As John West points out over at InsideHPC.com, the FAQ really didn’t live up to FAQ standards … very little was answered, and there are many more questions. But a pattern did emerge, that fundamentally suggests that we may have been (more) right (than we knew).
This acquisition was about MySQL and Java. And other software bits. But no mention of Lustre, and even more important to a larger number of HPC sites, GridEngine.
No mention of VirtualBox, arguably giving Sun and now Oracle similar virtualization capabilities to EMC and VMware. I said arguably … don’t thwack me for it.
Notice that the FAQ (I read it through) appears … well … repeatitive … in its answers. Which say very little by themselves.
Of course I have had a day to think about this, and have come up with some thoughts … which could be way off base, but they are at least some thoughts …
Oracle and Sun certainly doesn’t want Sun’s sales to tank during the acquisition, which is why it is coming out strong in terms claiming integrated stacks, and hardware growth, etc.
The reason for this is simple. You don’t want to scare off current customers, before the close, by telling them “yes, we are going to lop off the hardware and kill it or sell off what can be sold off”.
This gets to the question of whether or not Oracle is really committed to doing hardware. I am not convinced they are.
Their explanation of customers requesting fully integrated stacks from them? Well … my belief of this is … uh … not so much … the case.
From the acquisition, Oracle gets the value in Sun, which as I and a number of others have pointed out, isn’t really their hardware.
Sun, like most other pure hardware plays, didn’t have a clue as to how to go about making money from software. It either massively overcharged for it, or gave it away free. There was really no middle ground.
Sun also has a decidedly non-commodity set of architectures which are competing in a commodity architecture world. Curiously, this is where SGI was in 1999 and beyond.
But it has valuable software properties. The most installed database in the world, MySQL, which it has not really figured out how to make money from. VirtualBox, which it has not really figured out how to make money from. Java, which it has not really figured out how to make money from. …
Oracle knows how to make money on software. They will fix those problems.
But what to do with the software that is not core? Like all the HPC stuff? Lustre which is orthogonal to OCFS2 (the latter being in the Linux kernel and relatively well supported, and competitive to Redhat GFS), GridEngine (which is reasonably good, and competitive with LSF from Platform, and PBS from Altair, and to a degree with DataSynapse and others), Solaris (which is on the fast decline in all but a few locations).
I think Oracle could make a play in financial services by going after DataSynapse and others with GridEngine, though they would need to rework it a bit. But fundamentally they would need to tie it back into an Oracle product or two to make it work as a market. They could purchase a column structured database and tools company, combine it with their GridEngine revamped for financial services, and create an interesting offering that could be quite competitive with DataSynapse and others.
They could offer zfs on their version of Linux. Go after Redhat’s base. Oracle “controls” btrfs (a credible zfs competitor) as well as zfs.
My point from all this musing is that this acquisition was designed to free the value locked up in Sun. Value that has nothing to do with hardware. Value that has a great deal to do with a mishmash of software. Needing a firm/guiding hand and purpose. Which Oracle will provide.
The hardware will, if not spun/sold off to others, will likely be turned off within a quarter or two of the close.
This was a software acquisition, and competitive product absorption. I don’t think MySQL will be ended … rather Oracle will wield it the same way Microsoft wielded SQL-Server at them. And Oracle will make it a gateway product.
But, not mentioned anywhere, is OpenOffice. Oracle now has a direct competitor to Microsoft Office. And Oracle is many things. Dumb is not one of them. OpenOffice 3.x is good. Good enough to use as a daily solution for Office documents. And it runs on every platform. Including windows.
If I were Microsoft, I would be starting to get worried right about now. Oracle just bought a set of competitive software stacks that target its cash cow. Sun didn’t have a clue as to how to wield them. Oracle does.
What I am saying was, while in HPC this acquisition basically means an end to Sun product offerings specific for HPC, this acquisition was huge in terms of the SWOT analysis for Oracle’s direct competitors.
If they play it right, and I have no reason to doubt they will, Oracle will be making life in Redmond decidedly unhappy. Unless Redmond does the smart thing, and goes wide … makes Office available for all platforms. The likelihood of this is about nil though, so I expect Oracle to be offering effectively drop in replacements for Microsoft until Microsoft figures out that this is an existential battle.
But storage, backup, servers? No. Not likely to survive past December this year.
[update] Apparently Steven J. Vaughn-Nichols disagrees, and sees nothing of value here. His thesis is that all that goodness that can tackle the Microsoft stack, that Mr. Ellison has been wanting to do for years, will be discarded.
I think we simply disagree. Oracle is not stupid. Tossing out the hardware, not a bad idea. Tossing out the value … not a good idea.