According to at least one company, in order for Sun to gain credibility in the Linux market, they must acquire either Red Hat or Novell. Merrill Lynch hypothesizes that with one of these companies as an asset, the Opteron market is theirs for the taking, and without one, they are doomed to cater to mostly Solaris customers.
Analyst: “Sun Must Acquire Red Hat or Novell”
About The Author
Adam Scheinberg
Technology Executive • Web Developer • Father • Foodie • Music Snob • OS enthusiast
Follow me on Mastodon @[email protected]
Isn’t Novell just as big as Sun?
Redhat could be a possibility.
Hehe, that’s exactly what I was thinking! But if they could acquire Red Hat, it would certainly be a smart move. We’re in for interesting times…
two companies
#1 eat up a company
#2 cash (?)
#3 die
no company
I’d rather have two companies than one big one.
Especially as the player in the linux field are getting rare.
-mo
There was an article in IEEE Spectrum a while back about a hypothetical Sun+RedHat + Apple merger. Interesting read.
Novell market cap: 2.60B
Red Hat market cap: 2.94B
Where do you get numbers from?
What about stock size?
Wall Street is clamoring for Sun to buy Red Hat or Novell. SO WHAT! Frankly, this does not make too much sense. First, IBM has too much invested in Novell and SuSE. There would be too many obstacles there. As for Red Hat, IBM is pushing this too (as the primary OS for all Eclipse based applications, including Workplace/Lotus Notes), but there are other/different problems:
1. Red Hat’s virtue to many is its independence from hardware.
2. If Sun did buy Red Hat wouldn’t this be more to force Dell and HP to look at x86 Solaris (expensive way to do that!)
3. It would be very high price to pay for something they really already have.
Why wouldn’t Sun want to cultivate a more “open” model while backing Debian and Gentoo and forget the client. On the other hand, if they took an aggressive position on something like this based on a PowerPC solution with Solaris.
http://www.sun.com/sunray/sunray170/details.xml
It could *REALLY* shake up the market.
Apple + Sun could be interesting, but in that case why Red Hat and wasn’t the idea to get Solaris on the client? Another use of Apple hardware once branded exclusively with a bite out of the fruit of the tree of knowledge?!…sounds like a cardinal sin to me. 🙂 Poor Job(s)! In this case, why not YDL?
Don’t Sun and Apple both use open firmware?
Of course the performance and the application opportunity would be much greater and that would shake things up! ==> Surprise, surprise….;-) As suggested in the article, Sun’s reputation has always been against the herd and many of their customers are with them because they are different. This was the perception that has made them “hot.”
In this scenario, Sun Ray could BE the real BIG winner! Smart cards can go EVERYWHERE and this could tie to a whole new form factor for mobile devices that would be YOUR laptop/gamepod/video phone of the future. Not too mention what could be done with AltiVec enhanced Java as IBM marches ahead to completely control the ISA of the next generation CELL PowerPC CPUs. Keep in mind all the next-gen video game players are PowerPC based…PS3, XBox and whatever Nintendo has coming. What was done with Java, IBM wants to do in hardware and some new fangled way of packaging everything into a CELL (data and applications) so that all devices on a network have to have the same ISA to actaully participate.
There could be other PowerPC alternatives…Scott will sort it out. He has two choices “heaven or Dell.”
Sun’s whole stance is that they don’t think Linux should be on a server. Why would they go out of their way to “be taken seriously” as a linux server vendor? Merrill Lynch’s whole stance seems to be based around the assumption that Sun wants to sell linux servers, when the whole point of Solaris 10 is to be cheaper and faster than linux?
“If Linux is Unix, then Sun should dominate Linux”
Linux is not Unix. Unix is something based around standards compliance. Linux is a poorly implemented Unix clone that cares nothing about standards, doesn’t even have full posix compliance yet, and makes up APIs whenever the developers see fit.
Maybe if Merrill Lynch employed more people than MBAs, they would know the reasons everything about their prediction was crap.
Next is the fact that Red Hat is probably the worst possible distro you could ever put on a server. It’s buggy, wasteful with memory and CPU cycles, and generally contrary to everything Sun is trying to aim for in servers (which is why they make it a point to let each new release of Solaris run on things as old as the 10mhz sparc stations. And run well at that.) If Solaris can run just as well on a 15mhz workstation as it can a quad 2.4 Ghz Opteron server, or a 32CPU SPARC server, why on earth would they choose something that “recommends” a 300Mhz to run “marginally” well, with a minimum memory requirement that is twice the recommended memory for Solaris (going by Solaris 9 specs. I doubt Solaris 10 will keep it’s 512MB requirement, since it runs fine as a workstation on a fourth that much.)?
