The Red Hat ‘Network’ systems management tool is to gain the ability to manage software distribution and configuration for Solaris, with Red Hat claiming the features are the “the final nail in the coffin” for Sun’s operating system. However the software could be four months away from release.
Red Hat Aust claims Solaris on verge of eclipse
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Or is it the case that F/OSS people seem to hate each other more than they do MS?
Let’s face it, if anyone has reasons to mistrust Sun or IBM or younameit, the fact is that they couldn’t like that company less that they do MS… or could they?
Or is it just the Linux Borg mentality: everything else rather than Linux is ‘dying’ and in fact ‘should’ die? It would seem as though the aim of Linux is to eradicate every other OS by trying to do what it does better? Such an attitude is to be expected from dumb users, but not from people in positions of responsibility – like it seems it has now.
I mean, even Sun’s stupidest bloggers don’t realistically think they’ll kill Linux. The FreeBSD’s and DragonFlies don’t want to strangle each other nor do the Net’s and Open’s. And that’s people who’ve had strong differences. But Linux and Solaris?? When has Sun done harm to Linux?
“When has Sun done harm to Linux?”
Sun execs have made it very clear publicly that they believe Linux is an inferior operating system. They (for obvious reasons) believe that Solaris is a much better product and that Linux will never be able to live up to it. The normal target of Sun’s Anti-Linux venom has always been Red Hat. Compared to the comments that Sun has made regarding Red Hat in the past, this volley from them seems pretty minor, IMO.
-G
Sun execs have made it very clear publicly that they believe Linux is an inferior operating system. They (for obvious reasons) believe that Solaris is a much better product and that Linux will never be able to live up to it.
Exactly; but my point is that they’re no saying ‘Linux is dead’, ‘Linux should die’. When they offer superior features, they do it in a way that says ‘you’ll have to catch up to it and when you’re done we’ll be miles ahead’; whereas from the Linux side it’s always ‘we have this (or will have!) that you don’t have, so you’ll shrivel and die because once we do something better than you do you’ll never be able to recover’. For instance, Linux lagged behind FreeBSD in many areas for years. Now, with the mess that the 4->5 transition has been, Linux reportedly has the edge, so ‘everyone’ believes FreeBSD has entered terminal state and can only shrink from here. And tell me, when has FreeBSD done harm to Linux?
The normal target of Sun’s Anti-Linux venom has always been Red Hat. Compared to the comments that Sun has made regarding Red Hat in the past, this volley from them seems pretty minor, IMO.
It’s not the volume of abuse; it’s the perspective. On one side they (Sun or anyone else, MS lately with their ‘security’ analyses) bash Linux for its defects (being monolithic, being OSS, etc). The BSD people claim Linux is unelegant or insecure. But Linux… Linux people just ignore all the bashing and say ‘you’re dying’. That’s not pretty. To bash a product is one thing, but to continuously say it’s dying is on a different level. That’s what MS did to OS/2.
“Such an attitude is to be expected from dumb users, but not from people in positions of responsibility – like it seems it has now.”
This is the kind of tactics used by Microsoft to derail Novell a decade ago. Unless I’m wrong, thousands of responsible CTOs throughout the world fell for this trick and replaced Netware with an inferior product (Windows NT). The same goes for OS/2, WordPerfect, Internet Explorer, etc.
Apparently, in the software industry, good managers must resort to such antics to attract customers. By the way, I find this practice quite appalling.
Or is it just the Linux Borg mentality: everything else rather than Linux is ‘dying’ and in fact ‘should’ die?
Eh, I don’t think so. At least I don’t treat anything that isn’t Linux as though it doesn’t matter. I run Gentoo Linux on my Desktops, but I use FreeBSD for my e-mail server on one my Soekris boxes, a highly minimized version of FreeBSD called m0n0wall on my other Soekris which is a WAP/Router/Firewall, and OpenBSD on my general purpose server on my other Soekris.
I like the *BSD’s a lot. (Except for NetBSD. I don’t really see where it has a purpose except as a proof of concept OS to port to yet another platform. I’m probably missing something though.) But for Desktop purposes that include more than just e-mail and web surfing, I think Linux clearly has the edge. (Please notice the qualification in that statement if you feel compelled to reply. *Desktop purposes that include more than just e-mail and web*. I’m not talking about a development workstation, or a general e-mail/web box. FreeBSD can handle that just fine. Possibly better/faster than Linux.)
