Channel Zone Editor Steven J. Vaughan-Nichols thinks he’s finally figured out what Sun really wants to sell on the x86 platform—and it’s not Linux.
What Sun Really Wants to Sell in the x86 Market
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSAlert.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
… and the winner is: Linux on x86.
Good. Perhaps Sun will finallyfigure that out too, and reinvent itself (as J. Schwartz would put it – he likes trendy words) around this new idea.
Funny how previously Sun seemed to think (and acted like) Micro$oft was the Big Satan, and now that they “settled” with Micro$oft, they have turned their guns on RedHat…
It is amazing… SCO is having its marketshare eaten alive by Linux. So what does SUN do? uuuuu lets market Solaris x86. HOW STUPID CAN YOU GET!
But this was expected. While I was speaking to a Sun evanglist at a conference I saw that Sun just does not get it. Gee, if they do not like Linux, then do an Apple move and use one of the free BSD’s.
Oh well, it is about time that Sun is put out of its misery and Scott McNeily is doing a good job at that!
Maybe split the company? Sun hardware like the Sun Rays, and the other does Java.
Their market share is being eroded by Linux.
They can’t outprice Linux/x86.
Out-featuring has not seemed to be effective & will become more difficult.
Open sourcing to cut costs works for the underdog but for a company that already paid the R&D money open source would be suicude (think Adobe profits after making Photoshop GPL).
If these options won’t work and they are being pressured to do something what else do they do?
They won’t beat RH + IBM/Dell at Linux/x86 partly because SuSE is now owned by Novel.
Many (most?) Solaris developers already develop on Solaris/x86, so this makes the most sense.
When Sun releases the free version of Solaris 7.0 for x86 I was one of those who ordered the CD. I was intrigued to find out what a “real” UNIX was like. My curiosity lasted a very, very short time-because even though my PC had incredibly generic wide-spread components most things were unsupported. I know the situation has improved somewhat in the more recent releases of Solaris. But at that time I could a) not get on the interenet because the ethernet card was unsupported b) sound-forget it c) 640×480 16 color CDE was a nightmare and d) for desktop purposes the machine ran slower than molasses flowing uphill.
I am glad to hear that perhaps Sun is actually making a decision. I really am not that concerned with which decision they make as long as they make one. I imagine that the newest Solaris with GNOME/JDS will probably be a good workstation system and a great system in a Sun network. But Solaris will never have the hardware support for Solaris that Linux users take for granted.
Solaris x86 is, and shall remain, only a viable option on custom boxes built for Solaris-whether from Sun itself or from OEM’s, unless it is strictly used as a headless server. And this fact delimits it’s usefullness and it’s ability to compete with Linux.
There are no exact numbers out there but according to my experience Linux hardware support follow this rule of thumb-99% of all hardware greater than 5 years old is supported, %97 of all hardware greater than 3 years old, 95% of all hardware greater than 2 years old, approx. 90% for hardware produced in the last year.
Although I use Linux primarily for years now I still have managed to by many non-supported products for my PC-mistakenly having believed that it was already supported-my mistake. Two years ago I purchase a toshiba laptop 5005-S507. Although most of the system worked well sound support was weak, pcmcia was no go, and firewire did not work.
I have been in for a treat in the last two months. Recent 2.6 kernel improvements have made a profound difference-sound is not fully supported-better than in windows, pcmcia is now working, which means I can finally use the wireless netgear 54/22/11 mbs card I purchased 12 months ago. Firewire is still unfortunately not working. Last week I finally found drivers for an Intel webcam which I purchased two years ago.
Now some might wish to complain about Linux when confronted with such. I no longer do. It was Toshiba who made the decision to create this “BIOS free” , “legacy free” laptop, and I for being foolish enough to be oohed and aahed by it’s promising potential. Toshiba did everything in ACPI and to make things worse they did not follow the specs in implementing it and of course they did publish any info for developers.
Renegade Toshiba employees pissed off at Toshiba’s lack of Linux support got some of this information out and it just took time for this stuff to work it’s way into the kernel. This laptop has bad hardware support even under Windows XP-for which this laptop was designed. I will probably never purchase a Toshiba laptop again-unless they wake up and smell the coffee regarding Linux. To summarize:
Even though this machine was designed for windows XP the hardware support has been incredibly problematic.
Linux support at the time of purchase allowed me to use the machine as a desktop -but as a laptop-ie. no acpi-no sleep mode, power management , and no pcmcia. Approximately half of the machines functionality was unusable-although the most important things worked-nvidia graphics, sound did work albeit sub-optimal, the dvd/cd-bruner worked fine, keyboard worked minus the extra toshiba keys and so on.
Linux support now has almost everything working good (2.6.6-mm4+reiserf4fs). I am still trying to get suspend-to-disk setup(I am now quite confident that it will work), and firewire is still no go. But not sound is really good, power management is working fine, I now have programs which allow me to customize the use of the extra keys /in short it has become a great Linux laptop.
Even the newest versions of Solaris support less than 10% of the hardware which Linux supports. This means Solaris x86 is an absolute no go on most consumer PC’s. This is not bad thing for Sun. Sun never wanted to produce a commodity OS, and they never wanted to have their OS running on consumer PC’s. The overwhelming majority of PC’s and PC hardware run Linux good. Linux is primarily targeted at consumer PC’s. With Linux I can take advantage of the hardware I already have-as an administrator at the local university I have 20 ancient PC’s setup in a LTSP solution. The folks who use it would never believe that 486DX33’s with old matrox graphic cards form the basis of their desktop experience. Sun has nothing to offer with Solaris x86 for the situation in which most people and companies find themselves in-ie. already having legacy hardware and not wanting, or not being able to dish out money for the newest generation of hardware. This is true in the US and far more so in the rest of the world.
>> I imagine that the newest Solaris with GNOME/JDS will probably be a good workstation system and a great system in a Sun network. But Solaris will never have the hardware support for Solaris that Linux users take for granted.
That comment says it all. Solaris will not have the hardware support that Linux users take for granted. I use Linux and am mostly happy with the driver situation. BUT there is still room for improvement. Solaris will offer even less, makes me just scratch my head and think, “Sure Sun guys this is going to swim like rock in a lake”.
He’s probably right. I work for a Sun competitor doing Linux stuff, so perhaps I should be happy.
If I was an existing Sun customer and Solaris user, then I might consider going with Solaris on the Opteron. Otherwise, I can’t see any obvious reasons to “Switch”, as the Apple advertisements sugest.
Linux has the following obvious advantages: A) Better hardware support. B) Better software support. C) Open Source. D) Vendor neutral. E) Cheap.
Apple was able to create an utterly kicking operating system with OS X, so it can be done. Unfortunately, it can’t be done by Sun because they are not Organised. There is too much infighting going on at Sun and the CEO doesn’t have any kind of plan.
That comment says it all. Solaris will not have the hardware support that Linux users take for granted. I use Linux and am mostly happy with the driver situation. BUT there is still room for improvement. Solaris will offer even less, makes me just scratch my head and think, “Sure Sun guys this is going to swim like rock in a lake”.
So you some how have some “secret information from the inside” that SUN is doing nothing about Solaris’s hardware support issues? please, cut the bullshit. SUN does’t go out and hire a few hundred hardware device writers because they think it would be a nice idea, support IS being added.
Want confusion? look at Red Hat, one day they’re after the desktop, other day, “oh, we’re not ready”, then again they’ll say they’re going for the desktop. What are they? arthur or marther? are they coming or going?
> Linux has the following obvious advantages: A) Better hardware support. B) Better software support. C) Open Source. D) Vendor neutral. E) Cheap.
Wrong on pints B and E and on point C just being Open Source doesn’t necessarily count as an advantage since probably 99% of enterprise customers will want the vendor to resolve the OS-related issues, therefore being open or close source is a moot point in the overwhelming number of cases.
On piont B, Solaris x86 will have no worse and possibly even better software support than Linux considering the number of ties Sun has with ISV’s and provided in the most instances it will be just a recomile of a SPARC version.
O point E, Solaris x86 is already cheaper than any other enterprise Linux offering to purchase and support and I forsee it stay the same, since Sun intends to keep it very competitive with Linux and Windows.
It’s a no brainer… switching from Linux to Solaris is so obvious if Sun do walk the walk this time.
superior technology, consistent and powerful… why would a *nix user not switch?
solaris on sparc mmmmmmmm very nice
solaris on x86 mmmmmmmmm stinky stinky
hardware support is almost non-existant at the moment and I cannot see that improving much
but the worst thing is that solaris x86 is so slowwwwwww
and I mean really slow
in fact it is slower than a bagfull of slow things
Sun can’t make up it’s mind about linux & Open Source. They are as wishy-washy as Charlie Brown.
The choice is good, but Solaris won’t last. As SJVN points out Sun is only trying to sell Solaris x86 to get people to pony up for Sparc. Sun has a good piece of software that they are screwing up royally with. JDS while visually nice, doesn’t even mention that it’s an enhanced version of gnome with a new theme. Sun’s Linux offering doesn’t even mention that it’s Linux at all.
Choice is good, but Solaris is slow on anything not sparc, and JDS lies about where it came from.
As I said Sun is wishy-washy
> Choice is good, but Solaris is slow on anything not sparc
This is no longer true, Solaris might have been a little less efficient on x86 hardware before, but it is much improved with the recent releases, which makes Solaris to be pretty much on par with Linux. With release of Solaris 10 I expect Solaris to be even faster than Linux especially in respect to network performance — the antiquated STREAMS based TCP/IP stack will be completely overhauled. I already run Solaris 9 and 10 on my Intel machines and I have nothing to complain about, Solairs is just more comfortable to work with.