“Those customers and developers that doubt Sun’s Linux love would simply have to get over their concerns – fast.”
What linux love? Sun has stated, in so many words, many times that they believe their product is superior. What love are they looking for here exactly? To quote sun (not really word for word) “We provide Linux for those customers who desire it, but we believe Solaris is a better product for server use.”
On top of that, purchasing any Linux vendor would then put the financial strain on Sun to support developement of both Solaris and Linux, which, being GPL, would not really give them the same advantage that improving Solaris does.
Ultimately, however, Merrill Lynch’s latest musings will no doubt end up in the trash can made specifically for them by Sun years ago. At Sun, this is known formally as the “Loony Tunes Receptacle.”
First sound bit of thinking in the article…
I really don’t understand how a big company could be so oblivious of reality?
Apple + Sun could be interesting
You know, it’s a nice idea, except for the fact that PowerPC is made by IBM now, so buying Apple would be more likely to kill the Mac than create a PowerPC Solaris box. (Really, if you want a PPC unix server, just buy something running AIX from IBM? Like this for example: http://www-1.ibm.com/servers/eserver/openpower/hardware/720.html but with AIX and not Linux)
Sure, they could use the old G4 made by Motorola, but those are worse than the CPUs Sun makes themselves?
Indeed, another PPC version of Solaris (like the one with… SunOS 2.5?) would be rather cool… but highly unlikely.
Sun would never be able to by Novell – in terms of its market reach and its relationships with companies like IBM it isn’t going to happen.
Sun could buy Red Hat, but it would be an absolute disaster. Red Hat would resist it, and the takeover would be a disaster as it just wouldn’t work. Sun’s history of takeovers speaks for itself.
Anyway, this is yet another example of someone thinking that Sun is a big company. It isn’t. It may have a reasonable amount of cash, but that cash is sitting there doing nothing with no idea in site as to how they’re going to maintain it. Sun is not an aquisitions company, no matter how many people would wish it.
http://www-1.ibm.com/servers/eserver/pseries/hardware/entry/520expr…
that’s the one I was looking for.
There is no need to buy. And I certainly don’t want to see Sun take over a company and then destroy it through incompetence. Sun has not convinced me that it knows how to take a good idea and make it profitable. Sometimes they do, sometimes not.
What I would like to see is a formal business partner relationship form between Sun and Novell or Red Hat. This still has the benefits of showing a dedication to Linux and of good will, with out the huge expensive and without dragging other companies down.
Back when Sun looked like they might use SUSE for their distro I was pretty happy. Then they started trash talking Novell…
Probably what would make the most sense would be to team up with Red Hat and leverage (push) Java. Personally I’d rather see them with Novell (I wrote a post about this a week or so ago). But in reality, Sun has too much invested in Java to not somehow make that part of the deal.
For example, we could see:
Sun + Red Hat + Java vs. Novell + Mono
Linux is not Unix. Unix is something based around standards compliance. Linux is a poorly implemented Unix clone that cares nothing about standards, doesn’t even have full posix compliance yet, and makes up APIs whenever the developers see fit.
Mmmm, and yet all the software out there is written for Linux first. Full Posix compliance has been a stated goal of Linux, and developers do not make up APIs as they go along.
Did you not see the late 80s and early 90s? Unix was never based around standards compliance. Go tell that to the vendors that did make up APIs and extensions outside of any standards.
It’s buggy, wasteful with memory and CPU cycles, and generally contrary to everything Sun is trying to aim for in servers
Sounds like Slow-laris to me.
If Solaris can run just as well on a 15mhz workstation…
God, I thought you were being serious there for a second.
(going by Solaris 9 specs. I doubt Solaris 10 will keep it’s 512MB requirement, since it runs fine as a workstation on a fourth that much.)?
Sun’s workstations run Gnome as their basis. There is not a snowball in hell’s chance you’ll run it on a fourth that much.
“Sun’s workstations run Gnome as their basis. There is not a snowball in hell’s chance you’ll run it on a fourth that much.”
