It is Very interesting to see al these theories and musings on the subject of Oracle (aka Larry) and Microsoft (aka Steve) forming what seems to be a “friendship”. I think very few (if any) of them commentators is pragmatic (read: they are all wrong). (The other few new recent “friendships” of Larry are somewhat irrelevant). So here goes my version (which is right, of course).
That Pesky Wall Street
One day (most likely last year) some poor salaried soul has respectfully explained to Mr Ellison that Oracle results to be published (in 2013) will not be easy to be beautified any more, and that ORCL ticker, will start looking like one trouble too much, for many investors. It will be difficult to sneak in somehow under the Wall Street radars, most likely was the realisation.
Being Sun Tzu warrior as he thinks he is, Larry very simply realised he can not kill the messenger (at least not that lucky one). And Larry already new, he needs to be “on the cloud” and he needs to be there ASAP. And Larry knew he will loose the war itself if he is not on there rather quickly. And Larry knew he can not build his own cloud. It would look rather small and funny, on the sky populated with large global networks of the clouds, owned and operated by the “Big 3”.
But why exactly he made this conclusion? Why Larry did it? And this is where almost everybody is wrong. Not because of some far reaching and sinisterly slow moving dark IT geo-strategic movements, or some “who owns what” secret world ruling cabala. No. It is simply because ORACLE was (and still is) too expensive for what it delivers. Allow me to emphasize the point once more:
ORACLE is too expensive.
Got it now? For a given amount of data to be stored in its flagship RDBMS product, and to be managed by it, ORACLE database (and all the other supporting products) is too expensive. It is as simple as that.
More educated readers might think that the new database version, ORACLE 12c, is a quantum leap and very important release. It divides meta-data and storage. It is very clever and MSFT SQL Server (right now) can not match that. True. But sadly not important any more. And also this is not why Larry did it.
Please understand that these days, past and present ORACLE customers are offered (by competing hordes of VAR’s), a rich smorgasbord of alternative solutions to ORACLE. And some of them are really very tempting and not just because of the costs. They simply (and feasibly) are solving customers pain points, which no ORACLE release can, or ever will.
Yes. That is what I am talking about. Let’s not forget the whole NO-SQL movement: The one iffy big white elephant in the corner of every meeting room, with ORACLE representatives. For example. Particularly one, Large (if not the largest) user whose empire depends on super fast and super clever data handling is NOT using any kind of RDBMS for that purpose. Yes, it is Google of course.
Google Ante Portas
There is no larger data ocean on the planet than the one which Google owns. And all of that is managed in the most successful way and without anything like RDBMS present and in use. This is achieved by using the clever (and simple) MapReduce. Running on the wast network of cheap logical nodes running in parallel. Now, allow me please re-emphasize the second point, for which Larry decided to ask the “devil” himself for the help:
SQL is not the “Only-One”
Repeat: There are very valid and very successful alternatives to SQL. Not in theory, not in labs or hackers forums, but as mature products, in use right now, inside big and successful corporations. HADOOP, BigTable, etc. There is no (and never will be) RDBMS which can do (or will be used) for large unstructured data sets querying. Few have tried it with ORACLE, just to be made prominent later, as good examples why is it better and far cheaper to do it with one of the new data handling technologies. Also, not running on mega expensive (Sparc) servers, but running on the cheap and vast arrays of PC’s.
And yes, I know, there are sadly few orphans Larry has managed to collect: SUN, Java, WebLogic, Sparc hardware and a such. Small fragile and soft(ware) part of this sad bunch, will also find a temporary home on the Azure cloud. And, I am saying “temporary”. It won’t be long before they will be sold or simply “put to rest”. What will happen to the whole SUN/Sparc hardware saga nobody knows. And sadly anybody cares any more.
So here we go. Larry simply realised “it is going to hit the fans” if he does not make the (sudden) move to the cloud. Larry realised the only way to keep ORACLE relevant is to make it cheaper and to position it in the cloud. No matter what the cost is going to be.
And there is no way ORACLE can build its own cloudy data centres to match those of the big three. It is simply not possible for ORACLE, heck even Apple struggles to do it. So. Larry simply went to them enemies and asked: “How much is the ticket to the Cloud?”
if you can not beat them, join them.