So you some how have some “secret information from the inside” that SUN is doing nothing about Solaris’s hardware support issues? please, cut the bullshit. SUN does’t go out and hire a few hundred hardware device writers because they think it would be a nice idea, support IS being added.
Linux already has the required level of hardware support. What on Earth is Solaris going to offer? That’s the question people are asking themselves.
You cannot possibly hope to compete against the open source community by hiring several hundred hardware device writers – Sun are as broke as it is. You have to use the community, and one minute they are Sun’s friend and the other its enemy. Microsoft can do all that because they have the resources, Sun can’t. It is also pointless because customers and the IT community are looking at what is happening with Linux, and then looking at what Sun is doing and asking what the point of Solaris is. Windows is sufficiently different, and has a closed world of software, that has insulated Microsoft to a certain extent. Solaris offers nothing compelling, and the additional features Sun is touting aren’t making any difference either.
Want confusion? look at Red Hat, one day they’re after the desktop, other day, “oh, we’re not ready”, then again they’ll say they’re going for the desktop. What are they? arthur or marther? are they coming or going?
Red Hat are right behind Linux (they are a Linux business) and they use open source software and contribute to projects. They have the fundamentals right, and above that there has been some dithering about the desktop, but that is to be expected to an extent. Windows has over 90% market share, and it depends on what appetite there is for change.
Sun are trying to prop up an OS in Solaris that has really lost its way, and trying to use Linux in some way to channel people back to Solaris? I don’t know, but Sun need one OS and one base to work from.
If Sun said that they were moving to Linux and were going to amalgamate some of the Solaris system into it, then you’d see Sun’s share price and momentum go through the roof. It would remove a lot of confusion for customers and investors alike, and I think IBM and HP would be really worried because Sun would have everything under one roof.
Sun have also come up with a Looking Glass desktop, and no one has any idea what it will run on or what system context it will run in. Sun has also confused things by making it look open source, when it actually isn’t, and doesn’t seem to know what direction Gnome will take in their plans. It is no use whinging about Mono – if you want to make Java work on the desktop then do it!
Sorry Sun, but you can’t fake open source software.
This is no longer true, Solaris might have been a little less efficient on x86 hardware before, but it is much improved with the recent releases, which makes Solaris to be pretty much on par with Linux.
Can’t see that happening at all. Linux has been in development on x86 hardware since it was created (well over a decade). You can’t just bolt that onto an OS designed pretty much exclusively for SPARC overnight.
>>Apple was able to create an utterly kicking operating system with OS X, so it can be done. Unfortunately, it can’t be done by Sun because they are not Organised.<<
But apple didn’t create an OS for the x86. Controling the “whole widget” that makes hardware support, orders of magnitude easier for Apple. Sunw has no problems with hardware support for SPARC.
> You can’t just bolt that onto an OS designed pretty much exclusively for SPARC overnight.
Solaris was developed platform independent from the start, this is why Sun was able to bring the x86 version in line with Sparc so quickly. I don’t see any reason why Solaris would suffer performance wise-because of its Sparc roots.
Sun has nothing to offer with Solaris x86 for the situation in which most people and companies find themselves in-ie. already having legacy hardware and not wanting, or not being able to dish out money for the newest generation of hardware. This is true in the US and far more so in the rest of the world.
Very true, and Sun doesn’t really seem to understand the impact of this – x86 is commodity hardware, and there are many different setups and permutations. After the review of JDS many said that it was not necessary at all for Sun to include anything better than a 2.4.19 kernel, and that organizations were not going to use the latest PC tehnology.
That is true in many ways, but many replace their PCs on a ‘one-by-one’ basis, so the odds are that they are going to buy a new PC in addition to their old ones. What happens when the JDS doesn’t run effectively, or at all, on it, and what happens when they have to deal with Sun’s abysmal support? When you consider that this situation will be several hundred times worse with Solaris running the JDS, what’s the point of Solaris on x86? Yes Sun sells pre-configured x86 boxes, but you need the general flexibility from the OS to make a success of it.
It’s a no brainer… switching from Linux to Solaris is so obvious if Sun do walk the walk this time.
Because no one is switching from Linux to Solaris, as can be viewed by the financial situation Sun is in. What kind of walk is Sun going to perform?
superior technology, consistent and powerful… why would a *nix user not switch?
What superios technology? No one has experienced anything compelling enough to stay with Solaris, and that is why Sun is losing money.
> If Sun said that they were moving to Linux and were going to amalgamate some of the Solaris system into it, then you’d see Sun’s share price and momentum go through the roof.
There is one company called SGI that have already tried that, check how far their share price ended up… Sun has got by far the biggest 64-bit installed base in the world on Solaris, do you think they are going to just drop the ball and walk away from that?
> No one has experienced anything compelling enough to stay with Solaris
What the hell are you babbling about? Solaris is at least a mile ahead of Linux in pretty much any department (scalability, features, ISV support, you name it…). With Solaris 10 it will 2 miles ahead of Linux. There is place for Linux and there is place for Solaris, get over it…
Hey guys,
Woah, unleash the linux attack dogs…jeez. Before I start, let me just say that yes, I am a Linux user. Currently trying to get Slackware-current + swaret up and running, Gnome 2.6 is pretty cool (although I diagree with the decision on MIME sniffing in gnome-vfs in 2.6, but that’s just IMHO).
Anyway, my point is, how many of the guys here have actually tried a *recent* build of Solaris (e.g., Solaris Express 5/04), before spreading your FUD.
raver31: Sure, in times past, Slowaris may have been a deserved title (at least on desktop, not on server however) but from my recent experiences, CDE is incredibly fast and responsive – much nicer than either KDE/GNOME (although I concede it’s seriuosly fugly – Sun has to hire some better artists =) – speaking of which, what’s up with art.gnome.org – it’s been a few weeks already…)
So before you guys start posting crap, why don’t you check out http://wwws.sun.com/software/solaris/solaris-express/get.html, and download it for yourself (just get CDs 2 and 3 – only about 700 Mb zipped).
Hardware support: sure, official support in Solaris x86 is pretty pathetic, but if you’re prepared to read mailing lists and compile stuff yourself (and if you’re a linux user, you should), then it’s not actually all that bad. For example, http://www.tools.de/solaris/xf86/ has the XFree86 4.4 drivers for Solaris x86 – basically if it works under XF86, it will work under Solaris.
Also, Solaris 10 includes support for USB 2.0, wheel mice, more drivers etc. Sure, it took them a while, but they’re getting there.
I still use Linux (and my leaked longhorn March build, hehehe) for my everday stuff and general mucking around, but I’m seriously thinking about using Solaris x86 for real work. (at least when they release OO for Solaris 10).
Just my 2c.
I have been a sun user since 1993, my sun id is ddonley, look me up on there.
The last attempt I had at solaris on x86 harware was solaris 9. After I seen how bad it was I immediately removed it and reinstalled linux.
You said CDE is ugly.. I would say that I was almost sick when I used it, it might be that I am colour-blind, or it might have been the load of beer I had while installing, but I did not like the experience.
Anyway…Solaris belongs on sparc hardware, simple as that.
Sorry to burst everyones linux bubble, but most mid to large network admins cannot “bet the shop” on linux, they need a company behind their OS, they need to be able to call someone at a moments notice, and have support, or have someone flown in without hassle. Sun has that, and has always had that. Linux is only “eroding” the small company marketshare where they can not afford any support like that (be it Microsoft support, Sun support, IBM support or even Redhat support). When you start adding support to the mix linux becomes very expensive to the point where the savings on the hardware does not really matter, it comes down to the extent of the support and the ease of the tools that are available for the job the server will be doing.
Case in point…if you have a major problem on your 20 server cluster which was built on cheap x86 hardware and Slackware linux, and you cannot find a solution, with opensource what do you do?
You post on a newsgroup, mailing list or forum and wait for help…now, if you tell your boss that you posted on a newsgroup and do not know when an answer will be given I guarantee you will not have your job anymore or linux will be re-evaluated for any future projects. The “suits” (as I call them) do not want to hear something like that, they want to hear “I have put a call into Micrsoft/Sun/IBM/whoever tech support and they are on the problem with our 2 day SLA”
Linux, as a free solution to enterprise problems, is a fad (brought on by the crash of 2000…the rush for cheaper IT). Once the companies get burnt on support enough (opensource is only as supported as the amount of developers willing to devote their freetime to the project) companies will demand support and demand accountability for problems. Thats why Redhat-IBM/SuSE-Novell are rushing to be the first that can offer real support and real accountability even for opensource (probably why sun is rushing to get on x86 too). When that happens (hint-its already started), Linux will cost just as much as Windows, Solaris, MacOSX Server, etc.
—–
Thats my 0.02, dont spend it all in one place
Fair enough. I don’t have any experience with any of the older Solaris versions, and I concede to your greater experience on this. From what I hear though, it’s not exactly something that you’d miss. (Although I’m happy to be corrected on that last point).
Perhaps you should try Solaris 10 (link in last post, two posts up), it’s only a 700Mb or so download. From a complete Solaris n00b (me), who has some degree of familiarity with Linux, I didn’t find it all that bad (although some of the new features, e.g. zones are that documented – blastwave.com has a nice article on zones, though). Anyway, since you’re experienced with Solaris, you might make more of the new features, most of which go right over me – although the names sound cool =)
CDE: When I first started it up, in 256 and 640 x 480, it was completely horrendous. Anyway, I installed the XF86 porting drivers, and changed to 24-bit 1024×768, and things improved dramatically. Sure, it’s not a great looker (nothing on Plastik on KDE), but it does the job (which I suppose is its target market, versus people like me who like eycandy), and I find the UI quite consistent and well laid-out.