Well, you could, but I definitely wouldn’t recommend it.
I’m running Gnome 2.8 on 128MB RAM right now, which is 512/4, however I’m not running Solaris 10.
First Sun bashes Red Hat on a daily basis for months, and then they buy them because they have a better product? I don’t think so.
Sun doesn’t have the finaces to support this kind of deal. they have nearly 2 billion in debt, the 1.6 billion from MSFT, won’t last long and they are hemoraging cash almost as fast as SCO.
Sun if they could trim down, straighten out their managment, kill sparc for opeteron, and increase the x86 hardware that solaris supports 10 fold will finally be in a position as an option to Linux.
…if being standards compliant involves forking 8 ways from sunday without a hope in hell of ever bringing the code back in sync. I can’t say Linux has done that overly bad in this area.
IBM and Motorola/Freescale make the PowerPC. Most apple computers sold today use the G4 – many more than the G5! OK, the G5 (IBM) and MacOSX get great reviews in the WSJ (recently) and Apple’s stock has quadrupled in the last year with Merrill Lych types forcasting $100 stock targets. Sun might be a million iPods slow on that one…
On the other hand, was Motorola/Freescale in the Power.org announcement?
http://www.power.org/
Whats up with that?
Why not Sun and Motorola/Freescale? They are the world leader in embedded CPUs. Getting a sparc portable is not as likely as a Freescale one. They do it with Apple already.
The only reason Sun is playing around with Linux is to gain interest in its Java Desktop System so they can put it on Solaris x86. Once people gain more interest in Solaris x86 for a viable desktop OS, they will abondon Linux altogether.
That is a very good idea. Sun could promote a total solution and not have to worry about HP or Dell. Motorola gets leverage on IBM. Sun gets into the embedded space connecting all those Motorola controllers everywhere.
SMART
Think about it like this, would Red Hat be any better off if it were owned by Sun? Highly unlikely. If Sun wants a powerful Linux distro, they could build that today without needing to buy Novel or RH. Sun has different views of what they want Linux to be than other vendors. If they purchased Novel today, SuSE would just become JDS.
If Sun wants to have a mainstream Linux distro they could just ‘repackage’ existing code base and start offering ISO’s and boxed sets of it, then add their own code contributions as they start to buiild a programming team. Mandrake does this with only like 10 programmers. Why would Sun buy a multi-billion dollar a year company just to do that??
Sun needs novell or red hat more than Novell or Red Hat needs Sun.
Besides, for all their talk, Sun doesn’t truly believe in Free software. If they were in M$’s position, they’d probably be more despotic.
If Solaris can run just as well on a 15mhz workstation as it can a quad 2.4 Ghz Opteron server,
What????
if they acquire either company it is only going to make sun look like a monopolistic monster. sun can beat either one of them with quality, sun should not have to resort to eating one of them
Smart cards are instant DRM!
The cultures behind RedHat and Sun are different, IMO. RedHat always felt more like a fly-by-the-seat-of-your-pants OS, with things like RPM getting in the way and their installer is often so brain-dead I can’t stand it. I also hate having to learn so many needlessly-complex RedHat-specific things just to do the things I can do by rote on other systems.
Sun should figure out a way to aquire Debian (I know it isn’t really possible), so at least they would have a sane distro on their hands.
I think Sun will probably not aquire a Linux distro and will keep reselling RedHat, because they really don’t have a choice in the Linux market. It wouldn’t be a big deal in the future for Sun to provide a driver supplement for RedHat to make sure everything works all the time (currently they mainly use reference boards that RedHat already supports).
Novell first tried to aquire Red Hat but RH said stuff it. I really don’t think they would sell out to sun, especially since they have done nothing but BASH them these last couple of months (more so than ever).
I like Sun and solaris but I don’t ever see using it as my main os. I like my Linux/Slackware community just fine thank you.
I don’t get it, why does Sun have to buy RedHat or Novell to get into the Linux game? We hear over and over again that RedHat and SuSE do not equal Linux, yet the same people spouting those words, insist that Sun buys RedHat or Novell.
Now, how does this sound for a business plan? Sun opens a small office in India and China and hires about 40 full time developers and they take Debian Linux as their base. They use 10 developers to help out the Debian core distribution (installer, package manager, etc.) and use the remaining 30 to build and mantain their own Linux distribution. Then they could use their existing infrastructure to sell and support this distribution. Instant awesome Linux distribution at only $1 million / year. Beats the heck out of spending 2.9 billion to merge.
