Red Hat comes out swinging – will it be enough?

While Red Hat has been careful to measure its words when speaking in public about Oracle (still a valued partner, according to the "Unfakeable Linux" page on Red Hat's web site), it is in no mood to take any lip from pesky Linux competitor Novell. In a rare public interview, Red Hat's senior lawyer slams Novell and goes so far as to predict that its rival will be out of the Linux business a year from now.

Red Hat management's line is that neither the Oracle announcement nor the Microsoft-Novell deal affect the fundamentals of its business, and that its rivals are merely engaging in empty theatrics with the purpose of driving down the price of RHAT stock.

I believe Red Hat management is right on both counts. But they still have to make their case in public. Here it's the General Counsel putting out the message, but no doubt we'll hear from the CEO or CFO at the Goldman Sachs conference this week. In many ways, Red Hat is being punished not for any truly fundamental change in its business, but for its egregious failure to communicate clearly to the market the nature of its value-add. Management needs to explain to investors why the company is and will remain indispensable to the Linux ecosystem. While they're at it, they could also do a lot better job quantifying the growth prospects of that ecosystem compared to the competing Microsoft universe (it's much easier to get quantitative data about the size of the Linux market out of Microsoft than out of Red Hat, and that's a problem).

On the fundamentals with respect to Oracle (I'll come back to Novell in another post), it looks like the price gap between the Red Hat and Oracle support offerings will quickly tend towards zero, especially for larger customers. Oracle has several clever little gotchas in its pricing scheme that make it less advantageous than a first look might suggest. For example, with Red Hat you can mix and match between expensive and cheap levels of service on your various Linux servers depending on the needs. But with Oracle, if you want the premium version you have to get it for every one of your servers supported by them. Similarly, the difference in the content of the service offerings is not as large as Oracle claims, and will quickly disappear in a de facto sense as Red Hat covers whatever bases its clients want covered.

Then there is the issue of Red Hat's ISV and hardware certifications. Red Hat is currently taking the very legalistic stance that the bug fixes which Oracle proposes to patch onto Red Hat code will break these certifications. This is true, but how frequently or how badly the Oracle patches will break ISV apps is at this point anybody's guess. It may well turn out that problems are few and far between. Nevertheless, it seems a no-brainer from a CIO's point of view that if the price for RHAT and ORCL support is the same (which it will be), then there is no incentive to switch and incur even a slight risk of breaking an existing application. After all, why make even a trivial investment of effort if the ROI is guaranteed to be either zero or negative?

But Oracle will counter with the political leverage it undoubtedly exercises on certain clients (although it has been widely reported that only 5% of the Red Hat server installed base is actually running an instance of the Oracle database). Putting the enterprise edition of the 10g database on a four socket server will cost you $160K in upfront license fees prior to discounts plus 22% of that per year in maintenance. Compared to that, saving $500 bucks per year per server off the list price of Linux support (and maybe nothing after RHAT discounts) isn't very motivating. But if Oracle is willing to throw you a big discount on your database license, then that's another story. Realistically, though, why would Oracle want to do that? Larry doesn't like Red Hat (and he has good reason, as we will see in the next point). But he's not going sacrifice his database margins just to punish them.

Finally, we come to the likely hidden strategic motive for Oracle's move (beyond the obvious one of accelerating the spread of Linux in order to bar the advance of Microsoft's very dangerous SQL Server 2005). In my reading, Oracle is not aiming to kill Red Hat here – after all, they are the source of the leading enterprise Linux distribution, the very distribution Oracle now values highly enough to offer full-scale support for. No, but they are sending a shot across Red Hat's bow. The message is: you managed to snatch JBoss out from under our nose, but don't you dare go out and try to fill out your open source software stack by acquiring MySQL too (MySQL is the leading open source database, and a distant but real threat to Oracle's database franchise). If you do, we'll get really mad, and we'll do something that hurts you even more than what we've already done. So there.