Competition is a good thing. Even among OSS projects. I think the zealots that are in that “everything but Linux is dying” crowd are just comfortable with the one 1, non-Microsoft thing they know, Linux, and are too afraid to try anything else.
> Sun execs have made it very clear publicly that they believe Linux is an inferior operating system. They (for obvious reasons) believe that Solaris is a much better product and that Linux will never be able to live up to it.
That is absolutely true, Solaris is a much better OS than Linux, what’s so groundbreaking about that? RedHat is starting to get its panties in the knot because they’ve a product on their hands that can’t really compete with Solaris on either price or features, this is why they are starting to adopt the FUD tactics. I think Solaris will take a good chunk of market share from RedHat because picking Solaris makes much more business sense than Linux — Solaris is cheaper, more full featured, more scalable, better performing, more stable and more mature. Solaris can even run RedHat programs unmodified, why on earth would you pick RedHat? The only reason RedHat will get picked over Solaris is when the customer is so hopelessly hopped up on GPL cool-aid poured down their throats by RedHat et al, that they will pick anything that says GPL on it with no regards to price or features. BTW, I’ve already scrubbed RedHat off my machines not only at work, but also at home, Solaris 10 is awesome!
…as up2date is crap, rhn is commercial and yum/apt (or even yast2!) are infinitely superior.
plus, if redhat gets in on it, they’ll replace sun’s excellent pkg’s with crappy rpm’s!
It’s inferior to Solaris–ofcourse it is.
Mr. Schwartz loves linux. I have e-mails with headers saying that he does.
if i were red hat, i’d port my stuff to solaris and release an OS based on solaris and run with it. that would kill RHEL and their linux competitors.
I think Solaris is superior to Linux in the way you can easily fine tune security amf performace; what tp restrict serial ports or just make it so you can use a terminal one them? easy. Want to tighten down the system so the user cn’t even change their desktop wallpaper? Easy. All with Solaris Management Console. Nothing on Linux comes close.
That said, Red Hat is full of it. I mean, their update tools are way sucky and to think they want to use them on Solaris…ick.
Solaris is better, but still waiting to see a superior desktop system.
It’s funny that I attended a seminar today which representatives of Sun, AMD, Oracle, and RedHat were present. The RedHat engineer was “politically correct” and didn’t bash Sun in front of a group of mostly Solaris. The discussion was about the Management Server and RHEL 4. The first problem we (the group of admins I work with) was the starting price of $13,000 for the Management Server software. The other announcemnt I found interesting is RedHat saying that RHEL 4 would be CC EAL4 evaluated, well that is for Update 1 which will come out “later this year”.
During the discussion about the Management Server I kept saying “it’s a poor man’s Tivoli” and for the capabilities it has, can be easily duplicated with other F/OSS tools that even if you paid for them, would cost less than $13,000. The first thing that comes to mind in monitoring is Nagios.
I personally feel that some people at RedHat need to put the crack pipe down and listen to Solaris, AIX, and HP-UX admins. If RedHat is trying to convert existing networks to use their products, they not only have to win over management, they have to win over the administrators too. In that regard RedHat has some serious work to do, because I am not convinced their products are superior, especially above 8 CPU’s. And the solution to every computing problem is not a cluster.
All of the nix’es are going to grow!!!! There is room for all OS’s. It is only MS that will lose market share. Solaris, Linux, BSD, …, are all fine OS’s and will eventually interoperate with each other well because of open standards.
What I eventually expect to see is a variety of OS’s optimized for specific purposes. I would love to have a computer and OS optimized for the civil engineering environment I work in. Make it secure by throwing out everything you don’t need for an engineering office and give me GREAT graphics. I stare at a screen 10 hours a day.
Thank you.
Tom
and whatever Solaris’ technical advantages are, it doesn’t match Linux for usability. But, I haven’t tried Solaris 10 yet.
And, I think that the Redhat Australia is taking the piss if they think that Solaris will simply disappear because they’ve come out with one good tool.
> and whatever Solaris’ technical advantages are, it doesn’t match Linux for usability
Usability in what sense? Solaris is running the same desktop environments as Linux and is all around a better put together OS. I would rate Solaris above Linux in usability as either workstation or server.