Anyway, you obviously have experience with Solaris, as well as having running it on ‘real’ hardware (SPARC), as compared to our toys (x86) =). Could you point to any useful resources for learning Solaris, coming from Linux? (I already know about the Bigadmin site).
Thanks,
Victor
Linux will cost just as much as Windows, Solaris, MacOSX Server, etc.
The TCO of Linux and Windows is arguable; both sides claim lower TCO. IMO each company should determine for themselves which gives they a lower TCO in the long run. There are too many variables to make bold claims like some (sponsored) analists make. But more importantly, with Linux you have a choice in companies that provide you support. Don’t like the way Red Hat gives support? Go to Novell, and vice versa. It is not hard to switch with Linux. With Windows you are at the mercy of Microsoft.
A good example is the mobile phone business. Former state-owned telecom companies in Europe that had monopolies generally give customers better service nowadays because they are afraid that their customers will switch. The ability to switch, even if you don’t really want to, is a powerful weapon for customers and especially companies.
Solaris x86 is a leper port, Sun has had on/off again ideas about this product…you really have to wonder who is making a long term investment in it. Seriously, if you even are adopting Solaris at this point, chances are you are going with the Sparc variant.
you appear to have things totally incorrect.
you seem to know that there are companies like ibm/suse/novell etc who do offer linux support, but you seem to be equating this with installing a free version of linux (slackware) and expecting NOT to pay for support.
there can be NO comparison done between tco of windows v linux for one simple reason… you want to have support, you need to pay for support.
you said companies cannot afford to “bet the shop” on linux.. how safer are they betting on a single company, be it microsoft, sun or apple ?
any of these companies could burst tomorrow, then who would do the support/upgrades/patches etc etc> I cannot see apple fixing any screw ups microsoft make nor sun fixing any apple screw ups.
now with linux it is far different. you install redhat, you buy support from novell, (they might be cheaper or a local support team), or you install mandrake and get support from ibm. you miss the point that with it support it is the actual SERVICE that you are paying for, and if there are companies who need it, then there will be companies that charge for it.
NOW – with microsoft, you will pay their support fees
same with apple and sun etc etc
BUT linux enterprise support is still a young entity, so when it finally settles down, there will be a few companies offering support at a much lower price than the few that do it now… will people use them ? who knows, people might pay inflated prices purely because they think they will be getting better service, this happens in all industries.
I will stop typing now, to let you all think over what I said.
sorry victor, I can’t point you anywhere to learn the ins and outs of solaris.
I use a few of them x86 toys, and personally I do not own any sparc hardware. I just use them at work. How I learned it was from using it in uni, just loads of years experience in it.
If I have a choice, I always ignore solaris and go with an x86 solution, yep, sometimes even windows.
Sorry to burst everyones linux bubble, but most mid to large network admins cannot “bet the shop” on linux, they need a company behind their OS, they need to be able to call someone at a moments notice, and have support, or have someone flown in without hassle.
What? What on Earth do you think Novell, Red Hat, IBM and HP are providing, and have been doing so for years? There a huge number of mid to large admins doing exactly that, which is exactly why Sun are losing money. Please understand how Linux is being used in this context before posting.
When you start adding support to the mix linux becomes very expensive to the point where the savings on the hardware does not really matter,
There are many mid to large admins who wouldn’t agree with you, which is why Sun is in the red. Saying that isn’t going to change the situation.
Linux is only “eroding” the small company marketshare where they can not afford any support like that (be it Microsoft support, Sun support, IBM support or even Redhat support).
Nope. If Linux is only eroding the small company market share of other systems, and Sun is safe and sound at the high-end, why is Sun deep into the red? There are any number of commercial support options for Linux-based systems, from having someone come in to support you right down to doing it yourself.
However, you’re missing one crucial element in this. Sun is losing money hand over fist which means that these arguments aren’t working.
Case in point…if you have a major problem on your 20 server cluster which was built on cheap x86 hardware and Slackware linux, and you cannot find a solution, with opensource what do you do?
You don’t use Slackware for that purpose, and certainly not for clusters. You go to IBM or HP using a commercial Linux distribution and get it done that way. You can use free Slackware or something else to fill in the gaps where it is not cost-effective to pay for that kind of support. I cannot believe that people still don’t understand how this works.
The “suits” (as I call them) do not want to hear something like that, they want to hear “I have put a call into Micrsoft/Sun/IBM/whoever tech support and they are on the problem with our 2 day SLA”
Yyyerrr… That’s why IBM, HP, Red Hat and Novell/Suse do support for Linux-based systems – and they do exactly the above. You can use the flexibility of free software around that. There are many people who have already learned this basic stuff.
Linux, as a free solution to enterprise problems, is a fad (brought on by the crash of 2000…the rush for cheaper IT).
It is not a free solution – but it is as free as you want it to be and you get the flexibility you need from that. Software being free means a great deal more than it being monetarily free. Microsoft have come out with exactly the above phrase to describe Linux – and it hasn’t made a blind bit of difference.
And just in case you hadn’t noticed, there is still a rush for cheaper IT.
Once the companies get burnt on support enough (opensource is only as supported as the amount of developers willing to devote their freetime to the project) companies will demand support and demand accountability for problems.
You’ve just described the classic problem with proprietary software and hardware. Proprietary software is only as supported as the vendor that can be bothered to do so. And you only get one vendor.
Thats why Redhat-IBM/SuSE-Novell are rushing to be the first that can offer real support and real accountability even for opensource…
You’ve just invalidated your entire post. These companies are providing that support. It is the flexible nature of open source software that really brings the advantages and cost savings to everyone. Collaborative development on the Linux kernel and other software works.
When that happens (hint-its already started), Linux will cost just as much as Windows, Solaris, MacOSX Server, etc.
When what happens? Companies have been getting support from IBM, HP, Red Hat and others for Linux for ages. That started years ago. Your point is?
Considering that Windows, Solaris and MacOSX require that level of support as well (as well as Certified Partner Services, support contracts themselves, anti-virus software, separate software for mail and database servers and Client Access Licenses, all especially in the case of Windows), then no, Linux won’t cost as much at all.
King Canute never did make that tide go back.
While linux has thousands of drivers,
Drivers have the following disadvantages:
a) binary modules are detested even if they work better
b) impossible to develop binary drivers because kernel keeps changing on a daily basis
c) GPL and other restrictions prevent IHVs from releasing their own linux drivers
d) open source drivers can only be improved if docs are freely available.
The question is who do you think knows hardware better?. Open source driver developers or the guys who actually developed the device?
It’s like even if Nvidia gave the XFree developers all the docs, their binary driver will still outperform Xfree’s nv driver.
These are basically negative points for Linux – the fact that hundreds of drivers have been cobbled up by open source developers means squat!.
Solaris’s advantages:
1) Solaris provides IHVs/ISVs a stable api for drivers – you develop a driver for Solaris 7 and it works unmodified on Solaris 9 – can’t say the same for Linux even between point revisions (ie binaries from Linux 2.4.21 will not work on Linux 2.4.22). A lot of the nic vendors have drivers for Solaris – realtek, broadcom, intel, 3com etc.
2) IHVs don’t have to open source drivers and therefore can create excellant drivers for Solaris x86.
3) Sun’s going to show ISVs and IHVs that there’s money to be made because Solaris ain’t like Linux community where everybody wants just free stuff.
4) If Solaris took off, I think Sun could really get IHVs to support Solaris by throwing a few bucks here and there.
The bottom line is Solaris is a platform where money will exchange hands and create a vibrant economy. I’d definitely like to see this play out.
“Want confusion? look at Red Hat, one day they’re after the desktop, other day, “oh, we’re not ready”, then again they’ll say they’re going for the desktop. What are they? arthur or marther? are they coming or going?”
sun has stopped and started solaris three times on the x86.
Redhat never went for the consumer desktop market. get a clue. redhat linux wasnt for the desktop. fedora isnt designed for the desktop. only recently did they come out with their corporate desktop product.
x86 is a commodity market. solaris will not fly there
”
1) Solaris provides IHVs/ISVs a stable api for drivers – you develop a driver for Solaris 7 and it works unmodified on Solaris 9 – can’t say the same for Linux even between point revisions (ie binaries from Linux 2.4.21 will not work on Linux 2.4.22). A lot of the nic vendors have drivers for Solaris – realtek, broadcom, intel, 3com etc.”
a lot of drivers dont exist in solaris because sun has been changing their mind everyday unlike linux.
“2) IHVs don’t have to open source drivers and therefore can create excellant drivers for Solaris x86.”
read the gpl. there are lot of binary drivers for linux.
“3) Sun’s going to show ISVs and IHVs that there’s money to be made because Solaris ain’t like Linux community where everybody wants just free stuff.”
who do you think is paying for oracle redhat novell and ibm
free as in freedom not price
”
4) If Solaris took off, I think Sun could really get IHVs to support Solaris by throwing a few bucks here and there.
The bottom line is Solaris is a platform where money will exchange hands and create a vibrant economy. I’d definitely like to see this play out.”
does that imply that ibm making millions is a myth. redhat enteprise and stuff are paid. oracle costs money. there is a lot of money to be made in linux. commerical companies arent here in linux for donating stuff. get a clue. they are making money much more than solaris
The question is who do you think knows hardware better?. Open source driver developers or the guys who actually developed the device?
That’s a very good question, and invariably it is the open source developers, because they are taking the time to get it working properly. The only thing they don’t have is the docs.
Why is this? Because the vast majority of manufacturer drivers are literally thrown out of the door without adequate thought or testing. Wireless drivers are a good example of that.
These are basically negative points for Linux – the fact that hundreds of drivers have been cobbled up by open source developers means squat!.
It does, because a large proportion of them work extremely well. You’re touting Solaris and it has absolutely none of this, and won’t have any of this if they can’t attract IHVs.