Which unix?
Unix is just a trademark given to anyone who follows the basic posix standard and applies for the expensive certification
the now extinct caldera Linux did apply for it. its useless when it comes to compatibility since everyone build upon those posix standards in incompatible ways
“RedHat always felt more like a fly-by-the-seat-of-your-pants OS, with things like RPM getting in the way and their installer is often so brain-dead I can’t stand it. I also hate having to learn so many needlessly-complex RedHat-specific things just to do the things I can do by rote on other systems.
”
how much crap can people throw out. redhat is a enterprise quality OS that has marginalised solaris in the industry in a major way forcing Sun to market the next version of its OS heavily while trying to open source.
this is reality. period
“redhat is a enterprise quality OS that has marginalised solaris in the industry in a major way forcing Sun to market the next version of its OS heavily while trying to open source.”
Ha ha. Does RedHat even come close to Sun’s product lifecycle and support durations? Also, Solaris is taking shape into what RedHat could never be. The new feature list for Solaris 10 is amazing, complete with BSD-style packet filter, new TCP/IP stack, ZFS, etc. Solaris has good documentation that is generally task-oriented and thorough. Solaris doesn’t cater to the lowest common denominator, either, while Red Hat’s installer just drives me nuts.
And Sun is right that Red Hat is not free. Buying “enterprise” Red Hat is thousands of dollars. Sun’s new pricing model is dollar-for-dollar competitive with Red Hat all the way down to free Solaris against Fedora. Sun figured out Red Hat and get revenue from hardware, too!
Great, so Sun will buy RedHat, and then not to feel “left out” IBM will take over Novell. Sun continues to lose out to IBM.
But if either one would become a reality, it would just mean or me one of two things for me:
1. Stop development in Mono (before anyone bashes over this: ECMA + GTK only)
2. Stop deploying RH servers
Both decisions would be hard to make, but then again still better than being riddled with Suns daily fashion of making different moves with different (friendly and unfriendly) players, different OS, different licenses. Bt the most hard thing for me would be following daily fashion of McNeally and Schwartz (I doubt I would survive even for a few days)
Is that not what they want anyways? Is that not the reason for opening up the Solaris source?
Sun’s in the linux market, or linux in the commercial OS market?
Interesting hypothesis on the part of the genii at Merrill Lynch, but I wonder if any one at MR is on the record a few months ago practicing their hypothesizing skills on the decline of the all mighty dollar.
Anyway… back to my point. Sun, AMD, and RH, all great superstars, at different times, including maybe now, but also superstarts only in their own niche. But outside their niche, where other big boys compete, Sun, AMD, and RH are nothing more than underdogs. A collection of underdogs doesn’t make a pitbull.
how much crap can people throw out. redhat is a enterprise quality OS that has marginalised solaris in the industry in a major way forcing Sun to market the next version of its OS heavily while trying to open source.
Red Hat is isn’t marginalizing Solaris, it’s marginalizing Sun hardware, which runs Solaris. Red Hat succeeds because of the commodity nature of its platform, basically cheap x86 servers, and it runs “good enough” for most of the use cases on that hardware. “Why run SPARC/Solaris when Dell/Red Hat meets my needs for less $$$” is a perfectly valid statement.
Red Hat is not beating Sun because of overall quality, it’s simply the “worse is better” scenario.
Sun tried to fight back with cheap SPARC boxes but failed to make much of a dent because of the wide open x86 market vs a comparatively closed SPARC market.
So, now with the AMDs, Sun is embracing the x86 architecture, particularly on the low end where Red Hat has been dominant. Sun didn’t have much commitment to the x86 market with Solaris 8/9. Solaris wasn’t better enough than Red Hat to justify customers spending $$$ on it, and they didn’t have a hardware platform that could distinguish itself for the x86. But they do now with their AMD machines. If they’re going to support the AMDs, there’s no reason not to support generic white box Intel.
Now, with the renewed push towards x86 support, and with their new Solaris 10, Sun is going to be able to compete on a level playing field with the likes of Red Hat et al.
If you want to run Red Hat on x86, their AMD boxes are a compelling, high performance platform. I hear nothing but good things about Suns new hardware. Sun will happily sell their boxes to anyone who asks.