Well, I’ll in fact reply since you raised a number of interesting points:
I like the *BSD’s a lot (…) Butfor Desktop purposes that include more than just e-mail and web surfing, I think Linux clearly has the edge.
Well, you’re right. There’s just so much than can be done right away with Linux that can’t be done in the BSD’s without some effort or at all. And I do love the BSD’s (including Net!). It somehow saddens me that I don’t see that gap decreasing, as I think the somewhat barebones approach of the BSD’s is both their strength and weakness. It’s probably no coincidence that when Apple (NeXT) decided to build a full workstation OS, they did it as a separate layer. Linux went the opposite way, and the BSD’s are left profiting from a subset of what’s been built around Linux.
I think the zealots that are in that “everything but Linux is dying” crowd are just comfortable with the one 1, non-Microsoft thing they know, Linux, and are too afraid to try anything else.
What strikes me as appalling is that responsible people engage in that sort of prophesising too. As you said competition is a good thing, and I’ll give one more example: even MS has done good things. Look at Windows 95. It’s a stability and design nightmare. Look at OS/2 Warp 4, which is about the same time frame. Great OS. Great stability. Great efficiency. Now look at its UI. Great concept. Reasonable implementation. Ghastly looks. Video/Gaming API? What’s that? Look at the file system. What, everyone thinks it’s alright to dump files and directories in the root partition’s root directory?
On both accounts, Win95 and its successors went the right way. Where would we be, graphics-wise, if OS/2, driven by IBM to please the corporate user, had won the parade? I love OS/2, but I have to recognize it suffered from the same sickness as Windows or proprietary UNIX: the owner catered to certain features and just ignored demand for others. MS ignored security, stability, modularity and just about anything good, but focused on home user entertainment. And I think the user wouldn’t be so entertained today, running mplayer and 3d-nethack, if it weren’t for MS. No matter how much I dislike their practices and products.
because ultimately Microsoft is the loser. Open standards allow Solaris and Red Hat to play nicely on a network, while Windows requires teams of people to reverse engineer packets to create Samba. The path of least resistence is clear.
Sun and Red Hat definitely get some enjoyment out of bashing eachother (bring in IBM or HP for a complete set of stooges), but that is business as usual in the computer industry.
With minimal modifications the vanilla Linux kernel will scale to atleast 64 processors; also SGI ships a single system image Linux kernel to run on systems with 512 processors.
Linux can handle 8 processors, no problem.
On most servers, I do not see how Solaris 10 is superior to GNU/Linux.
For instance, Linux lagged behind FreeBSD in many areas for years. Now, with the mess that the 4->5 transition has been, Linux reportedly has the edge, so ‘everyone’ believes FreeBSD has entered terminal state and can only shrink from here.
I’m curious, in what areas has Linux lagged behind freebsd for?
> With minimal modifications the vanilla Linux kernel will scale to atleast 64 processors; also SGI ships a single system image Linux kernel to run on systems with 512 processors.
Right, Linux can run highly parallel workloads on a large number of processors, there is no feat in that and Altix with its Numa-flex interconnect makes a good use of it. But if you throw a workload that requires a high degree of communication between processing nodes at Linux, it will crawl. Linux will seriously suck on transactional database workloads with large number of processors. Even Linus Torvalds and John “Mad Dog” Hall explicitly stated that Linux will have problems scaling beyond 16 CPU’s on this type of workloads. Reason for that is Linux kernel is very coarse grained and lock management will grind the kernel to a crawl when a large number of CPU’s are involved. This is why SGI does not even attempt to target Altix as a database platform, it is just a highly parallel number cruncher. Solaris kernel on the other hand is fully multithreaded and is very fine grained lending itself very well toward systems with large number of CPU’s. This is exactly why Solaris has been the predominant choice as a platform for high end databases and see it changing any time soon. For business applications (databases, ERP, CRM, etc.) Solaris is still a much better platform than Linux as far as scalability is concerned.
> On most servers, I do not see how Solaris 10 is superior to GNU/Linux
Besides being a cheaper, more scalable, more proven, more stable, and more mature OS Solaris has such things as dtrace, zones, and soon ZFS — something Linux won’t have for long time. Solaris 10 is miles ahead of Linux on pretty much any comparison.