3) Sun’s going to show ISVs and IHVs that there’s money to be made because Solaris ain’t like Linux community where everybody wants just free stuff.
You need free stuff to get started, otherwise there will never be a market to attract the IHVs!
4) If Solaris took off, I think Sun could really get IHVs to support Solaris by throwing a few bucks here and there.
IF it takes off? What is going to make it take off? You are describing the chicken and egg situation that Linux is currently getting over. Solaris won’t attract IHVs unless there is the wider market, so the Solaris driver situation will remain the same.
This is the problem Linux has gone through and why there are many open source drivers.
“The question is who do you think knows hardware better?. Open source driver developers or the guys who actually developed the device? ”
buddy. a device is a hardware. writing a driver is software.
its not that everybody who makes hardware has a brilliant software development team too. given good specs i am willing to bet that open source drivers would be better
“The question is who do you think knows hardware better?. Open source driver developers or the guys who actually developed the device?
It’s like even if Nvidia gave the XFree developers all the docs, their binary driver will still outperform Xfree’s nv driver.
These are basically negative points for Linux – the fact that hundreds of drivers have been cobbled up by open source developers means squat!.”
SW, you must live on another planet. On this planet hardware support makes or brakes an OS. If Sun was getting drivers from Nvidia, ATI, HP, Canon, Epson,Creative Labs, etc. you might have point. But this is not the case. Sure there are very specific companies providing drivers for Solaris for their top of the line products. But how many? The fact is Solaris does not and will not support the incredibly broad range of consumer hardware that Linux already does. The “disadvantages” which you attribute to Linux have done nothing to change this situation. Sun can hire as manz device driver developers as they want-good luck catching/keeping up with more than thousands in the FOSS community.
a) binary modules are almost universally disliked-but probably a 1/3 of all Linux installs are using binary propietary drivers.
b) the igh pace of change in the kernel does proove a challenge for the propietary binary driver developers-open source drivers are usually very quickly fixed, within days, and of course the vast majority of drivers are not directly impacted by these changes.
c) There are many companies releasing the code to their drivers. Sure there are many who refuse, but less and less over time.
d) this is true. What a shame more companies ddon’t properly document their own products.
If Nvidia gave the FOSS community access to the documentation necessary to write drivers the FOSS community would quickly implement drivers which are a) more bug tested b) more stable c) more quickly fixed d) more integrated into the whole Linux system(FB, Cairo-opengl etc.). And it would save them lots of money and hastles. One day they will *grok* this too.
Like I said earlier- if you are talking about using Solaris x86 with custom boxes built for Sun thats’s one thing- if you are talking about Solaris x86 competing with Linux in commodity PC’s thats another thing-there is no competition.
It actually pains me to make this point-as much as I like alternative OS’s hardware support is the ultimate stumbling block- If alternative OS’s can’t use the FOSS drivers due to their licenscing then they must start out with tremendous amounts of cash to sign contracts with all of the major hardware manufacturers-like Apple does. Sun is not going to do this for Solaris x86-they might try, but they don’t have the money to make such happen. Besides since when has Sun ever expressed any interest in providing a commodity OS for commodity PC’s-never.
i think sun should concentrate on solaris/sparc because it performs the best for sun and due to the mounting losses sun probably cant afford to do both solaris/sparc and solaris/x86.
http://www.washingtonpost.com/wp-dyn/articles/A46497-2004Apr2.html
http://www.hoovers.com/sun-microsystems/–ID__14833–/free-co-fin-f…
Sorry to hear you are having trouble with fedora core 2(hint-they were the ones who released 2.6 kernel with 4k stacks). If you are having such trouble-recompile your kernel.
I doubt seriously fedora munges your partition table-it is more likely that you munged your partition table, albeit you have my sympathy if you trashed your partitions.
Never heard of USB crashing a system on disconnect-I have been following things pretty closely for many years now-sounds fishy.
You call it chaos-I call it creativity.
So you are a frustrated Linux user dreaming about Solaris x86 ?. Hmmm I almost feel for you. You probably have just about enough experience with Linux to be really frustrated.
There never will be a Linux monopoly-not ever, even if it were to deserver such through superior code and functionality. Linux is NOT *a* company. Only companies can form monopolies.
Linux(FOSS) has done more for increasing the diversity of alternative OS’s than anything else. The FOSS developers who contribute to Linux have also made things like Syllable, OpenBeos, Reactos, GNUstep, SkyOS etc. possible.
OSS is an integral part of MacOSX-the main system Mach + FreeBSD userland utilities-printing-take a look at CUPS, Safari take a look at KHTML.
Your experiences with *a* distribution of Linux does not represent Linux as a whole. The transition to the new kernel was and still is a massive undertaking-I have been compiling my own kernels tracking the development through the 2.5 series and am now running 2.6.6-mm4. I may be a bit obsessed-but it has been a fascinating journey over the past months. The 2.6 series will probably be able to put 2.4 out to pasture within the next 2-3 months-it is already 90% of the way there.
“linux apologist” wow- that’s new, how original.
“You Linux appologists better gear up for a good ass kicking by MacOSX, Longhorn, FreeBSD and Solaris in 3 years time.”
Gee I am so scared-I am sooo worried, the thought just makes me shiver.
In 3 years time Linux will have relegated MacOSX to a distant third place on the desktop. Longhorn won’t even be out. Solaris x86 will have finally caught up with Solaris sparc. FreeBSD will be even greater!
Linux is the only original operating system in many, many years which has managed to solve the chicken-egg problem with regards to hardware and software. If MacOSX had not had binary compatiblity with MacOS9 -the new OSX would have been a horrible flop. BEOS was probably the last propietary OS to really have a chance in the commidity PC world.
People have such short memopries-when Microsoft started there were not that many hardware devices for PC’s available-a tiny, tiny fraction compared to today. Microsoft became the standard OS because IBM chose to put it on their PC’s- and IBM PC’s became popular because IBM was THE name in computers, at least as far as the corporate world was concerned, in the early 1980’s.
That IBM embraced the PC opened the door to millions of small buisness to purchase their own computers- Compaq followed heals offering better technology cheaper-thus the PC market began. Microsoft was already standard. New technologies opened up the door for new devices and new companies sprung up to meet new demands.
There was only one OS for PC’s-and the sucess of the PC at the workplace carried over to home-this combination enabled MS to become the monopoly it is today. The combination was sufficiently large enough to support a burdgeoning market of new hardware and software companies. Those companies had no choice but to write their stuff for Microsoft PC’s.
It has taken thousands and thousands of developers around the globe over a time period of almost 15 years for something to coalesce sufficiently to be able to be a viable alternative to Microsoft. Linux has done this because it runs on everything-PC’s, Mac’s, PDA’s, Sparc stations and IBM Z series servers-to name just a few. Linux makes it it possible to create new kinds of computers today based on different architectures-99% of Linux software runs on every platform on which the kernel compiles.
No one today even speaks of PC’s anymore-anything which does not provide PCI/AGP is not even remotely viable in the commodity computer market today. This is because the the specs to the bus architecture of PC’s has long since been standardized-any manufacturer can implement it. Microsoft got carried on the back of IBM’s decision to create an open platform for personal computers-the only thing in the original PC that was closed was the BIOS-and Compaq together with Phoenix figured out how to open that up too. When people lament the Microsoft monopoly what they really mean is the IBM, Microsoft, Intel alliance which gave birth to the modern PC market. IBM became irrelevant-but it’s standards stuck around-still forming the basis of our motherboards-though countless revision and improvements. Intel, now accompanied by AMD and Cyrix, dominate the processor market absolutley-PPC is what 5% of the CPU market.
I miss the days where there dozens of different computer architectures(z80,8085, 8088,6809,6502,68000, TI 94XXX/commodore/Tandy/Apple/Atari/Timex Sinclair/TI). Funnily enough Linux makes it possible for such to emerge again…
I want choice and I just don’t want one monopoly (Windows) replaced by a different monopoly (Linux).
You think Linux is a monopoly?
It’s not even freaking funny when Linux 2.6 was touted as being stable and the next biggest thing to sliced breaks all kinds of apps. Right now my Nvidia or ATI drivers don’t work…
nVidia drivers work absolutely fine with the 2.6 kernel. The only thing you watch out for is specific drivers you need to have for 2.6, and the addition of uDev. Other than these slight bumps it is 100% stable and it does not break any apps – certainly not in user space. Sorry!
…because some bozo decided to add REGPARM and 4KSTACKS.
Rrrriigghhttt…
I suggest you go and buy a commercial Linux distro. Free distributions are fine if you can work methodically to get them up and running. They aren’t for whingers.
You Linux appologists better gear up for a good ass kicking by MacOSX, Longhorn, FreeBSD and Solaris in 3 years time.
And what do you think’s going to happen in 3 years time? The second coming?
“And what do you think’s going to happen in 3 years time? The second coming?”
Didnt you know?
LOL
Do yourself a favor and don’t make a fool of yourself by claiming that if Nvidia opened up their specs that the open source “community” would make better drivers than Nvidia themselves.
ATI had open specs for some time and the open source drivers were piss-poor compared to ATI’s binary drivers.
It’s absolute idiocy to think that a few kids with some specs is going to produce better drivers than Nvidia software and hardware engineers working together.
Don’t confuse an advanced graphics chipset with some generic ethernet card.
I don’t think ATI has fully released the specs to for their Radeon line of cards to the open source community. From searching the web, it seems that they’ve only done it with the Rage line of cards.
The Radeon still uses proprietary binary drivers because ATI has some NDA’s with certain parties. More info at http://www.ati.com/products/catalyst/linux.html
A check at http://dri.sourceforge.net/cgi-bin/moin.cgi/ATI confirms that ATI only released the specs up to the Radeon 9200. Still no excuse for why the open source drivers don’t hold a candle to ATI’s proprietary drivers though.
