Michael Dolan Dot Com

Tuesday, July 24th, 2007

Enterprise Linux Log: “OpenSolaris users group to Ian Murdock, Sun: You’re missing the point”

URL: http://enterpriselinuxlog.blogs.techtarget.com/2007/07/18/opensolaris-users-group-to-ian-murdock-sun-youre-missing-the-point/

If you watch my space here regularly, you probably would not be surprised that I find community reactions to something like this both interesting study and at the same time amusing. Jack Loftus wrote the article but it seems another colleague of his was in attendance. I was in attendance at this meeting as well (yes, shocking I know - my first one… and probably last). What I witnessed was in-line with what the article describes. In all fairness to the “Solaris camp”, it’s hard to see what value a Project Indiana is going to bring to the OpenSolaris table other than to make it more like Linux and perhaps closer to Linux in areas Linux users take for granted (package management, install, usability). But what does that offer the Linux crowd that Sun is obviously trying to steal (hint, Linux users already have these benefits…)? At the same time, does it really help the Solaris user? So while it takes Sun two years likely to get Indiana viable, Linux will only improve on what it already has in that time, and so what benefit will be there in 2009/10?? Further, this is just Sun - OpenSolaris is Sun-only - it’s one vendor… when did open source community users start contributing to vendor-dominated projects (hint, they haven’t).

I can say good luck Ian but I certainly do not envy his challenge ahead - especially when it was evident the NYC Solaris faithful were not exactly asking or waiting for Indiana to help them. I was not impressed by Indiana. I’d instead call it the Solaris Makeover Challenge. Oh wait, maybe IBM has already done that. :-)

I was slightly to actually offended at some of Ian’s jabs at Linux that are now the Sun “party line” and I suspect Ian knows full well these FUD lines are not entirely true (my how things have changed from being the LF CTO…).

Ian made it clear [now that he's at Sun marketing Solaris] that the only reason people use Linux is “because they couldn’t afford Sun workstations and Linux was free” and “that’s how Linux happened”. HA! I started using Linux because Windows 95/98 were horrible. Yes, free also played a part but just because things are free does not mean millions of people use them (look at all the BSDs, MINIX, Evolution, etc, etc). Then the other shameless argument was “Linux breaks compatibility” which is a small, small % of the use cases out there - sorry, doesn’t fly - not to mention that many of those “Linux breaks” are actually modernization technologies which the lack thereof are the same reasons Sun users are flocking from Solaris (hint, when did you last see a modern Solaris shell? Solaris in HPC?). Shameless because I also thought that was Ian’s job at the FSG…

Another interesting point from Ian was that he flat out stated that multiple distributions is the greatest weakness of Linux… which I tend to think of as a benefit. While we’d have to agree to disagree on this point, I’d like to think that if one Linux vendor is asleep at the wheel (takes 8yrs for the next release) or perhaps does something that I don’t like (i.e. partner with X company) then I have choice - I can switch with relatively very low pain from one Linux distro to another (perhaps just over time). I’ve seen customers do this successfully - they didn’t receive the support they expected and tests with other vendors proved other Linux providers better. It’s not that there was a problem with Linux or even the vendor, it’s that vendors interface customers through people and sometimes people make mistakes. Switching lets you avoid poor performing people.

Just think… what happens when you only have 1 source for you platform (think Windows, Solaris). Have you seen Windows or Solaris “fall asleep at the wheel” and maybe “milk customers without reinvesting”? I think that’s what you’ve probably seen since 1999. Remember when Sun launched Solaris 8/x86 only to take x86 support away? Then bring it back again in Solaris 10???

Sure, Sun flashed DTrace and ZFS onto the Solaris user base and wowed them with something modern… but if Solaris is so powerful and modern, why does the license prohibit publishing performance results? Linux is the opposite - Linux creates competition, drives competitive and cooperative investments and in the end customers get choice from competing vendors. And with Linux - it may not be the distributors competing that benefits customers - it could be Motorola competing with Apple needs feature X and to get feature X into Linux, Motorola invests in X which is later picked up by Red Hat and Novell. It could be the maker of the fastest network card on the planet trying to tweak Linux’ TCP/IP stack that makes Linux have superior networking performance… Something to think about…

Anyway, a snippet from the article that I recommend reading:

The user’s group meeting last night all but confirmed this was precisely what was happening - at least with OpenSolaris users in NYC anyway. Ken Milberg, Linux site expert and contributor to this blog, said some audience members started to notice that Murdock’s plans for OpenSolaris were starting to resemble Red Hat’s strategy with Fedora. It didn’t gel. “Sun pretty much admitted that this strategy made sense, and more or less were admitting that is was time they starting talking about an ‘innovation strategy’ too,” Milberg said. Thing is, the Sun users kept saying they were already satisfied with OpenSolaris.

And yet Sun and Murdock continued to lay out the case for Project Indiana, to “mixed results,” Milberg said. One unidentified attendee reportedly blurted out “all this does is help Sun, what does this do for someone using Linux?”