Red Hat is perhaps one of the most understated success stories in the software industry. Each quarter, the company registers roughly 20-percent year-over-year growth, fueled by partners that increasingly cement Red Hat's place at the center of the enterprise data center. Despite repeated efforts, no other Linux vendor has come …
I mean, yeah sure it could be Microsoft doing a complete about face and supporting CentOS in some form or other but really? You really think it's going to work out like that? History is laughing herself to tears behind your back.
Great soundbite but ...
"The most open solutions are those that embrace what the customer already has[.]"
Give me a definition of "open" that renders the above sentence true (and doesn't beg the question).
And what in the name of heck does "embrace" mean in a technical environment? My server ain't gonna embrace yours. It might talk to it, if your server is decent enough to use standardised protocols.
Divide and conquer
Personally I’m a great fan (and user) of CentOS. However I am aware that its development depends on work at RedHat, and is being paid for by RedHat’s customers.
Microsoft have never been a positive force for Open Source, and to support CentOS over the paid alternative can only be an attempt to undermine the most significant player, and thus weaken the competition.
Here's an idea..
.. why doesnt MS base their next OS on a unix derivative? same way that MacOSX is based on BSD. I dont think this is far-out at some point the browser is the OS... and yes I think chrome is way ahead of its time. It might seem 'too early' now but they're pushing the envelope on whats possible, not so with MS.
They did once
It's called Xenix, and you see how far that got.
Microsoft is still stuck in the FAT32 and NTFS era, whilst much more efficient filesystems have whipped by [ext3, ext4, etc.]. I don't see them ditching their precious DOS-based hierarchy any time soon.
think the point is that "browser is everything" cloudy stuff means that the client-side filesystem is irrelevant to the UX.
But if MS do as the OP suggests. they's deader than a very dead thing. So they probably won't.
They'll mess around buying and wrecking Skype and parading around their new biatch (that would be Nokia) until Google puts them out of their misery. Or did I just dream that last bit.
OS X not based on BSD
OS X is based around Mach, with portions of BSD userland.</pedant>
Re: OS X not based on BSD
Mach was based on the 4.3BSD Unix kernel. The original intention was to gradually move the core of the Unix kernel so that it sat on top of a microkernel, but for performance reasons this has never happened. Regardless of that, OS X was and still is based on the BSD - both the kernel and userland. When NeXT were paid to save Apple, the first thing they did was to update the userland with code from NetBSD and FreeBSD. The VM subsystem in the kernel was also improved with fixes from FreeBSD, which had previously adopted some of the adaptations from Mach.
The point is that OS X is still UNIX
... was sold of by MS and lived a very long life under the name of Santa Cruz Operation, or SCO. The only thing that killed it was that it was purchased by the ultimate patent troll: Darl McBride.
MS ditched Xenix because they went on with IBM to make OS/2 ... then they split from that project and went on to create NT.
MacOS is about as much Unix as Windows is.
Nothing significant that the user actually comes into contact is Unix. No Unix ideas or approaches are exposed to the end user. Anything that the common novice sees, will likely ever be exposed to,
or view as a valuable aspect of the platform is strictly proprietary Apple stuff.
Trying to call MacOS a Unix is a classic case of Apple fanboys badly misunderstanding technical details.
Are we really saying that people can't create a decent OS from scratch these days?
It's not so much that Windows is flawed because it isn't Unix, it is flawed because it tries to retain so much backward compatibility.
If Microsoft created a new OS from scratch now then it's possible it could be better the any Unix OS. It's just that would cost a lot to do.
I suspect most of you are running x86 which is just as bad as Windows for being full of quirks and backward compatibility.
Pointless title, which must contain letters and/or digits.
" MacOS is about as much Unix as Windows is.
Nothing significant that the user actually comes into contact is Unix. No Unix ideas or approaches are exposed to the end user. "
By this argument, Ubuntu isn't Unixy either, as no Unix ideas or approaches are exposed to GUI users.
Unix doesn't have a standard graphical user interface typically and there's no standard shell, so your arguments about what the user comes into contact with is pretty irrelevant.
Open a terminal on OSX and you'll see the bash shell and standard GNU type commands. It's good enough for me.
NTFS is light years ahead of ext-whatever, and nothing to do with FAT.
@JEDIDIAH - Actually OSX is UNIX
No it's a classic case of you not understanding the technical details or having the proper facts.
OSX is officially certified as UNIX along with Solaris, AIX and HP-UX - see http://www.opengroup.org/openbrand/register/xy.htm.
What has the fact that it's all wrapped up in a sexy GUI (and thank God for that) got to do with anything? The command line is readily available if you're so inclined. Makes for a fantastic desktop OS.
When MS sold Xenix back in the 80s, the terms of the deal were that MS couldn't develop another Unix themselves.
Any technical reports to that effect?
You gotta be kidding...
We need to run CentOS to support customers, so we're going to go pay Microsoft for the privilege of running CentOS on top of Windows? Whaaa? No. We're going to run CentOS on top of CentOS (well actually more likely Scientific Linux which has released 6, CentOS is still suck at 5).
Red Hat loves CentOS. It trains admins in Red Hat Linux and when they start to actually have to convince management to use CentOS 9 times out of ten they just go with Red Hat Linux.
What AC said...
There will always be some *nix on Win, but the momentum will go towards putting free on the bottom of the stack as win admins get more confident with Linux. Free is great leverage.