“Yep Linux is going down – all the companies that are cheering for Linux are going to get over taken by open source clones and then they’re going to come back to their senses and you’re going to see the likes of Oracle, Veritas and others going back to support Mac/Solaris/Longhorn/BSD just to give open source guys a taste of their medicine. ”
bsd is open source. mac os x has a bsd base. ibm is selling hardware and services which is not clonable. you got nothing to backup your claims anyway
“You Linux guys admit that UNIX is your enemy #1 (Linux effectively killed all UNIX except Solaris) and but you cannot afford to kill MS because you need MS for all their wonderful research otherwise you’ve got nothing to copy (.net/outlook/CIFS == mono/evolution/samba/) ”
when was this research
you dont think reverse engineering cifs is easy do you?
evolution has a lot more than outlook.
http://www.novell.com/products/evolution/
Mono has two stakcs
“http://primates.ximian.com/~miguel/tmp/two-stacks.png“
what about innovations like selinux and stuff
“Leave the work of innovation in software to professionals (go take some CS classes kiddo!) ”
i guess ibm, novell, redhat, ca , sun and others dont count as software professionals. good
Yes, I should’ve said “ATI had open specs at one time” instead of “for some time”. You are right, the radeon series is closed.
Still no excuse for why the open source drivers don’t hold a candle to ATI’s proprietary drivers though.
But yeah, it’s ridiculous to think that some sourceforge project or even the X guys are going to be able to produce better drivers than software engineers working together with hardware engineers 40 hours a week on this stuff.
I’m not denegrating the work that the open source guys do, but it’s just the fanboyism that open source automagically makes code better that irks me.
I’m just happy that closed-source drivers are allowed into the kernel or linux would still be relegated to the server and to the hardcore hobbyist who will tailor his machine to linux.
I am not making a fool of myself in stating that FOSS developers could produce better drivers than Nvidia. You may think that FOSS developers are simply a bunch of “kids”. You are mistaken. Many of the best programmers alive today work in FOSS. Some of these brilliant and talented developers code freelance for FOSS and some are on being paid to develop FOSS-as you probbly already have noted -many corporations have developers working on FOSS-IBM, HP, Novell, Redhat, SGI and many, many more.
Now I am not saying that the FOSS community could take freshly released Nvidia documentation(or code) and turn out a better driver in 6 weeks. Of courser not. It takes time for anyone-or any group, regardless of how good they are to tackle something of this size. But given time(ie. 6 months to a year) they could in fact do this. If people like Keith Packard, Owen Taylor, Robert Love, Andrew Morton, Carsten Heils(Rasterman) had access to such stuff-you would be amazed at what could happen.
Much of what has hampered the development of the next generation of desktop development on Linux has been based on the dependency of Linux on binary-only drivers for hardware accelerated 3D support on the most popular graphic cards. Sure developers can create new things,Packards kdrive(a.k.a Xserver) for example, but if Nvidia AND ATI don’t rewrite their drivers to support it(composite, damage,fixes extensions for X11)-nothing will ever come of it.
If the community had access to the documentation a whole lot more could be done with these incredible cards than is available today with the limited support in the current propietary drivers. Keith Packard and Owen Taylor were developing state-of-the-art graphics technology before Nvidia or ATI was ever even dreamed of. I am sure that Sun, IBM, HP, SGI and other corporations would love to be able to fully tap into the power of these cards for enabling incredibly powerful applications for film editing, CAD/CAM, etc.
You should not make a fool of yourself thinking of FOSS developers as “kids”.
Of course you neglect one other little thing:
If Nvidia AND ATI were to pursue open documentation for their GPU’s- FOSS developers and engineers at Nvidia and ATI would be free to work together, in the spirit of FOSS, to unleash the capacities which these technologies have. I guarantee you that the engineers at Nvidia and ATI would love to have the chance to have their hands in projects which are much more far-reaching than the projected “consumer-demand” which the higher-ups in these companies impose on the engineers. The point is if these companies were open in regards to releasing documentation synergenic energies would coalesce around these technologies creating new communities of incredibly talented people.
I think you totally underestimate what an impact such could have. Not to mention that if such were to occur-the Linux community would provide such drivers for all platforms on which Linux runs and all other developers of other alternative operating systems would be able to benefit from this- SkyOS, OpenBeos, MorphOS, AmigaOS, Linux on PPC, Solaris, YellowTab, GNUstep, Syllable, and on and on and on…..
“Gee, if they do not like Linux, then do an Apple move and use one of the free BSD’s.”
Why, you can download Solaris x86 for free.
Listen, even if Nvidia wanted to they can’t open up their drivers because of NDAs with intellectual property holders. ATI is in the same boat too now with the Radeon series.
As mentioned before, when the rage series specs were open, the open source drivers were still subpar to anything that ATI put out. There’s some talented engineers in the open source graphics world, but that’s far from being able hack fast drivers that are known inside and out by Nvidia, ATI hardware and software engineers. Even if open source programmers could be up to speed in 6 mos. or a year, which is optimistic for quality drivers, what happens when a new GPU comes out. They have to get up to speed once again and the open driver is always going to be of inferior quality to the ones created by the guys that know the hardware inside and out.
That is not to say there shouldn’t be cooperation between the X guys and the driver guys, but let the X guys do their thing and let the driver guys do theirs.
Certain people in the “community” need to grow up, accept that there is not always going to be source code for everything they run, and be happy that they even have high-quality linux drivers for today’s kickass video cards.
“Why, you can download Solaris x86 for free.”
because solaris on x86 sucks.
“I am not making a fool of myself in stating that FOSS developers could produce better drivers than Nvidia. You may think that FOSS developers are simply a bunch of “kids”. You are mistaken. Many of the best programmers alive today work in FOSS. Some of these brilliant and talented developers code freelance for FOSS and some are on being paid to develop FOSS-as you probbly already have noted -many corporations have developers working on FOSS-IBM, HP, Novell, Redhat, SGI and many, many more. ”
Actually the makeup of FOSS developers has been covered years ago. This hasn’t been the first time the “they’re not professionals” has been raised and shot down. In fact I bet that some of them are the self-same closed-source developers that apparently people admire when they say that one group can do this and another can’t. Skills are skills, and knowledge is knowledge, and it only favours the prepared. Be-fan goes on about how this group can do this, and another group couldn’t do that, even when it was pointed out that one of the X developers is also a Nvidia employee.
Also if the idea that only the closeness of the hardware and software developers can guarentee good drivers? Then how does one explain all the problems that Windows has had that were caused by drivers? Or the rep. that ATI drivers have that they still have trouble shaking?
Two things:
“Listen, even if Nvidia wanted to they can’t open up their drivers because of NDAs with intellectual property holders. ATI is in the same boat too now with the Radeon series. ”
Note we only have their word on this, two they could release register-level specs and we could write our own. We did this with the Mozilla code. What’s the excuse again?
Solaris on x86 (whether Intel or AMD processors) is the *only* truly affordable UNIX for the small or home-based business. Why? It has the lowest evaluation cost (none), it has solid networking support (with most common Ethernet connections, its as close to Plug and Play as UNIX gets) and, like Linux, it has a truckload of actually usable freeware (take a gander at http://www.sunfreeware.com; it’s THE source for Solaris freeware, and most of it is directly targeted at Solaris for x86), in fact most of the apps were ported from Linux.
No, Solaris isn’t Linux. Solaris isn’t (and wasn’t) MEANT to be Linux. Solaris is UNIX. The real deal. However, on common x86 hardware, Solaris is no harder to install than RedHat or SuSE (it may, depending on the hardware, be even easier). If you have GNOME (rather than CDE) as your default environment, you would have serious trouble telling Solaris from Linux as a user. However, if you go online with Solaris, Sun’s years and years developing Solaris makes for a comforting safety net. Add in that Solaris for SPARC and Solaris for x86 are absolutely feature-identical (not exactly something you would expect from a hardware company such as Sun) and that the software cost is identical (zilch for home use and SoHo usage), Solaris makes for a good alternative to Linux.
And that is precisely Sun’s point.
Sun *already* had Solaris for x86. Why in the world should it *reinvent the wheel* (as it tried to do with Sun Linux) when it had a good, solid, *proven* OS sitting on the shelf?
The one thing that Solaris for x86 suffered from is *apathy*…and that was entirely Sun’s fault. (Fortunately, *the revolt* woke up Sun’s software division to the lack of need to reinvent the wheel.)
Sun is now cranking the x86 Hardware Certification Program into overdrive; the list of *new* hardware supported by Solaris/x86 grew larger during 2004 alone than in any other single year, and the certification process is truly encompassing, as the BYOSB (Build Your Own Solaris Box) market is actually being seriously addressed, and not just by Sun itself. Motherboard makers are submitting x86 motherboards for certification in record numbers (ASUS subitted their entire line of motherboards, both Intel and AMD-powered; Intel submitted all their desktop, workstation, and Server boards, and other manufacturers have followed suit).
With the JDS (and GNOME), Solaris is actually a decent alternative to Linux, and not just in the server closet.
what about innovations like selinux and stuff
It seems as though a couple BSDs already have similar frameworks, better in fact, as the various MAC policies are stackable.
http://www.trustedbsd.org/
http://www.trustedbsd.org/sedarwin.html
http://www.trustedbsd.org/sebsd.html
And let’s not forget about Trusted Solaris:
http://wwws.sun.com/software/solaris/trustedsolaris/
And SELinux in fact is essentially a port of the security framework of a completely different OS, meaning that using it as an example of OSS innovation is quite silly.
http://www.nsa.gov/selinux/info/
Why would you use CDE? look on blastwave.org, there is GNOME 2.6, KDE 3.2.1 etc. etc. Couple that with the XFree86 4.4 porting kit from tools.de/solaris and voila, you have a pretty stable system
“And SELinux in fact is essentially a port of the security framework of a completely different OS, meaning that using it as an example of OSS innovation is quite silly. ”
it has been heavily modified and needs integration with the whole operating system to be effective. having a research trial doesnt count.
