The Economist on The Rise of Linux

nickco3 writes "The Economist is telling the business world that Linux is a worthy adversay to Windows and Unix. It is free, runs on almost any hardware, and generally more secure than Windows As result it is dividing the industry into winners that offer Linux (e.g. IBM and HP), and losers that don't, (e.g. Microsoft). Sun is probably doomed."

Comments

This made me confused.

[termos]

Sun is probably doomed
According to this article SUN is the one who's not doomed.

Re:This made me confused.

[jc42]

... SUN is the one who's not doomed.

Right. Note first that Solaris is highly POSIX-compliant, as is linux. This means that most software ports from one to the other with few if any problems, as long as you haven't used the private extensions of either. This isn't a conjecture; I and many others have tested this with our software. Portability between linux and Solaris is easy, almost as easy as between linux and *BSD.

And note also that Sun is actively supporting several linux distros. There was some confused news recently about Sun supposedly dropping their linux. What they actually did was drop the attempt to "rebrand" RedHat linux as Sun linux. This was mostly because customers got confused. And some of them wanted RedHat explicitly. But Sun seems to be going strongly into the linux support business, letting someone else supply the POSIX-compliant platform that runs on their hardware and has all of their software goodies available as options.

There is a strong contrast with Microsoft here: Microsoft has been moving strongly to a "total experience" platform which doesn't allow any software that isn't on their approved list. So if you're a software developer, you are facing a market in which you can only sell to Microsoft, on their terms. If you try selling retail, you'll find that your software constantly breaks, until you sign the rights over to Microsoft.

Sun, on the other hand, has a strong history of supporting independent software developers by sharing information about the innards of their systems while not requiring onerous licensing of any sort. As either a software developer or an IT manager, it's obvious which would be the wiser purchase. Why would anyone with half a brain go with a secretive, monopolistic computer system when there's another available that is open and cooperative?

And for a final note, we might observe that Sun has in the past objected to being called "SUN", since that refers to the Stanford University Network that they grew out of. They are officially "Sun", which isn't an acronym for anything. In todays environment of rabid copyright and trademark enforcement, it's important to get such things right. ;-)

as much as i like the

[minus_273]

economist.. i dont think MS is really a loser. Nor do i believe that anything else will be the dominant desktop OS. As much as we like to bash MS, they do do somethings right. Enough so that they will keep their customers. Heck sometimes they even make technological improvemsnts in their OS that others are just discovering. Look at all the research that went into Async IO in the open source world only to realise MS had done something better for nearly 10 years.

Re:as much as i like the

[Anonymous Coward]

Actually, the article, if you read it, did not say that Microsoft was losing. Rather, it said that M$ would come out less of a winner.

The poster of the root article needs to read the article. Doh.

Re:as much as i like the

[pVoid]

Windows NT's entire IO system has always been asynchronous if you want it to. And it's architecturally very sound and useful.

You aren't going to get an answer here on /. that will be sufficient for you (and if you expect one, you are just a troller/flamer)... the best way to know is to use it a bit.

A good example of why I love it so much is that *every* object in the Kernel is synchronizable. You can wait on a file just like you can wait on a thread, process, or mutex. Once you get used to the architecture, it's very sweet.

Re:as much as i like the

[NetMasta10bt]

MS is really doing less and less right. They are increasingly getting more greedy with their licenses. There security is to be laughed at. And every three years they release a rehash of the same old product.

Eventually, (and soon), Linux will have the usability that 51% of the population will find acceptable, and it will landslide from there. There will be no reason to pay for Windows XY or Office YX. Software companies will start producing the little annoying things that exWindows users like for Linux, and everyone will be happy.

Just wait until MS starts writing software for Linux. That will be a funny day.

Re:as much as i like the

[hughk]

As you know, the NT kernel is in a way a descendent of Digital's VMS. VMS in turned was based on a 16-bit operating system, RSX-11M. The interesting thing is that in the old days, propietary operating systems weren't particularly closed. RSX-11M was built from source code, so you always got the exec in source. Until VMS 4.5, the source listings were delivered on microfiche with the system. There were some very good books available on the internals of VMS too.

Personally, I think that the NT kernel seems quite ok. The trouble is that it is closed (so I can't easily understand what is going on) and it is surrounded by layers of crap. I agree that the Linux people could have learned a lot from the synchronisation objects in NT, but they could also have leaned a lot from older systems like VMS, which were extremely well documented. However, better asynch support is coming soon.

