I think mySQL should now probably come out with a licensing scheme that somehow make’s scenario one customers pay.

i.e IF Amazon says that it will provide support contract for mySQL instances. it should be prohibited unless there is a back back to contract with mySQL substantiating it.

So essentially - resellers of mySQL should be prohibited.

if you treat EC2 as a customer of mySQL then focussing on volume’s is a losing game . You will have to work like the GE’s of the world and learn how they do their Turbine business. You might give a prototype free but ultimately you will ask your customer to pay and your sales pitch is going to be based on performance and quality. Price performance ratio requires that neither the numerator or the denominator can be zero for any reasonable comparisons.

If you treat Enterprise’s and the traditional IT model as your customer then volume is what you want to go for. Now open source as a business is ultimately a market capture strategy. What happens when you have achieved the market share is the question mySQL needs to answer. You cannot suddenly charge charging for your product. I went to the JBoss website recently after about maybe 6 months. I got the feel that there the Buy word on every page. It is ultimately turning into a subscription based product. Same situation - what does a business do when it has got the market share it wanted.

I disagree with the first comment here - any vendor in the community could fork and screw up that vision (consider that MySQL is not like Apple on IP).

Also, consider the volume of the data or applications built for other database products already. Everyone keeps talking about the volume of MySQL instances, but really think of the volume of paid for database solutions. If I can get Database A under GPL from Sun who I’ve never bought a database from… ever, or I can get Database B from Oracle which every vendor has certified to and the Oracle Database is already all over my datacenter… am I really likely to switch?

My point? The volume of paid, production data out there is locked away in Oracle/DB2/MS SQL. It’s very difficult for companies to move that data. At best, MySQL is targeted at new data, new applications and that will be extremely difficult to capture enough share of to ramp up and provide returns on a billion dollars of cash invested in Jan 2008. The financial model does not work at $1B cash, even if you ascribe to the theory that volume eventually drives participation in the market, driving greater revenue opportunity, other investment options could have offered higher returns over the next 5 yrs…

Oh, and MySQL already screams on any other hardware gear. Most customers don’t care about performance unless it’s at least 10x. I don’t see how Sun’s processor platforms could ever achieve that (consider x86 and also the POWER6 for instance already kill Sun’s processors). Without Niagara/SPARC to offer a performance advantage, that relegates them to x64 based hardware from AMD and Intel that every other vendor already has. Sun could try filesystem tricks, but so can anyone else - they have the same code Sun does.

There’s a lot of speculation now, but I think the proof will be 2010 when MySQL is supposed to be accretive… by how much will it be accretive and can that ramp to provide any sufficient payback that would justify any credible financial model… only time will tell. Given they’ve been at this for quite a few years already and not made it work… it will provide an interesting challenge.

I must admit, this is a very valiant effort to recapture the award from Jeff Gould! A third recount has begun.

During the dot-com hay days, there was one platform everyone used their hard won VC dollars to backend their online flower boutiques with; Oracle running on Solaris. There was a reason for that, aside from insane sales tactics and a surplus of available VC dollars for licensing (and that it meant you got to throw Oracle and Sun up on your “Valued Business Partners” ppt slide). Sun and Oracle got together to make that platform *scream*. They locked engineers in a room together, told them to play nice, and printed out the subsequent cash.

In a short amount of time, MySQL will similarly scream on those cores within cores we keep hearing about. Jon’s going to lock that MySQL team in with his existing engineers, and in a year or so, not only is he going to have the software sales associated with any MySQL conversions, he’s also going to have the hardware sales associated with being the premier MySQL platform for all that beautiful open source market share. It’s right out of the Apple playbook (also, fundamentally a hardware company), and it’s fucking brilliant.