“t seems as though a couple BSDs already have similar frameworks, better in fact, as the various MAC policies are stackable.
http://www.trustedbsd.org/
http://www.trustedbsd.org/sedarwin.html
http://www.trustedbsd.org/sebsd.html
”
these are open source too and has been similar attempts to port from linux. so linux doesnt have any inovations is clearly false
there are whole lot of other stuff
Interesting that you talk about SUN server sales, just recently their server shipments grew by 26%, although their revenue declined; why? they’re retaining their customers by offering Solaris loaded on x86 boxes. Customers can now move to cheaper hardware and run all their favourite servers stuff on Solaris.
When Solaris 10 is x86-64 native, with Sybase and Oracle running on it, there will be little reason to be running Linux.
Linux’s claim to fame was because it is a cheaper UNIX, now Solaris is dirt cheap, infact, cheaper than Red Hat Enterprise, people are now saying, “hell, why not by an x86 machine from SUN loaded with Solaris?”. People are making that choice, and many customers are happy with the direction from SUN.
What I hope is that they finally market properly, Solaris x86 for the low end ( <8 CPUs) and Solaris SPARC for the high end ( >8 CPUs), with Java as being the unified API that allows the two to work seemlessly together.
these are open source too and has been similar attempts to port from linux
I never claimed them to be innovative, and I am well aware of the fact that SEBSD is a port of SELinux. I still have to point out the fact that none of them are good examples of Linux’s or BSD’s or OSS’s ability to innovate.
Chew on that.
“still have to point out the fact that none of them are good examples of Linux’s or BSD’s or OSS’s ability to innovate.
Chew on that. ”
its innovation despite what kingston says. Kingston is well known for his bad mouthing and trolling so others can ignore him
its innovation despite what kingston says. Kingston is well known for his bad mouthing and trolling so others can ignore him
You should have signed “Linux troll”
if i only knew i could run all my everyday apps, i’m sure they do run but the problem is getting them there.
is netbsd pkgsrc a good solution for solaris or how do you actually get all your everyday apps there?
currently i use *bsd.
> No, Solaris isn’t Linux. Solaris isn’t (and wasn’t) MEANT to be Linux. Solaris is UNIX. The real deal.
That’s such crap. UNIX is just a trademark. The only thing it means is that you paid the Open Group some cash and are basically POSIX compliant.
The places where Linux differs from the POSIX standard are either deliberate or not very important.
There is one company called SGI that have already tried that, check how far their share price ended up…
SGI were already in trouble and they needed to do something, fast. Sun are in that kind of trouble, but they still have a reasonable installed base to keep their heads just above water. However, it isn’t going to last.
Sun has got by far the biggest 64-bit installed base in the world on Solaris, do you think they are going to just drop the ball and walk away from that?
Because they’re in serious financial trouble, that’s why. They have no prospect of even revenue in the here and now, let alone future growth, and their R & D is producing absolutely nothing.
I don’t think many people here seem to realize the terrible financial situation Sun are in. They need to do something, and mucking about with Solaris on x86 and doing deals with a company who wouldn’t mind taking a lot of their revenue isn’t going to solve that.
Some people are still thinking of Sun as this great, huge company with steady revenues. They aren’t.
I’m not denegrating the work that the open source guys do, but it’s just the fanboyism that open source automagically makes code better that irks me.
You’re correct, there are some things that commercial development does better. However, once a commercial company ships a product and they’ve produced drivers they just don’t support the thing anymore. Open source drivers take longer to get right, but they continue to be developed and perfected over time.
The prism 54 wireless drivers are a good example, because you get better performnce out of those than the Windows drivers. With a few specs and iterative improvement it shows that the open source method works very well. The prism 54 project will still be around, but can you say the same of the Windows wireless drivers for the card that you bought that was obsoleted two weeks ago?
I’m just happy that closed-source drivers are allowed into the kernel or linux would still be relegated to the server and to the hardcore hobbyist who will tailor his machine to linux.
The vast majority of the drivers are open source, but it is good to have some closed source drivers where they’re needed.
Linux’s claim to fame was because it is a cheaper UNIX, now Solaris is dirt cheap, infact,
You’re correct in many ways. However, everyone is missing a crucial element in the way Linux is developed. Linux is developed collaboratively, so many companies (some large like IBM and HP) can share the burden of work in getting new drivers written and push forwards what can be done. They also indirectly benefit from new ideas being implemented at a grass-roots level.
Sun are on their own in this regard with Solaris, and cannot possibly hope to generate that kind of movement.
cheaper than Red Hat Enterprise, people are now saying, “hell, why not by an x86 machine from SUN loaded with Solaris?”. People are making that choice, and many customers are happy with the direction from SUN.
The problem is that you can look at this the other way around. You get more hardware vendors, and a much greater choice of hardware and options with Linux and a much greater deal of compatibility. With all the will in the world, Sun cannot provide that on their own. Only Microsoft have the resources to do that, and even they’re on their own.
What the hell are you babbling about? Solaris is at least a mile ahead of Linux in pretty much any department (scalability, features, ISV support, you name it…). With Solaris 10 it will 2 miles ahead of Linux.
OK then, and I’ll put this very simply. If Solaris is a mile ahead of Linux distributions in terms of features (which it isn’t – not on those that matter), then why is Sun in the red? People can’t be finding those features very compelling. That’s the cast-iron truth right there.
You can’t tout new features and tell people how great Solaris is if it isn’t producing a return. That’s the sort of babble people don’t seem to understand.
its innovation despite what kingston says. Kingston is well known for his bad mouthing and trolling so others can ignore him
You sir or madam are the troll, and not a very bright one at that. Citing SELinux etc, as evidence of Linux’s ability to innovate is about as good an example as proclaiming that it has a virtual memory system, or the fact that it uses shared libraries, or protected memory.
You seriously need to get a clue.
“You sir or madam are the troll, and not a very bright one at that. Citing SELinux etc, as evidence of Linux’s ability to innovate is about as good an example as proclaiming that it has a virtual memory system, or the fact that it uses shared libraries, or protected memory.
You seriously need to get a clue.”
no its not. MAC has never been in any OS as common as a vm. its not a essential part either. maybe you need to follow what you preach.
i think sun should split its software business and hardware business. Their software business would attract investments from oracle and IBM.. Software that runs on many platforms… Their hardware business would tout performance and usability. Promote SPARC for many other operating systems and applicaitons.. not solaris-only.
For sun dying, well thats just a stupid thing to say if you do not know the company and people have been saying that for years and now it finally seems like they will succeed. I bet after they report a profitable quarter people will be like “you were lucky sun, but you wont be for long, bye bye sun”
What sun wants to do on x86? they want to enter the OS market and sell software on it to increaes revenue and they will. IBM, HP, SGI, etc. all have operating systems tied to a particular platform therefore they don’t see the need to develope and market the OS as they care about hardware.. Sun is different. Solaris runs on many platforms now and Mr. Schwartz’s roots come in software and he is going to compete with Linux head on in the software market while embracing it in the hardware market to give users “choice”. When sun released Solaris x86 it was obvious that Solaris is an investment and its going to stay around.. Anyone who thinks they will get rid of it and embrace linux and leave their solaris customers high and dry must be really confused or something. Ohwell, this is a Linux-pro site anyway it seems
“Anyone who thinks they will get rid of it and embrace linux and leave their solaris customers high and dry must be really confused or something. Ohwell, this is a Linux-pro site anyway it seems
”
you will have to tell that to Sun. they have been changing their linux and solaris on x86 strategy pretty frequently now.
MAC has never been in any OS as common as a vm.
Certainly not in your favorite toy. MAC has been around in commercial UNIXes for quite a while.
Can someone tell me why everybody’s been buzzing about Linux and sounding the death knell for Solaris?
Is it because:
– Linux is free
o Solaris is free for up to 1-2 processors
o Is Linux really free if you’re getting Redhat or Novell?
– Linux has all the drivers
o that’s a problem with Solaris but they are solving it
o IHVs and 3rd party driver developers will support Sol.
– Linux has a lot of apps
o All the open source apps can be ported to Solaris
o Proprietary apps – VMWare/Win4Lin need to be ported
o Macromedia/Adobe/Oracle/Veritas are already there.
o Alias/Maya/Multimedia/Games need to be ported
– Linux has big names like IBM/Novell/Intel backing it
o So far Sun’s got Microsoft and AMD
o Sun needs to get Apple/Adobe/Macromedia to show support
– There is source code for linux kernel
o do you think Linux developers are going to jump to developing Solaris?
o Will open sourcing Solaris benefit Sun?.
o Apple’s come back from the dead without OSS’ing
Does this mean that there can never be a new operating system that will cause Linux developers to drop linux and migrate to it? How can one effectively compete against Linux?. Is the OS market dead for good?
“MAC has never been in any OS as common as a vm.
Certainly not in your favorite toy. MAC has been around in commercial UNIXes for quite a while.
”
now you started trolling again. whats the toy i am using?
MAC is not a common component of mainstream operating systems in anyway. its not the same as the vm of the operating system and its pretty much optional and not suitable for all scenarios.
trusted solaris is way too costly to be affordable piece of software for anything less cluster of sparc machines.
i have been using that for a couple of months on a edge server and it aint fun
comparing a vm and MAC is pretty stupid thing to do
“Can someone tell me why everybody’s been buzzing about Linux and sounding the death knell for Solaris?
