A question I touched on briefly ...

this person at Interop news goes into, in depth. The business person in me had difficulty understanding the acquisition. Sun didn’t have a missing technological niche the MySQL filled. MySQL had all the standard problems of growing a business, compounded by the Open source revenue model, which effectively eliminates distribution/redistribution revenues. There are many commercial outfits likely skirting the edge of legitimacy using MySQL in their shipping supported closed source products. They skirt the issue by claiming that the OS ships with the distro, so they don’t care, and won’t buy a commercial license. Why should they, its free … right? Well, no. And this is one of the great unsolved issues in open source business models. You effectively foreclose redistribution revenue. So how can you recover revenue for your efforts and pay your ongoing costs … like people’s salaries … never mind making a profit, which the shareholders/owners really would like you to do?

The only way to avoid this quagmire is to not distribute your software. Sounds odd, but if you distribute it, then anyone can use it. If you make it a service …. This may be, BTW, one of the huge drivers of the SaaS market. Back to that in a moment. What open source does for customers is to lower their cost of acquisition, and increases their control over the technology deployment and usage. You no longer have to consider “can I deploy this because I am not sure I have enough licenses” … you can just deploy it. This massively lowers the cost of ownership as you can amortize this cost much farther and wider than other models. So you can “buy” one license of Redhat, and you can, legally, deploy it everywhere, give it away, and so on. Well, you are bound by copyright rules, as the code does contain Redhat copyrighted material, so, as long as you follow the letter of that law, the GPL lets you deploy as many copies as you wish. This makes RHAT’s job harder. They have to show value for what you get in order to entice you to pony up for subscriptions. Here, they make officially supported patch sets, and package things nicely to make your job easier. At the end of the day, you are paying RHAT for packaging and support of the packaging. Not for the underlying bits. Which will often include “commercial” products such as MySQL. So RHAT is doing what is legal, effectively giving away MySQL as part of its offering. It will charge for, and handle basic support for that package. Which deprives MySQL of additional revenue. Adds to RHATs revenue. You see the problem? If everyone gets to redistribute the code, for free, without encumberence, then you cannot tie up your value in the distributed version. The real value has to be elsewhere, and not given away. There is the little issue of that revenue model, that you need to make work. What you can do with the wide scale distribution is effectively freeze out competitors. Whether or not your product is better, if it is “free”, well, you might capture most of the (niche) market for it if you push it early enough. They 80-20 rule. But then you have to continuously update it. And this requires, for many products, somehow, paying salaries. Not always. Salaries are sometimes decoupled from the revenue they might bring in. This is usually the case with universities and related groups that seem to really like open source for everything. They are willing to commit graduate students/post-docs to coding, paying low salaries on soft money, effectively decoupling the value->revenue chain. Which means that, as long as their benefactors continue funding the work, products will continue to be pushed out. A great example of this in the HPC market is the cluster distribution Rocks. It is tied to Redhat, which I am sure they are thinking in retrospect, is not a great idea. But the NSF happily continues to fund this work, so they can keep at it irrespective of their “revenues”. One might make a similar argument of the Microsoft HPC initiative, that the likelihood for sustained growing and profitable (significant) revenue streams is extremely low … it simply does not make business sense. But this decoupling happens, products are pushed out, and in the case of Rocks, the commercial competitors are effectively frozen out. New cluster distributions need to be free (as in zero cost) to get attention. This is another reason by the Microsoft HPC initiative as it is constructed has been in trouble from the get-go. It doesn’t matter if you are better than them, if you are not free, you won’t get the eyeballs. Considering the damage such a thing would do to Microsoft’s other business models, I would guess that they will never arrive at the needed conclusion. But that is a divergence from the discussion. Exactly how will Sun make money with MySQL? This is the issue. Most of the well understood revenue paths are frozen out. They are frozen out by the license. This license has tremendously lowered the cost of IT and development systems. Many/most of the startups I am aware of developing products are building them on or targeting Linux, as it adds nothing to the cost of the units they ship. Read that last sentence again. It adds nothing to their BOM costs. Shipping MySQL (as part of the distribution) adds nothing to the BOM costs. It gives you a huge number of MySQL users. Your job, as MySQL, and now Sun, is to convert these users into paying customers at a higher rate. So how do you do that? That is the catch. You can offer a non-OSS version of your code base… then you can’t integrate OSS patches (what Don MacAskill and others complained about) … You can offer online DBs (ala Amazon and others) and DBaaS. You can offer extended service/support/training (ala RHAT). It remains to be seen whether or not Sun can make this work. The authors note that Sybase is valued at 2x revenue is telling. This means that MySQL is ~$80M, which means Sun overpaid by 12.5x. Sadly, I think this may be right. Sun is going to need to be really creative to make the MySQL acquisition pay for itself. I don’t necessarily believe they will be able to. My thoughts are that it was more of a tactical acquisition, to give them control over a property, without really purchasing a business. Its what they do with the property, and how the leverage it that will count. It will be interesting to see what happens.