The future of Solaris is in question. Oracle certainly wants to keep is alive, as many of its database customers run on Solaris, but what about everything else Solaris is used for? The important question is one of strategy: you need to quickly decide if your Solaris usage should be reserved only for the essentials, if it should be eliminated, or if it should grow.
I admit it: I was a huge fan of Solaris as recently as two years ago. Solaris is really a great operating system, and when developers at Sun were given free rein they created some amazing components, notably ZFS and DTrace. Great or not, recent events should be worrying. Will Solaris continue to evolve and provide new and useful tools? Or will it stagnate and become a niche OS like the other two corporate Unix versions (HP-UX and AIX)?
Simply put, its place is: big servers running Oracle software. This nicely results in one-stop shopping for Oracle customers. Get your hardware, OS support, and database software from one place, and have only a single place to call when support is needed. It’s a nice idea, if you can afford it.
Here are a few reasons to seriously reconsider your relationship with Oracle-Sun products:
- No more free security patches.
- 150% fines when a support contract lapses.
- No more hardware-only support.
Sun had spent the last few years focusing on lower-end products and building quite a large SMB customer base. Oracle has no interest in that market, and the few immediate changes Oracle made after the acquisition result in Solaris being too expensive for the SMB market. This is probably necessary, recall that Sun was failing, but it means a lot of customers are stuck with migration work.
The finance industry and universities have long since been a big portion of the Solaris user base. Running Solaris on servers that don’t necessarily need to be Solaris, though, is going to end. For example, the Sun x86 storage servers, such as the x4500 series with 48 disks, found a big audience. Combined with ZFS, they make excellent storage appliances. Mandatory support contracts and hostile customer relationships will result in a decline in this area. Not to mention the fact that ZFS will not be a compelling advantage in the near future, when btrfs in Linux matures.
Overpriced and underpowered Sun servers still provided some benefit under the Sun regime. But now that Oracle will require insane support contracts, it may be worth the effort to move as much as possible to Linux.
There is also a major concern about the future of Solaris as an innovative operating system, which I will get to shortly.
Yes, there are still Solaris workstations. Industrial design, university students, electrical engineers, and researchers represent a large portion of Solaris workstation users. In various press releases, Ellison has publicly stated that Solaris will remain a component of Oracle strategy because it does so well on high-end servers (Oracle DB licenses). This seems to be Oracle’s core focus, so the viability of Solaris’s future as a workstation is questionable. If I had to make a prediction, I’d say the user-land utilities and desktop experience will stagnate.
Furthermore, as the community dwindles, so will efforts that make Solaris functional from a year-2000 GNU utilities perspective, such as blastwave. Strides have also been made to have OpenSolaris run on laptops, and the list of supported hardware grew surprisingly large. Laptops and Workstations simply don’t fit into Oracle’s focus, and Linux does it (much) better anyway. The main reason Solaris workstations exist is because ISVs have been slow to port their applications to Linux. For example, circuit design software, such as Cadence and Mentor products, both support Linux now. As a result, the use of Solaris workstations had been decreasing even before the Oracle acquisition.
One might think that because OpenSolaris is a community effort, like Linux, it will survive and even transcend any business decisions. Except, OpenSolaris is not. It is very hostile to the community and the only thing that’s open is the source code, which is open but not really reusable in the pure open source sense. Ben Rockwood (supreme blogger and OpenSolaris guy) explains how OpenSolaris really works, in this blog posting. And in the very next post, talks about the end of the Silicon Valley OpenSolaris user group.
Solaris may remain open, but there is little chance a community of passionate users and developers will congregate around OpenSolaris. In many ways, Oracle doesn’t want that. It’s too much hassle, and Solaris can meet Oracle’s needs well enough by focusing on one thing: running a database. Much like the other two Unix variants that are still widely used, Solaris only needs to do a few things well.
Is there really any reason to use Solaris? The answer used to be “ZFS,” but as btrfs in Linux improves, that reason dwindles. I honestly cannot think of many reasons to run Solaris over Linux these days. While I do work for a Linux company now, I still quite like Solaris. Favorites aside, it makes the most business sense to seriously reconsider your Solaris usage.