If you want to run Sol 10 for its new features on your current server hardware, go ahead. Sun will happily support it. On the server side, their HCL is pretty good.
Basically, Sun is now pushing a “better mousetrap” for a competetive price against Enterprise Linux distributions. They’re taking both Red Hat and Dell head on for the low-mid range server market. And the Janus Linux layer in Sol 10 will enable users to migrate many current applications to Sol 10, so that switching will be a reasonably cheap operation for those who want to do it.
Sol 10 x86 gives Sun access to the lowest end server markets on a white box home grown machine asssembled by a VARs high school son, while giving continuity to the highest end uber (currently) SPARC data center multi processing monsters. The adminstrators see the same thing for edge servers and the inner core super servers. Same management, same tools, same training. Only have to manage the binaries. If Solaris handled Fat Binaries ala NEXTSTEP, then even that problem would go away.
No reason to switch to Solaris if you’re happily running Linux today. But it will be a compelling option for any company that needs to buy new machines, or upgrade their servers (say, from RH 7.x to the RHE 3) to support new binary application versions.
Unless it doesn’t support a specific application that you require, Sol 10 compares very favorably with RH.
Seriously, I think it’s a whole new ballgame.
Sun doesn’t have the finaces to support this kind of deal. they have nearly 2 billion in debt, the 1.6 billion from MSFT, won’t last long and they are hemoraging cash almost as fast as SCO.
You must be thinking about a different Sun than the rest of us. According to Yahoo today, Sun has $3.52B in cash and
$1.16B in debt. According to their 2004 10-Ks, they have better than $7.5B in cash and marketable securities; up slightly better than $2B from 2003. RedHat’s market cap is $2.88B. Sun could give a 20% margin and still be able to pay cash out of retained earnings. Of course, they wouldn’t, they’d do a stock swap out of treasury stock.
You know, no one seems to understand finance at all. Just because Sun shows a loss doesn’t mean they are hemorraging money — their cash and retained earnings *increased* by over 2B last year. A large part of their “loss” was restructuring charges. Restructuring charges are monies that they retain for expected future expenditures due to restructurings; it is not even money that they have spent. Another charge was for “impairment of goodwill”, a decrease in the company’s value based on a change in their reputation. This is not even real money.
Someone else mentioned RedHat marginalizing Sun. Sun has revenues of $11.28B last year. RedHat had revenues of $124.7M, just barely over 1% of Sun’s revenue. RedHat has a cummulative total of 206,000 subscriptions as of February. Sun shipped nearly half that many just in the 2nd quarter of ’04 (an increase of 37% over the same quarter the previous year).
Every poster so far has ignored the fact that the article
on El Reg is Ashlee Vance’s analysis of something
that Merrill Lynch put out.
Now that we’ve got that out of the way, did anybody bother
to read the second page of the article?
It makes no good sense for Sun to purchase RedHat or Novell,
those companies are worth far far more to Sun as competitors
than as divisions of Sun. (poundsmack already made the point
about monopolistic practices).
Apart from anything else, does anybody care to know that
Sun already has a business relationship with both RedHat
and Novell? Surely you understand that you can purchase
RH’s distribution from Sun when you purchase an intel-
or amd-based server or workstation. What more do you want
to show that the marketplace is working just fine?
I’m more of a fan of Linux, due to it’s affordability, open source colaborative development, and now it’s ease of use and wide hardware support.
But Sun is coming on strong on the server side with Solaris 10, and it’s new pricing structure. I reckon with superior performance and features and with a competitive pricing against Red Hat and Novell, they will start winning back some market share, and take away some of Red Hat and Novells business.
Solaris has really added killer server/enterprise features and performance. However, I’d like to see Sun do that with Solaris on the desktop side. JDS (whether SuSe or Solaris based) is a nice looking, good featured Gnome based desktop. However, Solaris is badly behind any Linux distro in terms of hardware support.
Now if Sun adds wide hardware support to Solaris (no small task), and if they add some other cool features for desktop users, they would have a very compelling product for end users and enthusiasts, and then win some mind share.
In spite of McNealy’s and Schwartz’s idiotic FUD, Sun has it’s act together in terms of R&D and putting out compelling products for competitive prices. Again, I’m more of a Linux fan, but more power to Sun if they can compete well. It’s all good for customers in the end.
and the Janus Linux layer in Sol 10 will enable users to migrate many current applications to Sol 10, so that switching will be a reasonably cheap operation for those who want to do it.