The funny thing is that for Solaris’ supposed scalability, I’ve have never seen it used by any institution for super computing. Maybe I haven’t searched hard enough. But today, I’d wager there are more super computers based on Linux than there are on Solaris.
Yet some mindless bigots have the balls to dismiss Linux as a scalability wimp. It boggles my mind.
Anonymous (…iup.edu) should know better — there is a whole
lot more to scalability than just HPC.
Why do you think Oracle runs their backend servers on high-end
Sun hardware and Solaris? Because it scales to service the
business load that they throw at it.
It’s great that linux is so popular for HPC — perhaps that
is where its true market really is — HPC most definitely
allows linux to shine as evidenced by the top-X lists of
supercomputers. However, that’s just HPC. Sure, big oil,
big geotech and big pharma (and Defence/No Such Agencies all over
the world) run HPC stuff, but they also run Solaris for
their business purposes. That’s why Sun has enhanced it
over the years to provide the near 1-1 scalability that
customers keep coming back for. That’s why Sun claims that
Solaris is more scalable.
Oh, and when you can spec and purchase a 32-cpu intel or
power box running linux for business purposes which is not
a custom order with a customised kernel, and that comes
anywhere close to a 32-cpu ultrasparc box please let the
world know because we’ll all be mighty interested in your discount!
Hi James,
I might not agree with all the comments above, but the whole point of the GPL bigots as you seem to call them is that if needed, the kernel can be modified by anyone to suit their problem/solution. Till recently, you could not do it for the solaris kernel. Even now, the traffic will always be oneway. You can always use other open source O.S. technology to improve solaris ( if necessary), but the other way is barred. That is not playing fair atall. So if they play foul, why can’t the others ;^)
Solaris for sure is a better OS that Linux. I have used it and found it to be much more stable and efficient. Solaris 10 looks promising with features like dynamic healing and dtrace.
Solaris is not going to die. It is the best unix and it will stay here…whether red-shyt like it or not
OSnews is now my favorite RedHat vs Sun news site. Even through all the BS, the nice thing is that end-users win. Sun wouldn’t be nearly as congenial today if things hadn’t taken a turn for the worse. Remember, they chose profits over marketshare growth during the dotcom era!
It’ll be interesting to see how RedHat responds. My guess is they’ll start providing free downloads of RHEL. They’ll be without any support, but CentOS and Whitebox notwithstanding, it is now easier to for amateurs to acquire Solaris than it is RHEL or SuSE Linux Enterprise Server (SLES).
I laugh when people complain that KDE and Gnome waste a lot of effort by producing two graphical desktops. The fact remains, KDE and Gnome have pushed each other tremendously in the last couple year. Competition breeds excellence. Let’s see what the New Unix Wars will give us.
Why do you think Oracle runs their backend servers on high-end
Sun hardware and Solaris? Because it scales to service the
business load that they throw at it.
—
oracle development platform is redhat linux for a long time now
Not that I even have considered RedHat for anything earlier, but this is just one more reason to stay away. Lame FUD marketing speach, couldn’t care less.
James didn’t say development platform. He said backend servers. Oracle runs their business on Sun.
alan.
> oracle development platform is redhat linux for a long time now
Yeah, you wish, Solaris has been the primary reference and development platform for Oracle database for a long time and I know it for a fact that it still holds true. You know why? Because two thirds of all Oracle shipments are still on Solaris. It is going to be a cold day in hell before Oracle starts putting RedHat in front of Sun. Sun is the platform of choice for Oracle DB even within Oracle — if Oracle chose to run its applications on Solaris, that there must be a pretty good reason for it.
I guess most of you will have read the original article, Sun has responded through the same journalist at http://www.zdnet.com.au/news/software/0,2000061733,39181514,00.htm
Link offered sans comment.
Alan.
“I’m curious, in what areas has Linux lagged behind freebsd for?”
performance, stability, usb support, p&p, …
I also blogged a response to the initial article before I saw Sun’s formal one. Anyone interested can have a read at http://blogs.sun.com/roller/page/tpenta/20050216#red_hat_vs_sun
Alan.
> It’s great that linux is so popular for HPC — perhaps that is where its true market really is — HPC most definitely allows linux to shine as evidenced by the top-X lists of supercomputers.