However, the next big area for Linux is resource locking (needed for clusters). NT does it, but not very well. VMS did it better (and still does, from last time I used it).

Re:as much as i like the

[rmohr02]

And their anti-piracy measures don't work. I'm downloading FP and Publisher XP now, and I've cracked Office XP Trial!

Are you proud of yourself for pirating software? Personally, I don't pirate software--there's generally a better (or at least acceptable) open source alternative. That said, I hope Microsoft keeps with the "product activation" and makes it harder and harder to crack their software, so that people like you will be forced to actually pay for MS Office, or go with a free alternative.

Re:as much as i like the

[VZ]

> Threads do not have message queues

Have you ever programmed under Win32? This is simply false. Read the MSDN docs for PostThreadMessage() function:

The PostThreadMessage function posts a message to the message queue of the specified thread

...

Messages sent by PostThreadMessage are not associated with a window.

Sun is NOT probably doomed

[Ubergrendle]

Sun is a hardware company; as much as they like to trumpet their 'value add' components and services, their bread winning business is SPARC. Java, a great language, generates little revenue and is a tool that drives need to purchase their hardware. Remember iPlanet? Approaching 0 on the latest web surveys as Apache dominates that space.

IBM and HP are making smart moves adopting Linux business models. As Linux matures and benefits from a gazillion different implementations, AIX and HPUX will begin to look less and less desirable.

Getting back to Sun, Solaris is not a revenue piece for them either. There was alot of complaining in the Slashdot crowd and Sun's commitment to Solaris on Intel has waned, but really, would you like to be running Solaris instead of Linux or Debian? Thought not...

Re:Sun is NOT probably doomed

[Temkin]

Sparc performance has always been and still is a laggard compared to the rest of the industry.

Not true. When the SPARC first came out in 1987, Sun enjoyed roughly a 4 to 1 performance advantage. Most PC's at the time were 16-bit CISC, and running between 4.77 and 12 Mhz. The Sun 4/110 was 32-bit RISC, and running at 20Mhz. The 386 came out shortly afterwards, and ran at a blistering 16Mhz. Intel PC's didn't take the lead until the Pentium rolled out. Even then, Sun's are optimized for different uses. A PC is like an F-16. It's really good at taking one person somewhere really fast. A Sun server is more like an airliner. It's really good at taking 300 people somewhere reasonably fast, and effeciently.

Temkin

Re:Sun is NOT probably doomed

[aphor]

I have to disagree.

The only thing that differentiates Sun hardware is that Solaris runs on it, and thus the multitude of Solaris hosted applications runs on it. Take away Solaris and Sun doesn't lead in anything.

What Sun sells is a scalable platform. It is a *hardware* target for developers to *write software* for. It is not *primarily* an Integrated OS/Hardware stack as you suggest. It is trivial to port human-readable Linux code to Solaris, and thus the success of Linux as a development platform does not preclude anyone's ability to implement that software on Solaris. If you can write good, MT code for Linux, then you can test it on a small sub-$1000 UltraSparc system (Running Linux, FreeBSD, OpenBSD, NetBSD... or even Solaris), or maybe a *little* bigger system if you're paranoid about testing CPU concurrency. Once compiled, that software's performance can be made to scale with the harware it runs on in a near linear way. Code for 2 CPUs and 1GB RAM, but the unaltered software will support 32x the load on a 64CPU 32GB RAM box, without re-engineering anything. Instead of worrying about the MPP (Beowulf) style clustering architecture and optimizing the software for that type of system (if it can even be done), Sun provides a stable ABI across all tiers of servers.

"Take away Solaris and Sun doesn't lead in anything." Well, that's true as far as you know anyways. It's a hardware performance niche that Sun occupies.

Sparc performance has always been and still is a laggard compared to the rest of the industry. Pricing at the low-end is wiped by PCs, and features at the high-end, where the profit margins are still fat, are wiped by IBM and HP in the commercial sector and SGI in the technical markets.

