Slashdot Mirror
Sun's Linux Killer Examined
gnaremooz is one of several users to mention Thomas Greene's look at Sun's supposed 'Linux Killer'. From the article: "If Sun gets very serious about Solaris 10 on x86 and the Open Solaris project that it hopes will nourish it, Linux vendors had better get very worried. That's because, in the many areas where Linux is miles ahead of Solaris, Sun stands a good chance of catching up quickly if it has the will, whereas in the many areas where Solaris is miles ahead, the Linux community will be hard pressed to narrow the gap." However, he goes on to describe many more difficulties with an install of Solaris than I seem to remember having with just about any recent Linux install.
You can't kill something that's non-commercial
Unix has been around since Linus Torvalds was in short pants.
Yeah, and Solaris x86 has been around since 1992. Hasn't killed Linux yet.
Open Solaris is Free Software, yes? So if it becomes a "Linux killer", then the Linux vendors will simply become Open Solaris vendors. It doesn't matter if Linux dies if what is replacing it is just as free. Hell, the user-space applications are 90% the same anyway.
If Linux isn't successful because something else is better at doing the job and just as free, then that's a cause for celebration, not worry. The only people who need worry about this are the zealots and PHBs who have latched onto Linux for its buzzword value and not its merits.
Bogtha Bogtha Bogtha
I already posted this on TechNudge.com:
I'm not a big reader of The Register, and having just finished the article, I remember why. The article's premise: Solaris didn't crash *as much* as Linux, so Linux had better look out.
Oh, but he couldn't even detect a NIC without the manual editing of conf files, and wasn't really unique or remarkable in any discernable way.
How tone-deaf is the writer to the PC world, anyway? It doesn't take a Bill O'Brien to see that the OS market is supersaturated, and anything short of the second coming of MacOS X will be greeted with a great big yawn from the collective computing community. (Well, a very small band of users will love it and sing its praises. I mean people are still clinging to Amiga OS, for crying out loud.)
This is aside from Sun's remarkable in its ability to ruin every good technology it creates through corporate nonsense and heavy-handed tactics (read: Java), and really, Solaris wasn't really all that thrilling on Sparc. (I spent my entire undergrad shackled to it.)
Neither the article, nor Sun, answer the most critical question in the OS world today: Why should x86 users switch? Why should I leave my comfortable XP or Debian or Red Hat or SuSE for Solaris?
Wait, let me guess: because Sun is including (insert Java widget here).
Note to Scott McNealy: the magic Java dust has lost its power.
Pomme de Terre!
I guess this just seems like a non-issue. Linux Killer? No way. Linux's Friendly Competitor? Welcome to the club!
Dewey, what part of this looks like authorities should be involved?
I was taking the article seriously until I got to this line:
KDE is certainly more popular than Gnome among Linux users, and most would agree that it's by far the better of the two desktops.
We've experienced a number of application crashes since we began playing with Solaris 10, but none capable of pulling the kernel down with it. On the other hand, we've had sloppy JavaScripts immobilize totally, and at times actually re-boot, our Linux box (especially with 2.4.x series kernels); we've seen X oddities do the same, and have experienced several wacky incidents using Microsoft bugware with Wine that required a hard reset. While we haven't taken a systematic approach to blowing up our Solaris 10 installations, one gets the impression of a pretty bulletproof kernel and shell.
That's basically it. The article goes on to basically say driver support sucks and it was kind of a pain to configure, make sure to use the Xorg server and app support is ok. But that kernel, rock solid! Without really mentioning what is happening in 2.6 kernel development or how that argument extends outward toward a better development platform overall.
It's a lost cause, there can only be one. Read all four pages of the article, and ask yourself... would I be interested in creating a disk partition or two and running Open Solaris just to see? I did... and the answer was no... I'd rather spend my time working on my Debian system.
If Linux isn't successful because something else is better at doing the job and just as free, then that's a cause for celebration, not worry.
You're making the mistake by assuming that everything to do with linux is free, open source, and can be ported by a simple recompile.
Do you expect hardware vendors to ever write drivers if the community switches a few times over a few years? What if a commercial vendor says sorry, we don't support that OS, either stick with Linux or lose our product (contrary to some of the opinions here you don't just switch products at the drop of a hat in the real world, a product doesn't just have to be better, it has to be better enough to warrant the pain of migration)
There's a fine balance of amount of choice that's good, and an amount that's counter-productive.
Linux is to Solaris/sparc what the Mac platform is to the Intel platform. (At least before the whole Intel/Apple deal)
You have Solaris/sparc which is rock-solid on its Sparc platform, with integration using the OpenBoot PROM to 100% compatibility with its Sun arrays, Sun NICs, Sun hard drives, Sun video cards (rebadged, but still labeled as Sun)
Then you have Linux doesn't have a specific hardware platform so it is made to be as compatible as possible, and while a lot of hardware is known to work great with Linux, the QA team at Sun who is able to directly interact with Brocade, QLogic, and other vendors to address one-off issues provides a value-add that CIOs like which Linux does not offer, yet.
> ...in the many areas where Solaris is miles
> ahead, the Linux community will be hard
> pressed to narrow the gap...
After all, it's not as if Linux had the backing of a major computer company with a three letter name.
Oh. Wait...
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
My feelings exactly.
Think of it this way:
Linux is free. That means if you dont like the direction, just fork it and improve it. You can still call it Linux. The better fork will win in the community.