——
except that its not ready and wont work unless extreme amount of careful library versioning is done. If there is one project within the solaris 10 timeframe that will fail this is it
RedHat doesn’t have anything that SUN couldn’t take for free or redevelop. Caldera tried the linux and commercial UNIX business and didn’t get very far. SUN could just take (not even buy) a BSD and port all their stuff becoming something like Apple. Now .. SUN could also buy SGI which could add a lot of cool technologies to their camp.
why is sun talking about open source then. for 30 years a huge market worth billions of dollars with every major software company with an exception of maybe MS has a big stake on open source software and you still want to ask the same old questions. weird
Sun has more cash and investments than either of the 2.
oh lord, may we forget that the SPARC is much more popular than the opteron?
Why shut something down that’s profitable. HMMmmm
Redhat is bloated, who would want to buy them?
Remember, Jonathon Schwartz hinted ‘what if sun buys novell?’
He may be onto something. Novell is mainly a networking software company.
Redhat is bloated, who would want to buy them?
—-
all of redhat’s millions of enterprise user base
its unsubstantiated silly claims that void reality like this that happen to make advocates look childish…
“Remember, Jonathon Schwartz hinted ‘what if sun buys novell?’ ”
are you talking about the same guy who lied about redhat being proprietary and not LSB compliant. yah. he still has much credibility left now..
Ha ha. Does RedHat even come close to Sun’s product lifecycle and support durations? Also, Solaris is taking shape into what RedHat could never be. The new feature list for Solaris 10 is amazing, complete with BSD-style packet filter, new TCP/IP stack, ZFS, etc. Solaris has good documentation that is generally task-oriented and thorough. Solaris doesn’t cater to the lowest common denominator, either, while Red Hat’s installer just drives me nuts.
Just reading this, you discredited yourself because you didn’t specify which Red Hat product you compared with Solaris 10. If I remember, Solaris 10 should be compared with Red Hat Entreprise Linux 4 which is based on Fedora Core 3.
And Sun is right that Red Hat is not free. Buying “enterprise” Red Hat is thousands of dollars.
You failed to mention that the thousand dollarsthat come with Entreprise version of Red Hat include support and the source available to the public can be used by anyone i.e. CentOS, Whitebox. Try that with Solaris source codes. As an large entreprise, will you take a risk to run a product without a technical support? This is large business league you talk about.
Sun’s new pricing model is dollar-for-dollar competitive with Red Hat all the way down to free Solaris against Fedora. Sun figured out Red Hat and get revenue from hardware, too!
Fedora is free since it is an open source product i.e. Linux OS without any proprietary softwares unlike Solaris. It appears you tried to spread FUD rather than saying fact.
would be for Sun to buy the right to open-source those bits of Solaris that are licensed from 3rd parties. Then Solaris could compete with Linux on an equal basis. Solaris is already less expensive than Red Hat Enterprise Linux. The only magic ingredient that Linux has, and Solaris doesn’t, is its license.
why is sun talking about open source then. for 30 years a huge market worth billions of dollars with every major software company with an exception of maybe MS has a big stake on open source software and you still want to ask the same old questions. weird
Because their primary competition is “free”.
Because it can speed adoption.
Because it can speed hardware compatability.
Because it provides Good Will to the community.
Because they have nothing to hide anymore. AIX is dead. HP/UX is dead. Who’s left in the general purpose server space that’s innovating anymore?
They can Open Source OpenSolaris, yet still keep closed proprietary integration features and utilities (like, say, high end clustering and fail over, for example) that are layered on top of the system (like Apple and Darwin).
They can give away things like deep dark internal SMP kernel bits knowing full well that a) just because you know how to do something doesn’t make it easy to do b) some of the advanced bits will take a long time to engineer into another system, and it will take even longer to actually make a market impact, and c) if I have a choice of reimplementing those bits vs just using OpenSolaris, umm…why bother?
For example, take porting DTrace to Linux. I imagine smart kernel guys already have a pretty good idea how this system works, and if they put their heads together and worked on it, they could figure out what they actually needed to do in a few months. But they would still need to actually re-engineer the kernel to pull it off, and even then, how long before Linux DTrace would show up in the market in something like Red Hat Enterprise? (Because that’s who Sun cares about — y’all who build your kernel often, change distro’s often, or whatever, are simply not Sun’s market.)