This is all very much subject to debate. First of all, you seem to have very narrow definitions of "performance" (which seems like you mean price/performance in comparison to PCs), and "features". IBM can't provide a platform with the scalability potential of UltraSparc. HP doesn't even have a hardware platform any more! Compaq killed Alpha, and HP killed PA-RISC, and even since buying Compaq, HP can't guarantee a niche for Itanium as they are *more* vulnerable to AMD Opteron competition (high-end to low-end) than Sun is to low-end-PCs in their workstation market. This situation is made MORE significant as Linux gains credibility as an OS. Sun is sitting pretty in the niche they currently occupy.

News Flash: Linux still not ready for the desktop

[lpret]

The most likely outcome is that customers will face a choice between Linux, which is cheap and cheerful, and Windows, which offers more bells and whistles, is tightly integrated with other Microsoft products and is easier for unskilled staff to use, but costs more. In short, Microsoft will be not so much a loser from Linux as less of a winner. In the server market at least, Linux is providing Microsoft with some much-needed competition.

I think this sums up the consensus of the article -- Linux is coming, but not to the desktop.

Sun isn't doomed

[Sir_Bill_William_Jen]

Sun will be around for quite a some time, because they offer specific product for one market on which they have quite a monopoly.

Linux will probably take some share from Microsoft but It won't be so huge that it will be used more than Windows. Microsoft is too smart to let Linux take hold of them. They will do everything to stop Linux from taking their share, whether it be through Advertizement, Deals with companies, or Palladium.

However Unix might fade away, BSD and Linux provide great alternatives for Unix... But they won't be completely gone.

Sun increase market share against x86 last year

[ChrisRijk]

In the "low-end" of the market (systems costing under $100,000) Sun increased their revenue market share last year (2002 over 2001). Ie they grew faster than the industry average. Almost no Intel, Windows or Linux based systems are sold for more than $100,000.

So if Sun was being hurt by Linux (or x86 based systems in general) why did their market share increase?

It wasn't anything to do with the LX50 (dual P3 Solaris x86 / Linux) systems they launched late last year - they only shipped a bit over $1m worth by the end of the year.

Oh

[KoolDude]

Sun is probably doomed.

Umm... but... I thought Doom was going to be MS exclusive

*logs off and runs away*

Re:News Flash: Linux still not ready for the deskt

[drooling-dog]

I think this sums up the consensus of the article -- Linux is coming, but not to the desktop.

Well, after years of dual-booting with Windows, Linux is the now the only thing going on my desktop, and I've gotta tell ya, I'm doing just fine (better than ever, IMHO). Maybe I'm dreaming? Or maybe the Microsoft fuddites don't know what they're talking about.

Microsoft Prices and Competition

[GregBildson]

Thankfully, Linux has prevented Microsoft from dominating the low end server market like they wanted. From the low end, Microsoft was then going to push into the high end. Can you imagine how much Microsoft would be making off this area if Linux had not stopped them?

On the desktop, we just bought 5 rather powerful developer PCs for $600 each. Of that price, $150 was for Windows 2000 (not XP thank you). 25% of the price is a rather large part of the cost. The decision between Linux and 2000 was pretty close on these boxes and getting closer all the time. Pretty soon, developer workstations could well be all Linux with OpenOffice and the like. I think the competitive threat to Microsoft will soon restrict their desktop and office pricing.

A Lesser Form of Unix

[MBCook]

"But for many applications, Solaris is overkill, and Linux, a less capable flavour of Unix, is good enough."

I don't know much about Solaris, so I'd like to ask you guys out there. What makes Linux less capable? What does Solaris do that Linux can't do (at least well enough)? Just wondering.

Re:A Lesser Form of Unix

[jc42]

While it's not nearly the whole story, an interesting benchmark is the top 500 supercomputers list. Sun and linux are both on the list. Sun seems to be the main supplier of supercomputers these days. But they first appear on the list as number 156. There are only two linux systems, but they are numbers 5 and 45.

I guess that doesn't tell you all that much about the top end, other than that linux and Sun are both quite capable of supporting raw, top-end gigaflops power. The actual computing is done by processes, of course, so this benchmark really just tells you that neither OS has any showstoppers. Neither interferes with the ability of a process to crunch bits and bytes.

Now on to the other benchmarks ...

Re:A Lesser Form of Unix

[sql*kitten]

I don't know much about Solaris, so I'd like to ask you guys out there. What makes Linux less capable? What does Solaris do that Linux can't do (at least well enough)? Just wondering.