Now Solaris is free (kinda, I have reservations about the license). That means people have Solaris code available to them. If Linux is generally good, except for some solaris features, they'll just port those features to Linux. If Solaris is awesome except for some Linux features, the same will happen. In the end we'll have code that is good, does cool things and is free. Whether you call it Solaris because you think it was 'descended' from Solaris or Linux, is a political matter. Linux wasnt threaded or ran ELF in the beginning. It wasnt SMP. Now its all those. Can we say it is a Solaris with the Linux name?
"Give orange me give eat orange me eat orange give me eat orange give me you." -Nim Chimpsky
Meanwhile, Linux has been around since 1991, and it still isn't ready for the desktop. If upstarts like Firefox and OS X can increase their market share so much faster than Linux, why not a revamped/free-as-in-beer Solaris x86?
Probably because almost all of the desktop software available in Solaris x86 is exactly what is used on Linux: Xorg for X11, GNOME (or possibly KDE if you so desire) for a desktop environment. StarOffice (which is to say OpenOffice.org) for office applications, Firefox as a web browser, Evolution as an email client... the list goes on. What does Solaris 10 offer that Linux doesn't? DTrace and excellent developer and server performance tuning tool. Zones, and excellent server security and partitioning system. Really crappy hardware detection and configuration. A severe lack of drivers for standard consumer hardware. A packaging system that's great for updating servers but even worse than what Linux offers for desktop use.
Solaris 10 will be ready for the desktop a sometime after Linux is ready for the desktop and not before. The desktop software stack is the same, and Solaris offers nothing new for desktops at the lower level. It does have nice features for servers, but then so does Linux. I would expect Solaris to gain back some ground in the server space slowly, but I don't forsee how it could manage to somehow shoot up in market share any faster than Linux already is.
Jedidiah.
Craft Beer Programming T-shirts
We're about to see a major war between three very large computing firms: Microsoft with Windows and .NET, Sun with Solaris and Java, and Apple with Mac OS X and Cocoa.
.NET and the marketshare of Windows.
Frankly, I think this desktop/workstation market conflict will make the UNIX Wars of the late 1980s and early 1990s look petty in comparison. In one corner there's Apple, offering extreme multimedia and usability via Mac OS X and Cocoa. Then there's Sun, with the extreme stability of Solaris and Java. And finally Microsoft, with
It isn't just a battle over which operating system is better. It also involves three competing development environments involving three separate (yet similar in many ways) languages. I'd like to consider it more of a Systems Stack war. The vendors are competing on their ability to provide a coherent operating system/programming platform composition.
I believe we will really see things heating up in the near future as each system attempts to draw the best features from the other. Windows will obtain the stability and security of Solaris; Mac OS X will obtain the enterprise connectivity of Solaris; Solaris will obtain the multimedia mastery of Mac OS X. We're living in very interesting times, folks!
Cyric Zndovzny at your service.
IBM discovered the hard way in the nineties that a hardware manufacturer trying to get competing hardware manufacturers to support their OS is a dead end. Discussions between IBM and the other PC vendors sounded a lot like similar conversations will if Sun tries to get PC manufacturers onboard the Solaris wagon:
Sun: Hi, HP, what do you think about preloading Solaris on your workstations?
HP: Yeah, right! Why would we want to license or support our competitor's operating system for our hardware?
Sure, Sun might be able to get a few PC peripheral vendors on board. But, honestly, what kind of target market can Sun tempt them with? Solaris x86 has a smaller presence than Linux and you've already said that these same vendors aren't getting on the Linux bandwagon.
I agree, Solaris is clearly the desktop operating system of tomorrow!
And it always will be.
All movements for social change begin as missions, evolve into businesses, and end up as rackets.
Meanwhile, Linux has been around since 1991, and it still isn't ready for the desktop.
It's close enough, and I'm formerly a member of the "not ready for the desktop" camp.
I installed Ubuntu on a laptop last weekend. It configured everything automatically except the sound, which I had to tweak some config files for (no worse than when I've had sound problems in Windows).
The only reason I had to do cliched Linux stuff like recompiling the kernel was to get my Orinoco card working in monitor mode. Desktop users don't care about that, only people who want to run WiFi hacking utilities.
Keeping the system up to date is actually easier than Windows, since I can run a single apt-get and upgrade everything (OS components + apps) to the current version.
There are definitely some gaps in terms of things like no Photoshop on Linux, but the OS itself is fine for desktop use now IMO.
"...always new atoms but always doing the same dance, remembering what the dance was yesterday." -Richard Feynman
"Killer" just means "70% of the features and sort of works"...
What I'm listening to now on Pandora...
- Solaris 10 and Open Solaris (which you build and install on Solaris Express) are both very nice, Linux-like operating systems.
Linux is no longer "Unix-like", people. It's Unix that is "Linux-like".Its closing the gap argument missed some really important issues; for example, developers. There are some things that Linux doesn't do, and will never do because the benevolent dictator doesn't believe in them.
For one, POSIX compliance. OpenSolaris IS compliant, so as a real-time junkie who loves his shared-memory mapped files, I'm bouncing up and down. Linux shared memory stubs some calls, doesn't implement the POSIX suite, while barely implementing older shm. How many MAN pages can you find that tell you "This isn't implemented." in OpenSolaris?
/\/\icro/\/\uncher