I really respect Jonathan Schwartz, he has made Sun and Solaris worth considering again. He seems to combined the best qualities of a visionary big company leader who understands technology. A geek who can explain things to big business. Not an straight-forward combination.
For along time Sun was the heart of Unix, but with the advent of GNU+Linux and building of the open-source community in the late 90s Sun has fallen well behind the innovation cycle. Currently people do not (in general) create and use products like Xen, Zimbra, Alfresco, or Asterisk on Solaris. Companies like Google and Paypal use Linux at the edge to drive deployment and increase their innovation productivity, maybe leaving Solaris to the backend traditional well-structured heavy loads.
However comments like this from Smugmug are indicative of the gap Solaris has to leap over to make it a first consideration vs Linux for many sysadmins:
Which brings us to Solaris. Solaris is now open-source, so it’s on my radar again. I love ZFS, I love the fault-tolerant stuff it has for when memory or CPUs go bad, etc. Sure sounds great. It’s a commercial OS, with free updates, and (I assume) good support options. But the last time I gave it a shot (last year), I was lost in userland. Solaris userland and Linux userland differ so greatly, there’s a steep learning curve for someone like me with 14 years of Linux under my belt. But I’m considering taking a closer look.
The ian (Murdock) of Deb-ian is joining Sun as their new Chief Operating Platforms Officer, and he comments about his view on the past acceleration of Linux past Sun.
Everything I know about computing I learned on those Sun workstations, as did so many other early Linux developers; I even had my own for a while, after I joined the University of Arizona computer science department in 1997. But within a year, the Suns were starting to disappear, replaced by Pentiums running Red Hat Linux.
I experienced this exact same situation during this period, using Sun machines at University up until 1998 then moving completely to Linux for home and work as it easier and possible to use. Ian further comments:
As a newborn Sun employee, Murdock is thinking about making Solaris more Linux-like. “When people say Linux what do they mean? Linux is a kernel. Cool apps are not written to the kernel. The OS powers higher levels of the stack. What we want is an open OS platform and to make sure that the existing skill sets and knowledge and training investments are leveraged. We don’t want to make them learn a new product or rip and replace,” Murdock said. “You can make a real argument that Solaris innovated more than Linux in the last few years—such as DTrace and ZFS—but usability stands in the way of appreciating that,” Murdock said. “Part of what we are working on is closing the usability gap so that it doesn’t stand in the way.”
Although quite rightly not everyone agrees with the above comment:
In fact, over the past few years Solaris has spent significant amounts of effort copying the innovations within Linux. One only need to look at their packet filtering and security layer offerings and development.
Solaris still doesn’t even have proper TCP Segmentation Offload support, and that’s a nearly 5 year old feature Linux has had and network cards have been providing this hw acceleration for even longer.
It could be quite easy to play a tit-for-tat religious feature war. I think though it might be better to think of Sun is being a walled garden, and evolution spread up outside when some of the flowers escape. The forest that grew around this garden has started to over-shadowing. Now there is a new gardener and he’s deciding how to tear down the wall and the best way to get the sheltered flowers the garden has been hording back out into the sunlight.
Regardless Ian states exact what Sun needs to do with Solaris. At the moment Solaris is too monolithic, feels like you need a PhD just to install and use it. Even if you only want one feature. Comparatively, Ubuntu is almost to the stage where my grandmother would be able to install and use it. Make it easier to install and use. Produce more publicly available docs on sysadmin tasks.
Lose the need for a PhD to run Solaris.
Here is my wishlist:
- Port rBuilder to Solaris – this vcs like packaging system might work well with Solaris’ monolithic system-style.
- Create and build howtoforge.sun.com – sun docs aren’t really very googlable at the moment.
- Investing in slimming Solaris – making porting appliance projects like openfiler or freenas on Solaris possible.
- Move the PDFs into a wiki
I think Ubuntu is a very clear example of what is possible. A technically brilliant platform that was getting bogged down by a slowing innovation cycle and fragmentation of vision. The progress of integration and documentation that has occured over the last three years generating mindshare is something Sun should be hoping to achieve.