Linux succeeded in the HPC market not because of some magic scalability cheese that Linux kernel hackers put into the kernel. Actually Linux scalability has absolutely nothing to do with it. There are two primary reasons why Linux is popular in HPC:
a) Linux was free of licensing costs, which adds up when you have hundreds or thousands of machines in the cluster and
b) Linux was capable of running on cheap commodity hardware, which also adds up if you have hundreds or even thousands of compute nodes in the cluster.
Actually my view on this is even lowly Windows could be just as popular in HPC space if the per node licensing cost where low enough. HPC workloads are highly parallel and the scalability of the OS in the single image matter very little, so that fact Linux is used very widely in compute farms say absolutely nothing about scalability per se.
Solaris and OpenSolaris in particular on the other hand have some pretty good potential in the HPC space and should take a chunk out of Linux simply because OpenSolaris is free just as Linux but also has wonderful tools like dtrace. Dtrace can be a very big reason to choose Solaris over Linux for HPC applications simply because you can identify the bottlenecks quicker and make your apps run faster.
“> oracle development platform is redhat linux for a long time now
Yeah, you wish, Solaris has been the primary reference and development platform for Oracle database for a long time and I know it for a fact that it still holds true. You know why?”
solaris was the development platform
linux is the present development and reference platform for oracle. make no mistake about that. I am working with both oracle and redhat linux as part of our ISV relationships. this is the larger part of the unbreakable linux campaign and also the reason why redhat linux and oracle have combined centers at many places
“http://www.infoworld.com/article/05/02/15/HNredhatsolaris_1.html“
the other side of the same story
> this is the larger part of the unbreakable linux campaign and also the reason why redhat linux and oracle have combined centers at many places
I’m sure Oracle will change sides pretty quick to Solaris again, since RedHat’s value proposition is pretty weak compared to Solaris. Solaris overall seems like a better platform for Oracle to attack M$ and SQL Server.
”
I’m sure Oracle will change sides pretty quick to Solaris again, since RedHat’s value proposition is pretty weak compared to Solaris. Solaris overall seems like a better platform for Oracle to attack M$ and SQL Server.”
first it was that solaris was the development platform. the value proposition is not based on a particular product but as a platform for an ISV. the end charm for an ISV remains higher for Linux much more than solaris now and wont change in a hurry if at all anytime.
in turn you cannot compare solaris as a product with a redhat software subscription which includes support. if you want the product and not support similar to the free solaris binaries model then you can get it from centos.org
> in turn you cannot compare solaris as a product with a redhat software subscription which includes support.
Factor in support for Solaris and the comparo becomes pretty clear — Solaris is 40% cheaper to license and support than RedHat. This is where the value proposition comes from — much better product at much lower price, don’t tell me ISV’s are not going to jump on that (Oracle included).
“Now, with the mess that the 4->5 transition has been, Linux reportedly has the edge, so ‘everyone’ believes FreeBSD has entered terminal state and can only shrink from here. And tell me, when has FreeBSD done harm to Linux?”
Unix is unix is Unix.
Anyone with a bit of experience in a few of them will realize that and not give 2 whits about a popularity contest. The ones that do think there’s that much of a difference outside some implementation details can be ignored (unless you work for them!).
Right now, there are 3 types of operating systems being developed actively; Unix/unix, Windows NT+, and small nitch operating systems (often similar to unix). OSX is unix. Solaris is Unix. Linux is unix. The BSDs are unix. Windows NT+ and Windows for pocket PCs is Windows — with the NT+ branch having POSIX and other unix extentions. QNX is unix. Palm OS is propriatory moving to a Linux kernel (and thus unix). Symbian…is Symbian. There are other operating systems, though the ones that are actively developed tend to be for embedded systems and can work on devices too small for even a stripped down BSD or uLinux.
I have been a linuxuser for years and I don’t think I will drop linux for any other OS in the near future, but it feels good to know that another OS has been opensourced, and I will probably install and try it. I think poeple wouldn’t like Linux if it was an proprietary operating system. Linux is likely to replaced by another free OS someday and that’s nothing to be sad about. OS’s come and go but we have chosen to protect our freedom by supporting the free/opensource system and by standing firm behind the GPL and compatible licenses. It isn’t a war between OS’es but rather a war between those companies and individuals that try to steal out freedom by locking us into their systems and by aquiring harmful softwarepatents and those of us that are aware of these implications and try to fight back our freedom.