Well, scalability is one major thing. Linux struggles with more than 4 CPUs and more than 4G of memory; Solaris handles hundreds of CPUs and a terabyte of RAM. Linux lacks the manageability of system resources offered by Solaris 9, which allows system resources to be prioritized for different tasks, with a guaranteed minimum available. And even if Linux could do this, it doesn't run on hardware than can be dynamically partitioned, unless it runs as a guest on z/OS, and in that case it's z/OS doing the work. Tight integration with the underlying hardware is another advantage for Sun; they know precisely every component in every system that Solaris runs on, because they designed and built it, so there are never compatibility issues. Solaris' high-performance, high-reliability filesystems are proven, not just betas (yes XFS is also proven, but in IRIX not Linux). Speaking of filesystems, Solaris has ACLs, whereas Linux just has the relatively crude user-group model. Linux doesn't have remote shared memory or IP multipathing (IIRC).

Sun doesn't support Linux?

[Biolo]

Ok, Sun is not 100% behind Linux (yet), but that's because Linux isn't ready for the high end (yet). By high end I'm talking about F15
and F12K
servers. It is pretty close to having the capabilities to run on the Sun midframe stuff, for example I'm sure it would run fine on a 3800,
maybe even the 4800, but you start to reach its current limit with a fully stuffed 6800 system.

Now, step back for a minute and think why Suns UltraSparc and Solaris solution is so strong. Simple, at the risk of repeating the marketing guys the lure is that you can give your development and deployment guys a bunch of cheap Sunblade 150s or some cheap UltraSparc blades and whatever they come up with can be moved straight onto anything up to and including an F15K without recompiling. Put yourself in the place of a big corporation. Your putting together a new system, you have no idea just how big a load it will eventually have to take (say in 5 years). Today, sure you could run it on high end Linux box, but what happens if 6 months in the system needs a bigger box? If you chose Sun in the first place you simply buy the bigger box and move over. No porting, no redevelopment, and you know there is always a bigger, faster system you could move to. It buys you severe scalability that Linux isn't placed today to provide.

Now, about not supporting Linux, what about the LX50, the Sun Open Desktop that is coming soon, the Lintel blades (Coming Soon(TM)) the fact that the entire Sun One stack (web, directory, identity, etc, etc, etc) is either available now for Linux or coming soon, not to mention Star/OpenOffice.

So what is the perceived issue? I think people don't see Sun offering Linux on the UltraSparc range and thing they don't get it. Sun does get it, but look at their selling point for the last 10 years, total scalability. Linux doesn't provide this yet so they can't buy into it. What they are doing is making Solaris as compatible with Linux as possible, whilst at the same time helping Linux by providing software (openoffice, SunOne and much more) and I believe some kernel code too.

Believe me, when Linux is ready for the F15K class systems Sun will be ready for Linux to be there.

Disclaimer - I work for Sun, but nothing I have said here is not already public information.

Sun, not dead, just a change of direction.

[Zebra_X]

Sun really needs to rethink their strategies.

Solaris for Intel? None of the sparc binaries run on it, it's not any faster than linux. Linux 1, Solaris 0.

Solaris for web applications... absoultely not. Tried and true OS for sure. Though web sphere, atg, web logic and most other large scale app servers have linux ports. java's relability will never exceed the uptime of an OS i.e. the JVM or app server will crash before the OS does. That then precludes having a bullet proof OS. Web applications need redunancy, both from a geographic perspective and application. Doing so requires a duplicate hardware investment. Not such a good deal with sun. Linux 2, Solaris 0.

Maintainance... will be cheaper with linux rather than solaris. The reason being that Linux and it's friendly varients are all freely avaialable. To learn and use linux is not a big deal, solaris on the other hand need solaris hardware to run. As a result of easier access to hardware and software (linux) labor costs go down because the skill sets require to administer and maintain linux and linux apps are more freely available. The same is not true of solaris. Linux 3, Solaris 0.

Solaris, and it's hardware IS good for massive multi-proc applications. Data Warehouse with Multi-tera bytes of data? Linux and Intel are not suited to such tasks. Large transactional databases that require nearly 100% uptime and reliability, i.e. the database is nearly as reliable as the Operating system. Solaris is the OS for that application. Linux 3 Solaris 1

Sun is no longer suited to playing in the high(er) growth markets of dedicated servers, web applications, IT support devices (dns, dhcp, network management) and such. Their role is increasingly being boxed into ultra highend applications where a large number of processors, ultra high reliability and what sun has stood for still means something. Where the applications are almost as reliable as the OS, and that the OS and hardware is required to be up nearly 100% of the time and never unexpectedly. The difficulty they face is that that the role just described is not in particularly high demand. As IT budgets continue to shrink - decision makers are going to continue to look to linux to solve their problems.