”
solaris is alive and kicking on the sparc however with this non consistent strategy on the x86 which is a commodity platform the integrated hardware and software approach of Sun doesnt really fly well.
”
Does this mean that there can never be a new operating system that will cause Linux developers to drop linux and migrate to it? How can one effectively compete against Linux?. Is the OS market dead for good?”
not at all. Windows is not any going anywhere anytime soon. there are all those bsds and new hobby operating systems coming out everyday.
with an extensive amount of open source apps which can be ported without any lockins I would say that the number of operating systems that is available and used in the market is bound to be more diversified than ever.
They can even benefit each other. For example the paid and commercial development teams provided for gnome and kde also benefit freebsd. the recent enhancements to freebsd current took some ideas from linux 2.6 vm. rik van riel from redhat is known to have a good understanding of the freebsd vm and has contributed to linux from that. selinux has been experimentally ported to bsd.
skyos recently ported gtk and a couple of gnome apps to their operating system. khtml is being used by safari in mac and skyos too. mac has a bsd base now and so on.
basically there is the posix family of very diverse operating systems and windows competing now.
whats the toy i am using?
You’re right. I have no idea what you’re using, and I should have chosen my words more intelligently. I should have said “advocating.”
i have been using that for a couple of months on a edge server and it aint fun
Right. Gotcha. SELinux and TrustedBSD are far more “fun.”
comparing a vm and MAC is pretty stupid thing to do
Nope. MAC has been done far too often before to be considdered overly innovative nowadays.
In other words, lowest common denominator OSs (Windows in the larger world, and Linux in the open source world) playing catch up isn’t innovative in a grander sense.
“I should have said “advocating.”
ok. whats the toy that i am advocating. scared to mention it by name or what?
“i have been using that for a couple of months on a edge server and it aint fun
Right. Gotcha. SELinux and TrustedBSD are far more “fun.”
figure of speech. not literal meaning
”
comparing a vm and MAC is pretty stupid thing to do
Nope. MAC has been done far too often before to be considdered overly innovative nowadays.”
who is talking about it being overly innovative. it hasnt been done far too often and it isnt a common thing like the vm so your argument is pathetic
“In other words, lowest common denominator OSs (Windows in the larger world, and Linux in the open source world) playing catch up isn’t innovative in a grander sense.
”
not in general. they are both playing catch up in different ways. same goes to any other operating system.
they are both playing catch up in different ways. same goes to any other operating system.
Wow. We agree on something (two things in fact)! I’m shocked!
:Wow. We agree on something (two things in fact)! I’m shocked!”
i am not. the last statement was pretty reasonable unlike you other stuff. if you quit calling other operating systems toys and carry any dicussion in a intelligent way without outright name calling we are bound to agree on things too
i am using windows, solaris, freebsd and linux. none of them are “toys”
whats the toy that i am advocating?
You’re advocating the idea that the SELinux and TrustedBSD projects were innovative (see quote below).
what about innovations like selinux and stuff
“You’re advocating the idea that the SELinux and TrustedBSD projects were innovative”
that doesnt parse. an idea can never be a toy. you will have to use something to call it a toy. i wasnt advocating selinux at all and you were the one bringing up trustedbsd, not me. you better go read what i said
and carry any dicussion in a intelligent way without outright name calling
Yeah, something I need to work on I guess.
you will have to use something to call it a toy
You were using it as an argument ;^)
There are a lot of differences between selinux and traditional MAC frameworks
if you have used something trusted solaris and selinux it would very evident.
for example there is no difference between a domain and a type in selinux. you can generate runtime policy definitions. there are a good number of demo machines with a public access including root username and password for demonstration purposes
fedora integrated common utilites like su to change types when switching user identities.
http://people.redhat.com/kwade/fedora-docs/selinux-faq-en/
To access my Fedora play machine ssh to cable.coker.com.au port 222 as root, the password is “fedora”.
check this out before dismissing it as something non innovative and we can discuss about this
I had read about all of that. You’re not quite so clueless as it appeared from quite a few of your prior posts. My appologies!
> not at all. Windows is not any going anywhere anytime soon.
> there are all those bsds and new hobby operating systems
> coming out everyday.
Linux came in through the back doors. It could have been BSD but it wasn’t. How does Sun get those developers/sys-admins/office-geeks to bring a Solaris x86 box through the back doors?
”
Linux came in through the back doors. It could have been BSD but it wasn’t. How does Sun get those developers/sys-admins/office-geeks to bring a Solaris x86 box through the back doors?
”
not in areas like webhosting. linux came there with a bang. solaris cannot be pushed the way linux does it. it will have to follow a different route. i dont think the current way is the right thing to do.
I think it would be very easy for sys-admins to bring Solaris in. Since it is free and if you have hardware it will run on, Solaris x86 could be used for things like DNS, FTP servers, DHCP servers, etc…
“. Since it is free and if you have hardware it will run on, ”
thats the only problem right now
I only use pretty standard hardware and I don’t have any problems with it. I have it on an old Compaq DP4000 PPro 200 with 128 MB of RAM and a Cirrus video card and I have it on a white box PIII 600 with 256 MB of RAM and a Matrox video card. I had no problems getting it to run on either box and it runs just as well on that hardware as RedHat 9 and Fedora Core 1.
Personally, I think Sun simply needs to disappear off the radar for 2 or 3 years, while they cook up something really good.
Their stable is full of mediocre, or ‘good enough’ products.
Java is ‘Good Enough’ but has poor graphical performance, a lack of unified media APIs across platforms, slow application startup times and a lack of desktop integration.
Solaris x86 is ‘Good Enough’ but has major problems with hardware support, and has been forced to follow Linux’s lead as far as UI goes.
SPARC hardware is ‘Good Enough’ but is slow and expensive compared to x86 and PPC, POWER or Itanium.
Sun Ray etc. is ‘Good Enough’ but is expensive and a Sun-only solution, which has seen a terminal lack of focus from Sun in recent times
Their ‘Server Software Stack’ changes names/branding so often it is difficult to figure out what it actually provides. I don’t even know if this is ‘Good Enough’ or not.
Sun just seems unable to make products that can be classed as ‘excellent’.
The only exception to this might be 32+ CPU SPARC machines running Solaris/SPARC in a business-application-server role.
Sun should make their choice – either disappear for a couple of years, and put something truly excellent together on the low end – just like Apple, or dump that market and focus on the high end, producing an excellent product for a smaller, but higher-margin market.
For an example of the contrast between the 2 approaches, look at NeXT and SGI.
Both of these companies were heavily UNIX based, both doing hardware and software, and both were doing pretty well, until x86 came along, and basically destroyed their business models.
Both of these companies attempted to ’embrace’ x86, and ended up being beaten out by vendors of crappier, cheaper stuff.
However, NeXT’s persistence and dedication to ‘excellence’ eventually allowed them to be recast as a core part of ‘the new Apple’ – and the quality of MacOS X really makes the excellence that was demostrated by NeXT from day 1 shine through.
SGI thought it had a real market selling x86 machines with custom SGI graphics hardware, only to find that NVidia’s whirlwind R&D/product release schedule doomed their low end solutions to mediocrity.
SGI have been essentially forced out of the workstation market, which they used to be a significant player in, and have chosen to focus on the high end. They decided to take this choice, instead of letting indecision and lack of focus kill them slowly, and their Altix supercomputers are the among the most powerful computers you can buy for scientific/technical computing, and they are seeing a return to profitability.
They were able to realise that IRIX and MIPS had fallen from ‘excellent products’ to merely ‘Good Enough’ and allow them to fall by the wayside instead of letting arrogance and pride lead them . Perhaps they realised this too late, but they realised it and acted on it.
Neither NeXT or SGI was able to survive without help – NeXT from Apple (they got none from Sun), SGI from Intel and the Linux community (They got none from Microsoft). – Both now have world-class, excellent products, and neither of them are constantly bleating about TCO, ROI, Indemnification, Proprietary Linux, FUD, FUD, FUD like Sun are.
They have excellent products, and those tend to speak for themselves.
Sun should either shut up for 2 or 3 years and produce a ‘Longhorn killer’ – that is, a desktop OS that is literally 5 years ahead of anything Microsoft and Apple have to offer, with deep and broad support for Java at every level from filesystems and hardware support to scripting and UI frameworks, and put your faith in the fact that your software R&D team is better than Apple’s, Microsoft’s and the Free/OSS Communities’. This will either sell you boxes directly, or allow you to tax others box sales, as microsoft do. If you make a product that stands head and shoulders above the competition, you will get sales.
or
Forget about anything with less than 8 CPUs, focus on making ‘big iron’ and Solaris/SPARC, and put your faith in the fact that your hardware R&D team is better than IBM’s, AMD’s, Intel’s and HP’s. If you can make a CPU and platform that outperforms x86/PPC/POWER/Itanium, you will get sales.
If they don’t have faith in either of these, well, they better start shifting to either a software services model (IBM Services) , or a commodity sales model (Dell).
The only space left is volume hardware for the masses ‘with a twist’ e.g. Sony/Nintendo consoles, Apple iPod, TiVO, Palm, Nokia etc.
Sun has forgotten themselves – ‘The network is the computer’ should never have been replaced with ‘the dot in dot com’ – I don’t see any point in them carrying on this way now that the bubble has burst.
They don’t seem interested in moving computing forward any more, wanting to be seen as ‘todays big player in the server space’ not the ‘company that started the revolution’.
“If you can make a CPU and platform that outperforms x86/PPC/POWER/Itanium, you will get sales.”
There is more to performance than just the speed of your CPU.
“Sun has forgotten themselves – ‘The network is the computer’ should never have been replaced with ‘the dot in dot com’ – I don’t see any point in them carrying on this way now that the bubble has burst.”