Well, the reason why Microsoft were so successful with tactics like this is that they knew just how and when to use them.
Red Hat may allow Solaris to be supported by it and then Red Hat can slowly consume it over time, but given that people are and have been replacing Solaris with Linux for some time, what’s the point? I’ll be interested to see whether Red Hat get the balance of this correct.
Another plus point of Red Hat is their philosophy:
Robertson did not seem to be impressed with the practicality of Sun’s demonstration of good faith, saying: “Red Hat’s policy is to be an open source provider of technology. Our competitors are providing a hybrid of open source and proprietary software – that’s not the model we’re going after”. He went on to say: “We’ve had a very clear focus on Open Source technology and have been doing it for over ten years, now other organisations are seeing that this is successful and have started to copy what we’re doing. But this is not something that you can start doing overnight”.
I think Red Hat have that exactly right, as opposed to Novell’s stupid and confused both-source strategy. Red Hat don’t need to employ any protectionist tactics in relation to their software as it is all open-sourced and mostly GPLd. What happens when Novell’s Groupwise or Zenworks is threatened by open source alteratives? There’s no such problem for Red Hat as they’re already there.
performance, stability, usb support, p&p, …
FreeBSD has not led Linux in any important field of performance in years. Before 5, FreeBSD lacked a multi-threaded network stack and relied on LinuxThreads for a efficient threading library.
GNU/Linux is just as stable as FreeBSD, in maintaing performance.
i love it! massive tech companies bickering over who has supported OSS longer. when i see stuff like this, all i can think is “Man, eric and the osi did one hell of a job”. it wasnt that long ago when free software was thought of as something commie hippy hackers did behind closed doors…
BSD has always had superior tcp/ip stack performance. This was by a wide margin during linux 2.2.x and a narrower but still distinguished margin in 2.4.x onwards. Where FreeBSD may have lost leads in some areas since 2.4.x, TCP/IP stack performance has never been one of them.
As for BSD “relying” on Linux threads, it has the option to run them but it is hardly dependent on them. The native threads is more efficient for some purposes, less efficient in others. Which you use depends on the task at hand. Just like how you should be choosing operating systems.
And as for stability, even the most conservative distributions like Debian would have a tough time proving they can match the stability of the production branches of any of Free/Net/OpenBSD. If you take Linux as a whole and include distributions like Gentoo and Mandrake, there really is no contest.
The Solaris fans totally discredit Linux.
The Linux fans totally discredit Solaris.
Bottom line, both are great OSes, both with strengths and weaknesses. Both have strong corporate support.
Solaris is probably better for the really big-iron, mulit-processor servers, as well as some very advanced administration and multi-threading capabilities.
Linux is probably better for clusters, desktops, a larger variety of hardware support, a larger community, and a more open license.
Linux has eaten Solaris’ lunch over the last 3 years or so, as Sun was a bit asleep at the wheel. However, Sun has woken up and Solaris is now positioned to gain some of that back. Game is on. Competition is good. Choice is good.
Both Linux and Solaris should aim at MS. Why would anyone in their right mind choose a Windows server over a Solaris or Linux server?
Fanboys in both camps need to chill out about each other, and gang up on Windows. Solaris and Linux are both great Unix implementations, and both much better, as servers, than Windows.
“Fanboys in both camps need to chill out about each other, and gang up on Windows. Solaris and Linux are both great Unix implementations, and both much better, as servers, than Windows.”
Amen. How strange that the proponents of either OS forget who the real enemy is, about those folks in Redmond who still own the marketshare. You want to impress us, Sun? Start taking marketshare from MS instead of worrying about Linux. Show us you can make an all around, more useable desktop OS than Windows.
Why do they need to gang up on Windows? That’s no different than what’s happening now. I want to be able to choose between multiple Operating Systems for different tasks. Windows is one of my choices for certain things. What really needs to happen is that all involved need to chill out and quit trying to engineer a way to remove all choice and just focus on making sure that they are chosen by building a suporior product. Sun seems to be the only player involved who “gets it” in this area. MS certainly doesn’t, and the attitudes displayed by most of the pro-Linux camp lately show that they have more in common with the folks in Redmond than they like to think. At this point, Sun and Apple seem to be driving up the on-ramp to the high road, while everyone else is still sneaking around the parking lot looking for tires to slash.