See, knowing it can be done tends to be the “hard part”, doing it tends to be simpler. If some party (say, IBM) wanted a DTrace-like functionality in the Linux kernel, actually seeing the DTrace implementation in the Solaris Kernel would probably only shorten the actual development time by a couple of months, say 10% of total time from start of implementation to market.
Even lifting a piece wholesale (say, ZFS), and simply porting it, will take a good amount of time, just because of the complexity of the task. (If they can, since no doubt OpenSolaris will probably not come out under GPL, it may not be “liftable” into Linux, but maybe into, say, the BSDs.)
Sun doesn’t care if Linux “catches up”, because in many areas it simply can’t. Because that implies that Sun is going to say “Here it is! Sol 10/Open Solaris! We’re done!”. Sun is going to keep moving forward.
Linux folks says “If only We did X, Y, and Z, we could beat [Sun/MS/IBM/Ford/Chrysler/GM/etc.]”. But as a mobocracy, when people say “We should do…” it really means “THEY should do…”. So, Linux users float on the good will of They. They who do simply what catches their eye. That do whatever They want. They that chase the prettiest butterfly in Their meadow. Using the Million Monkeys rule (and some generous corporate support…) we have a usable system in Linux, but can’t seem to get X, Y, and Z done.
But, see, even on something like OpenSolaris, Sun can watch the marketplace, and can point to its army of coders, testers, documenters and engineers and say “Do X, Y, and Z”.
In Linux we have Brownian Motion incarnate, in Sun you have a force with direction. If something useful pops out of the teeming swirling mass that is Linux, bully! And a lot of stuff has, don’t get me wrong. But Sun has the resources to where it simply does not have to wait for something to pop up. It can look at the market, listen to its PAYING customers, and point OpenSolaris in a direction where its customers want the product to go.
So, now, the Linux vendors will have to play catch up with Solaris 10 and perhaps OpenSolaris, because S10 with their new pricing model levels the playing field. Save an application dependency or installed base, there is no reason to not give S10 serious consideration, and many reasons to prefer it over Linux. And if OpenSolaris is close enough to S10 itself, then Open Source fans will have one less excuse to not consider it, and users can have less worry about the potential for a Sun bait-and-switch down the road. (GPL fans will have to stick with Linux, of course).
In the short term, OpenSolaris may pose the biggest threat to FreeBSD. If Sun has a solid build system, and someone ports a BSD Port and Package style system, FreeBSD could be in for some serious trouble. The system build and management features, the centralized development, and the port system are some of the features that make FreeBSD appeal to many. I’m pretty confident OpenSolaris is going to have a similar centralized development meme, but who knows about the build and package system.
Interesting times, friends! Interesting times.
Your whole rant ignored that point that the reply was directed at someone who asked how opensource could make business sense and since you agree that it does it makes your detailed descriptions completely useless and redundant
Too bad solaris sucks on x86 (OpenSolaris? pfft), otherwise your comment might have been remotely valid, eh? Linux has to play catch up on SPARC, sure, but that’s neither Linux’s primary nor target platform.
I was going to make a substantial response to
your complaint, but I recognised you.
@Will: you seem to have done some thinking about
Sun’s strategies and have expressed your thoughts
very well.
Clearly you have no understanding of what Sun has
actually achieved with Solaris, no appreciation of
the engineering effort involved, and no concept of
innovation.
Saying Too bad solaris sucks on x86 indicates
that you are not interested in rational discussion.
fyi, it’s a discussion where you can’t fall back on
“but linux is soooo cool” — facts are what count.
In Linux we have Brownian Motion incarnate, in Sun you have a force with direction.
Well, if you learned nonlinear dynamics you know that both the fractal dimension and surface coverage of a Brownian motion is higher than a linear motion.
If something useful pops out of the teeming swirling mass that is Linux, bully! And a lot of stuff has, don’t get me wrong. But Sun has the resources to where it simply does not have to wait for something to pop up. It can look at the market, listen to its PAYING customers, and point OpenSolaris in a direction where its customers want the product to go.
Maybe you should have a look in the the Linux kernel changelogs where all those developments are coming from.
@Will: you seem to have done some thinking about
Sun’s strategies
Maybe, but rather inefficient for the discussion here if you do not do the same level of thinking for Linux.