Managers approve purchases, managers know about the burst of the dot com bubble, managers often base decisions on marketing material, thus it is important to drop anything referring to dot com.
“They don’t seem interested in moving computing forward any more, wanting to be seen as ‘todays big player in the server space’ not the ‘company that started the revolution’.”
Java, 3D computing, chip multithreading, etc…
> I think Sun simply needs to disappear off the radar for 2 or 3 years, while they cook up something really good.
If somebody has to disappear off the radar, it should be arrogant FUD-happy Linux Nazi that try to spoil the reputation of otherwise good company. Your opinions are totally unfounded and extremely biased. From your post all I can say is that you know absolutely nothing about Sun and its product, but you nevertheless are brave enough to post the regurgitated crap you’ve picked somewhere else.
At this point in time Sun has got one of the best if not the best product stack out there.
> Solaris x86 is ‘Good Enough’ but has major problems with hardware support.
Hardware support is not even the most important issue in the enterprise, since the enterprise environment is pretty homogeneous by the most part at least as far as new deployments are concerned. And by the way Solaris x86 is already ‘Pretty Good’ and with Solaris 10 it will ‘Excellent’. I will pick Solaris over Linux any time of the day as it is, with Solaris 10 it will not even be a question. Solaris is more feature rich, more stable, more proven, will have better ISV support, and is ultimately cheaper to purchase and support.
> SPARC hardware is ‘Good Enough’ but is slow and expensive compared to x86 and PPC, POWER or Itanium.
Sparc hardware is ‘Excellent’ as it is and by no means is more expensive compared to x86 and PPC, POWER or Itanium. In the 4+ processor space Sparc machines are actually much cheaper and having better performance per buck than equally spec’ed boxes powered by Pentium, Power, and definitely Itanic which is going the way of dodo anywhay. Sparc is still the way to go for database and ERP applications.
> Sun Ray etc. is ‘Good Enough’ but is expensive and a Sun-only solution, which has seen a terminal lack of focus from Sun in recent times
I guess this is why Sun is addining more and more customers to its Sun Ray list. Check with US Army, US AirForce, FAA, Q-West and shit load of more BIG companies.
> Their ‘Server Software Stack’ changes names/branding so often it is difficult to figure out what it actually provides. I don’t even know if this is ‘Good Enough’ or not.
Java Enterprise stack is ‘Excellent’ compared to similar offering from Sun’s competitors and is almost an order of the magnitude cheaper at some points. Let’s see, Sun has got the:
-best Directory Server out there (inherited from Netscape)
-one of the most scalable and best performing application servers (almost linear scalability up to 100 CPU’s)
-most feature-rich portal server out there (definitely better than what IBM, BEA or Oracle has got) supported by an extremely good Identity Server
-best performing and definitely most secure Web Server out there.
Oh, and speaking of SGI, if it actually kept on developing MIPS and IRIX, it could have been in a much better position than it is right now. Itanic is not going to take them far, since even Intel and HP are about to ditch it. SGI was doing fine and was a brilliant company before Rick Beluzzo (the Microshaft sellout) and his retarded ‘Intel’ strategy drove the company into the ground by freaking out the clients by trying to adopt Windows and Linux on Intel.
” solaris on x86 mmmmmmmmm stinky stinky”
And what makes it Stinky? Solaris x86 is a very nice OS and With GNOME its as usable as Linux. People have said that Sun cannot outprice Linux, well Sun does that as well, you can get it for free from the Solaris Binary program with no support, well you get no support when you download a Linux ISO, you do get support from the Media kits. If you are an Enterprise user you pay for the support and you have a nice alternative. As with Microsoft Windows, most of your Open Source software applications will work.
” hardware support is almost non-existant at the moment and I cannot see that improving much ”
Hardware support has gotten better since Solaris 8 and since Sun has decided to go this route I see hardware support getting even better, as with Linux, you just have to try it and see, My Solaris x86 box without virtualization is an older 566 and the only thing on Suns HCL is the Video card, an ATI vid card and Solaris x86 runs on it.
” but the worst thing is that solaris x86 is so slowwwwwww
and I mean really slow
in fact it is slower than a bagfull of slow things ”
I dont experience any lack in speed, maybe thats what I get for actually TRYING the product and not playing follow the leader and flaming Sun for not Open Sourcing Java. As a consultant and a developer I would reccomend Solaris x86 to anyone who wants an alternative to Linux and who are migrating to x86 hardwate.
“There is more to performance than just the speed of your CPU.”
Thats why i said ‘outperforms’ not ‘faster’ and added the ‘platform’ qualifier.
“Managers approve purchases, managers know about the burst of the dot com bubble, managers often base decisions on marketing material, thus it is important to drop anything referring to dot com.”
No, I just think ‘the dot in dot com’ doesn’t mean anything. Not only does it expressly trivialise any non-american, non-commercial use of Sun technology, but it makes no sense to anyone who is technically savvy. The ‘dot com bubble bursting’ puts a bad light on it, but it is stupid regardless.
‘The network is the computer’ exposes a fundamental property of the computing movement that sun was a central part of (the internet, client/server and distributed computing) in a succinct and clever way. It expresses an aspect of Sun’s design and engineering goals in a single, simple, precise statement.
I just think ‘the dot in dot com’ is a clever use of the english language if you think of the internet only as the place where you buy your stuff, while ‘the network is the computer’ is clever no matter where you are from or what you do.
“Java, 3D computing, chip multithreading, etc… ”
Exactly – Java is merely adequate – its a nice language, in the way that Python is a nice language – as a computing platform it has people waiting for something better to come along.
They’ve hardly set the world alight with Looking Glass, in fact preferring to show it as little as possible becuase they know themselves its a frickin waste of time allowing people to ‘write on the back of their windows’
Chip multithreading is being done today by Intel, IBM, and even you can even get multicore ARM chips – A chip composed of a bunch of slow SPARC cores is not going to be any better than a chip composed of fast POWER cores.
Instead of replacing CDE with GNOME and making GNOME the best desktop available anywhere, they sit on their hands.
Instead of making Java the best App framework available and taking it to Microsoft’s .NET they sit on their hands.
Instead of investing heavily in SPARC and ‘betting the company’ on it, they hedge their bets with AMD and pretty much – yes – sit on their hands.
The reason these Sun-bashing threads attract so much attention is because we all used to look up to Sun and think their hardware was – to use modern parlance ‘the bomb’
Now we’re seeing one of the companies who were our ‘heroes’ fade away and die because of sad bastards like Schwartz who think that ‘dot com’ is more important than ‘network computing’
I have a point and a question.
My point is, right now Project Looking Glass seems like a big waste of time to some people, but what if it makes people think of computing in a different way? What if it sparks a new way to solve computing problems just by having the idea out there and looking at it?
My question is, if you think everything Sun is doing is merely adequate, please list some companies/people that are putting things out that you think are revolutionary?
“Linux’s claim to fame was because it is a cheaper UNIX, now Solaris is dirt cheap, infact,
You’re correct in many ways. However, everyone is missing a crucial element in the way Linux is developed. Linux is developed collaboratively, so many companies (some large like IBM and HP) can share the burden of work in getting new drivers written and push forwards what can be done. They also indirectly benefit from new ideas being implemented at a grass-roots level.
Sun are on their own in this regard with Solaris, and cannot possibly hope to generate that kind of movement.
So you complete ignore the fact that the vast majority of drive development that is taking place isn’t being done by SUN but by third party hardware vendors?
SUN is working on issues like AGP support, PCI-Express etc. etc. What the vendors are supporting are things like RAID controllers, exotic devices. Having seen the type of hardware they’re developing drivers for, I’d feel rest assured that when Solaris 10 is released, you will be pleasently suprised at what will be on offer.
Btw, regarding Linux drivers, most of them are half-baked, beta and in some cases alpha quality. Just check out ALSA and how many audio cards are fully supported or how about features on hardware in regards to drivers included with the mainstream kernel release.
When it boils down to it, when you start ripping out the beta and alpha quality drivers, Linux’s hardware support is just as crappy as Solaris.
With that being said, if they Solaris atleast up to FreeBSD 5.x hardware support, then they’ll atleast be half way there.
cheaper than Red Hat Enterprise, people are now saying, “hell, why not by an x86 machine from SUN loaded with Solaris?”. People are making that choice, and many customers are happy with the direction from SUN.
The problem is that you can look at this the other way around. You get more hardware vendors, and a much greater choice of hardware and options with Linux and a much greater deal of compatibility. With all the will in the world, Sun cannot provide that on their own. Only Microsoft have the resources to do that, and even they’re on their own.
Microsoft doesn’t write a bloody driver, not one line of driver code. Every bit of driver code is written by third party vendors and then incorporated into their product, plain and simple. Microsoft does VERY little low level or application programming, most of it is licensed or bought. Nothing wrong with that, infact, it is a very economical way of conducting business, but lets not try to make out that Microsoft does everything themselves.
> When it boils down to it, when you start ripping out the beta and alpha quality drivers, Linux’s hardware support is just as crappy as Solaris.
Man you are soooo right!.
How to convince IHVs to support Solaris?. Give them incentives. Sun’s getting 2Billion from MS and there’s no time like the present to get IHV commitments (with today’s economy these can be had cheaply).
Sun can either use that money wisely or just keep holding it as they go down. If I were Scott, I’d be throwing a millon to NVidia and a Million to ATI telling them to Support Solaris. I’d throw a couple of million to buy up some of the small Solaris driver companies and get a headstart with graphics/pcmcia/usb/sound/video drivers. Id then start throwning a few million at Adobe, Macromedia, VmWare, Win4Lin, Codeweavers getting them to commit. This effort will take only 100Million dollars – you’re still left with 1.9 Billion!.