Hey RedHat…that nonsense just cost you a sale! Are you listening? I use Linux and Solaris and I want it to stay that way. I will not support your efforts to restrict my choices. Novell, you’d better be paying attention too.
Neither Sun or RedHat is interested in supporting the typical home user desktop, and why should they? There is much more money to be made in selling support, services, and solutions to medium and large customers.
Sun is positioning the Java Desktop System as an Enterprise Level desktop using configuration management tools to “standardize” desktops. If Sun or RedHat is going to “put the hurt” on Microsoft, they are going to have to start at the corporate desktop first. The vast majority of home users use the same thing they use at work. And it is going to take a lot of effort to convince “joe six pack” that using a *nix variant is better than using the friendly “point and click” world of Windows. Apple has a better chance of conquering the home desktop than either Sun or RedHat since MacOS X has its Unix insides totally hidden by a great GUI.
I am a Solaris Admin but I spend my free time rooting for Linux as long as it is not RedHat. I do not share the vision of RedHat and I do not believe their business strategy will survive. RedHat has tried to profit from the sweat of others by passing off open source software as proprietary at worse than shylock prices. In doing this they have put themselves in a position where they neither have the clout of Microsoft nor the hardware foundation of HP and Sun.
I have come to the conclusion that the only workable Linux strategy is to sell hardware appliances(server, desktops, firewalls etc) pre-installed and pre-configured with a 100% open source, freely downloadable and shareable Linux software. A little known company by the name of Linaire has figured this out I hope that McNealy will let go of some venom and do the same.
If I were Sun, I would adopt this stategy for Linux. Give away the software for free to anyone that wants to download, make a true commitment to the spirit of give and take that comes with open source and Linux but agressively sell Intel, AMD, PowerPc and Sparc boxes already pre-installed with their own Linux distribution—everything from supercomputers to desktops.
If they do this, RedHat will die in 2 years and Sun can use the power of 10,000 open source developers to fight Microsoft and their enablers like Dell to a standstill in every niche of the computer market.
FreeBSD has not always had superior TCP/IP performance, Linux 2.2 series could beat FreeBSD in efficiency and scalability.
I said FreeBSD relied on Linux threads for efficient threading. libthr is not efficient, neither is KSE.
As for stability, I would say Debian is by far more stable and conservative of their release quality, as of late.
First, please go read some detailed material on the inner workings and performance of FreeBSD vs Linux historically. Until Linux 2.6, FreeBSD was indeed better than Linux on the kernel level in many ways.
The only thing Linux had on FreeBSD was better SMP support. Probably the biggest advantage FreeBSD has had over linux performance wise in the past is the VM subsystem. It has also had many features ahead of Linux. For instance, FreeBSD had USB before Linux. FreeBSD had software RAID before linux.
In terms of stability, it doesn’t matter if it’s Debian, Gentoo, Redhat or any other distribution. Under extremely heavy load it used to be that you could not even get a login shell to open on linux. FreeBSD is still usable, albiet slow, under massive overload. There are other examples of stability as well. For instance, if a user runs the file descriptors out on FreeBSD, it only effects that user. On Linux, suddenly no one can login, deamons start failing, etc.
You know, it IS possible to be better than Linux. Why you insist on demanding proof as if Linux is somehow holy is beyond me. Sorry to burst your bubble, but every peice of software has faults in given areas. Linux, FreeBSD, Solaris, all of them. I highly doubt that it will ever be the case that one single kernel will be best at every possible application.
First, please go read some detailed material on the inner workings and performance of FreeBSD vs Linux historically. Until Linux 2.6, FreeBSD was indeed better than Linux on the kernel level in many ways.
Given actual tests, such as the MySQL one, FreeBSD just does not stand up at a kernel level even against Linux 2.4. Just look at it on SMP systems.
The only thing Linux had on FreeBSD was better SMP support. Probably the biggest advantage FreeBSD has had over linux performance wise in the past is the VM subsystem.
I don’t think that’s the case now, but yes, FreeBSD has always had a decent VM system.
FreeBSD had USB before Linux. FreeBSD had software RAID before linux.
Yes, and Linux has now surpassed it.
