this is one of the places that Sun has fallen short in the past. They have been trying to do the MySQL business model, but have failed to really understand where their various offering reside in the value curve. I could be wrong or rather that could be changing – perhaps they understand it very well and it’s just more complex than a first glance. Perhaps software is the commodity and hardware is the value add and they are looking for MySQL to be the catalyst for adoption much like Hibernate was for JBoss.
LG> There are currently about a half dozen ODBMS products sold with conventional licensing and a similar number of open source ones. As in the early days of ODBMSs, where there were about three times more products than the market could sustain, I doubt that many of the open source ones will survive in a crowded, highly specialized market. RDBMSs
need a lot more support, e.g. for database administration, than ODBMSs, so the split between license sales and services is dramatically different for the two technologies. The open source ODBMSs will need to spread out into applications to earn significant revenue from services. At that point, if they’re in the wrong vertical they’ll be competing head on with the big players.
RC> An excellent question. Many companies have admired MySQL’s success and wondered how to emulate it. In my mind, MySQL is the only company in the last 20 years to successfully challenge the domination of Oracle, IBM, and Microsoft in the database market. Other database companies have failed, have been acquired, or have been relegated to a smaller niche
Object database companies have generally been in the last category. Partly that was a matter of timing: object databases came out too early to “ride the wave” of the Internet, Java, and open source. Customers feared compromising the integrity of their databases using an “unsafe” C++ object database, and object databases met stiff competition from the
big relational players, both in marketing dollars and in inertia behind existing relational database installations.
I see that changing somewhat going forward. Although I think it’s too late for a new database contender to ride the “open source” wave in the way that MySQL did, and I still don’t see object databases challenging relational databases in mainstream markets, I do see that an open-source Java object database system could grow significantly, especially in
applications where relational databases are not well suited.
Q6. Specially to your company, do you see MySQL as an example you wish to follow?
RCG> Absolutely! The first company that comes along and offers 1B to acquire Versant, we will accept the deal on behalf of our shareholders Seriously, to remain viable, companies must constantly consider where they reside on the
value curve and adjust business models accordingly. Again, the future is especially hard to predict, should Versant decide it is prudent to switch business models, we would most certainly inform the public as is required by any publically traded corporation.
LG> No, not for pure ODBMSs. It will be interesting to track the commercial progress of db4objects, objectdb and other open source ODBMSs.