Nichibotsu (日没) is the Japanese for sunset (as far as my Japanese goes)

I developed a bit of an interest in what will happen to Sun, especially Sun’s hardware, in the context of an enterprise environment. Does it still make sense to invest (long term) in Sun machines or we should really be looking for other solutions? This post is about Oracle‘s acquisition of Sun.

Where did Sun go wrong

Sun used to be quite popular in datacenters in the old times and in the more recent times it was still popular with the database folks.  While the company was doing ok-ish on hardware sales and had plenty of cash in the bank they kind of fell out of fashion with the street and share prices kept falling.  Everybody started noticing the lack of focus of the company. In a short period of time Sun changed strategy a number of times, focusing either on networked systems, Java, database software or so on. Nobody really knew what Sun was about anymore.

While Sun was brilliant that it could offer pretty much a whole computing infrastructure all the way to the application level, it’s lack of focus made it’s offer uncompetitive.  IBM started driving Sun out of it’s traditional hardware market and Microsoft was challenging Sun quite a bit in the OS and databases field (not that Sun ever made a coherent point about what it wants to do in the DB space anyway).

Last but not least Linux put a big nail in Sun’s coffin by providing a viable OS alternative running on the same hardware and also being able to run on other hardware also, thus freeing any vendor lock-in that people might have felt with Sun’s computing platform.

Bottom line is that Sun’s lack of leadership made it uncompetitive in the market and everybody else seized the opportunity.

Will a new day begin

It is a bit unclear what will happen to Sun now. Oracle’s CEO boasts that they will now be able to offer a SYSTEM (i.e. a hardware-software integrated platform) to run the world’s financial systems and airline booking systems. This is doubtful by itself in my opinion, for a number of reasons:

• Oracle would be pretty darn late to the game on this one. Vertical integration started a while back (everybody has been doing it for a while including the late comers like Cisco, EMC, vmware) and most of them are realizing that hardware is not the key. Hardware will be a commodity and people will be buy a piece that can be replaced with another piece without worrying who and how made it (like light bulbs).
• Integrating hardware and software is a bad move. Software is agile, quick to evolve, patch and upgrade. If you tie it down to hardware you will constantly be behind the curve.  Everybody is building clouds now. Clouds abstractize the hardware and allow you to focus on the important stuff: the software. This further stresses the point that hardware will be a commodity.
• Oracle has never sold hardware. They always sold software. Selling hardware is a completely different ball game and requires a different development and QA cycle, different sales channels and strategy and an all together different corporate culture. It’s unlikely that Oracle will manage to integrate these under a single umbrella and if they do, both the software and the hardware division will suffer.
• What is the market for Oracle’s “Systems”? The really big financial firms will not upgrade to an unproven (and currently non-existent) technology any time soon so Oracle will have to dump money and time in this integration. This will make it even later to the game. On the other hand smaller companies are likely to go for Linux based solutions or cloud providers, they will not spend money for big iron and big licensing fees.

Given the above it’s not really clear why Oracle really made the move but my guess is that it has something to do with software. Sun must have something that fits in Oracle’s overall SOFTWARE strategy.

Virtual sun

One area where Oracle could benefit from is Sun’s expertise in the operating systems and virtualization area. Oracle might just become the next big thing in virtualization, building on Sun’s experience and integrating solutions such as ZFS, Sun xVM and VirtualBox. The market is still fresh, dominated by vmware which is loosing market share and seems to be incapable of innovation. This might just make a quick win for Oracle. They have already made some moves in the arena with Oracle Linux and Oracle VM (and Oracle Enterprise Manager) but the products are lacking any real competitive advantage as of now. This is one of the more obvious areas where Oracle could build on Sun’s software expertise.

Bottom line

For the hardware division the future does not look so well. While some demand is still there, Oracle will have to invest heavily in optimizing the channels and improving support (something that Sun has been particularly bad at in the past couple of years) in order to regain customer confidence and hopes for any profit. However they have little reason to further invest in that division so my guess is that soon we’ll see the hardware division for sale pretty much the same way IBM sold the harddrive division to Hitachi and the Laptop division to Lenovo.

Unless you have a very particular business stick to the trend: hardware will be commoditized, software will be virtualized. While Sun’s hardware division will continue to exist under one name or another it does not make sense to make any long term commitments to it, you will be better of sticking to any x86 platform and enjoy the sunset for now.