Linux is cheap - costs less to maintiain - and the hardware can be repurposed. Sun just can't argue with that. Sun needs a change of direction.

Sun's Doom is like Apple's Doom...

[Bug-Y2K]

Content profitability in a comfortable niche market.

Too long have people labored under the delusion that one company MUST dominate the "computer market"... only because one company (Microsoft) seemingly *has* dominated. Just because one does today, doesn't mean that is the natural order of the market place. If anything it is unsustainable, as Microsoft is beginning to find out.

Sun makes some excellent high end gear, and in that market niche they are by far the largest player. They aren't even competeing with M$ in that space... and there is plenty of money to be made there. Sure CEO Jeff won't get to cross-check Bill in the teeth as often as he'd really like, but hey... that is NOT what Sun is in business to do.

Shake the current "one must dominate" worldview out of your heads /.'ers. It won't work. Microsoft's whole strategy, both internally and externally, is "For us to win, they have to lose." You WILL lose if you play that game with Microsoft because they play it better than anybody... but if you play a different game... Steve Jobs' game... where "we need to make something of quality that some percentage of the market wants and not worry about Microsoft" then you will do fine. There are billions of dollars to be had and significant percentages of market to be owned. Sure, you won't have dominance, but you don't really *need* it.

The Economist didn't get it

[Alomex]

Traditionally, the Economist has been one of the earliest media outlets to get technology. The first printed press reference to the Internet outside the tech press is from the Economist, ditto for Linux, but this time around they are way off the mark.

The reason Linux is so popular is not that is free. BeOS is also free. Linux is successful due to convergence of many different factors:

(1) Free
(2) Open source
(3) Unix compatible
(4) X-windows (X11) compatible
(5) designed for x86 (yes it runs on many other chips, still Linux is an x86 project from the get go)
(6) Multiple vendor supported
(7) Plenty of third party support

Moreover each of these things feed of each other. That is why Linux is so popular.

Apache a CLONE??

[cmacb]

"According to Craig Mundie, Microsoft's chief technical officer, as the open-source movement grows, it will get better at producing free clones of commercial software."

I give up. What is Apache a clone of? I wish they had included an actual quote on this. Maybe the journalist was "interpreting".

Most of the article is on target though. The easy way to evaluate the strengths of the companies mentioned is to look at how diversified they are (or aren't).

IBM is no longer primarily a hardware company. They have a strong consulting division, they do fundamental research and grab pattents on REAL things (rather than new parsing algorithms like some companies I can think of), they have a very strong software development component, they farm out hardware manufacturing that is no longer profitable (disk drives) while hanging on to things that they do best and can make money on (chip fabrication).

Sun is primarily a hardware company. Their operating systems are (almost) exclusively sold to customers who use their hardware. Java and Star Office are far from being cash cows. Their weakness is that as Intel, AMD, etc, chips get cheaper there is less incentive to use Sun's higher priced hardware components. Supporting Linux helps them a bit, but it is the price of their hardware that puts them at a disadvantage.

Microsoft is a software company. They are trying real hard to become something else too, but like Sun they are having a heck of a time making anything else work. They don't really do fundamental research, but instead try and grab patents on programming concepts so that they can bully other companies in court when it suits them. They don't really make any hardware, but instead stamp their logo on a few things to make it seem that they do. They do select good subcontractors for mice and keyboards, I'll grant them that. Everything they do except Windows and Office lose money. Prospects for either of those (because they are already so successful) can only go down. They currently have a scatter shot approach to the "next big thing" which consists of trying everything at once and seeing if any of it takes off. Few companies have the money to do this. But they will bleed themselves dry rather quickly if they are not carefull. Something tells me they are not going to be carefull.

Apple is trying to diversify too. Since they are starting small the only way they have to go is up. It would be nice to see them further popularize the power-pc server. My personal experience with OS X is that they are rushing versions of it out the door too fast. I've decided to wait for XI (or whatever they call it) and switched to running Linux on my iBook. Those gel buttons are cute though.