FreeBSD is still usable, albiet slow, under massive overload. There are other examples of stability as well. For instance, if a user runs the file descriptors out on FreeBSD, it only effects that user. On Linux, suddenly no one can login, deamons start failing, etc.
This isn’t the case now – a 2.6 system handles this scenario very well. However, this is one thing where I can vouch that FreeBSD (and other BSDs) have indeed been very, very good.
Because they’ve apparently decided to toss out some staff!
http://www.theregister.co.uk/2005/02/16/sun_os_layoffs/
Bummer. I’ve been laid off before. Not fun. Especially when CEO’s get “fired” and get a $42 MILLION severence check…
Some interesting facts about Solaris..
1. Oracle uses E15k’s for it’s ERP/CRM environment. That’s right, they use Solaris/Sparc systems to run the business backend.
2. Pixar uses Solaris/Sparc equipment for rendering films! That’s right, look at the credits for “Finding Nemo”.
3. The Philadelphia Stock Exchange makes extensive use of Solaris/Sparc systems.
4. In many military command centers, you’ll find Sun Rays. Even Air Force One makes extensive use of Solaris/Sparc and Sun Rays.
5. Many financial companies make extensive use of Solaris/Sparc systems for databases, number crunching, and datawarehousing. Just look at job postings on monster.com.
6. You’ll find more Solaris/Sparc systems being used in hospitals. Lots of MRI systems are controled by as Sun workstation:)
7. You’ll find lots of Solaris systems running company backends. Sun doesn’t have to sell a new system to the same customer every year, the equipment lasts and the software is solid. Sun provides a complete solution for businesses with a Unix requirement. Red Hat does not do this, remember.. they only make software.
The point being, that Sun makes excellent products that companies can rely on. With the way things have gone in the hardware world, it makes sense for Sun to support x86, Opteron, and cheap Sparcs. Sun can compete now, they just need to get the message out. With Solaris 10, they have an advantage that’s challenging Red Hat and Novell. This is good for everyone. It’ll force everyone to work harder to match/exceed such features. Of cource by the time that happens, Sun will release Solaris 11. Remember, it’s taken Sun 3 years to make Solaris 10.
As for spreading FUD. Well guess what, everyone does it in every industry. Get over it, that’s how marketing is done. The bottom line is “Did the money you spent on a Dell running Linux save you enough money and give you more than you had before?”
With the pricing that Sun has for Solaris and the fact that they are the number one Opteron server distributor, should open people’s eyes to a change around the corner. Oh, and for those who make the point that most Sun Opteron systems are running Linux.. that’s because Solaris 10 with AMD64 support wasn’t around last year. Now that it’s out, things will change:)
Anonymous (IP: —.dq1sn.easystreet.com), please do not
conflate HP’s severance package for Carly Fiorina with Sun.
By way of comparison, McNealy’s salary is USD100,000 a year.
This is what he gets from Sun if – under his guidance – the
company does not meet the performance targets that the board
sets for him. So for the last few years, he’s been exercising
a lot more of his options than previous — USD100k only goes
so far.
This is all a matter of public record, btw, on finance.yahoo.com
(and all the others) “insider” reports, and in official
filings with the US SEC.
And by the way, yes, I’m p*d off that folks got laid off.
I’m glad it was nobody in my division, but nobody is
immune….
FreeBSD has not always had superior TCP/IP performance, Linux 2.2 series could beat FreeBSD in efficiency and scalability.
The 2.2.x branch lagged way behind FreeBSD in performance. Every benchmark testing raw TCP/IP performance always conclusively favoured FreeBSD, and by a wide margin. One of the most celebrated points when the 2.4.x branch came out was the significant improvement in the TCP/IP stack, which then came close but still could not match or beat FreeBSD’s. But by all means, feel free to support your statement. I’d be interested to see what kind of benchmarks you can list.
I said FreeBSD relied on Linux threads for efficient threading. libthr is not efficient, neither is KSE.
What exactly do you base “efficiency” on. Some applications run faster with native threads, others run faster with linux threads. I don’t know how you conclude that one is efficient while the other is not. There is a choice for a reason.
As for stability, I would say Debian is by far more stable and conservative of their release quality, as of late.
Debian has always been very conservative. I don’t see how this could have changed as of late unless they started backpedalling to earlier versions.