As for CentOS, development is looking decidedly shaky IMHO. Version 6 should have been out last year. Big problem. Scientific Linux is waiting in the wings as a strong and sturdy contender for the "free Red Hat" crown if CentOS don't manage to recover from their dev stumbles.
CentOS (for now) is doing for Red Hat what Windows piracy in China is doing for MS. Free while it suits, paid once it doesn't. Grows the market no matter what.
CentOS 6 - availability
Yes, I have been somewhat concerned that this has taken as long as it has. However it is now in QA and seems to be imminent, see:
Microsoft? Give up on pushing the "fat client, fat server, oodles of integrated everything else all interdependent on each other so that you can't escape" model?
Are you MAD?
There is absolutely no way on this earth that Ballmer is capable of even contemplating that. I don’t disagree that this is what Microsoft has to do if they want to survive – let’s face it, how does Windows compete with 16 hour battery life? Still, we’re talking about Ballmer here. BALLMER. Saying Ballmer is about to embrace interoperable software is as (pardon my French) fucking batshit goddamned bonkers bananas as claiming Jobs is about to pioneer an open hardware initiative.
Microsoft is not capable of thinking outside of “lock-in, lock-in, lock-in.” They aren’t capable of thinking outside of “traditional fat clients.” They sure as hell aren’t capable of thinking about heterogeneous environments or interoperability excepting as a method of locking you into their dated fat-client lock-in.
The scorpion will always sting the dog. Because it’s his nature.
+1 for lolz and being completely correct
Lock in, lock in, lock in
Everyone wants lockin. On some level, seems to me that every commercial software outfit is all about the lockin wherever possible. Its where the real money is. Anyone who think they can get away with it will do it; they'd be daft not to.
Red Hat LIKES CentOS
CentOS is NOT the "enemy of [Microsoft's] enemy." Red Hat is pretty clear that they are perfectly happy to see "unpaid" Linux installations such as CentOS out there. They've gone on record saying that if it weren't for the open source community, they would not exist -- it is NOT a one-way street. And the piece Microsoft will NEVER be able to understand is that it is NOT a zero-sum game. For Microsoft to win, everyone else must lose. Not so with Red Hat.
Funny you should mention that, I've been expecting Windows X to be more-or-less gnu based entirely ... Microsoft stole their networking code from BSD, and have been drifting ever-closer with each release to the linux/BSD way of doing things ... that gawd-awful "Documents and Settings" (aka /home) folder, adding multiple users and the concept of Admin (aka root), and the remarkable similarity between a modern KDE 4 desktop and Win 7 is rather amazing ... I no longer know who is copying who... All they basically need to do is to get rid of that registry, support ELF and let KDE run as front end.... oh, and get a decent file system. Lots to choose from. :-)
"...Microsoft stole their networking code from BSD..."
That's incorrect for three reasons:
1) You can't steal BSD, as long as you keep within the licence conditions, which they did. It's FOSS, free for everyone to use, you can't pick and choose.
2) It was only the IP stack, not the networking code as a whole. It was only in NT 3.1/3.51, after that all of the Network stack is MS code.
3) It was actually an external company comissioned to write an IP stack for MS that used the BSD code.
I think you are confusing concept and implementation.
It is quite clear that pretty much all of what you quote has been available for many OS's, with UNIX as one of them. For example, the segregation of authority has been in almost all multi-user OS's since computer-time began, and most implementations follow at least in part the Multics model. Ditto 'home' directories (and I can quote from direct experience VAX/VMS, RSX/11M, RSTS, PrimeOS, MTS, Harris VOS, AEGIS and Domain/OS [the earlier versions were UNIX like, rather than actual UNIX], and probably several I can't remember, and - of course - UNIX's from Bell Labs Version/Edition 6 onward and Linux)
In recent history, there has also been a common model of GUI, so I would not want to call whether GNOME/KDE was influence by Win95, with parallel evolution, or whether there is a common work-a-like (for example, whether both of them took anything from the Apollo/VUE/Motif/CDE developments). I certainly would not want to say that Windows 7 desktop was a copy of KDE any more than KDE being a copy of the Win95 desktop.
Many people who look at such things believe that the Genetic UNIX kernel (i.e. derived from AT&T code) has a number of fundamental deficiencies that need fixing by being replaced. Some, though not all, have been fixed in BSD derivatives and Linux, although both have other issues.
If you look at it, Rob Pike, Ken Thompson and Dennis Richie, all movers and shakers in original UNIX, have moved on, and developed (with others) Plan 9 as a replacement for UNIX.
Microsoft believed that they could develop a new OS from the ground up which would be better than anything then available. They did this by using IBM and people from other OS families (including Dave Cutler who was a VMS architect). Unfortunately, whether it was because of the short-sighted view of what it would be used for, or the requirement of backward compatibility with older OS's, Windows NT (new technology) and derivatives ended up where we are today.
If you look at the privilege separation, NT file system and multi-tasking ability compared with other OS's of the time (and Linux was not one of these), and integrated GUI, WinNT should have been quite a capable OS, but something important got lost along the way.
I blame the application model of being able to write where-ever you want on the filesystem (required for non-NT Windows application compatibility, and carried on to NT through poor education and practices by application writers) as the primary problem. If they had put a DOS/Windows 3.1/95/ME comparability mode that used something like chroot to isolate such application from the main OS, and enforced system directories that would never be written to by non-privileged applications for system libraries, utilities and DLLs, used specific User ID's (not Administrator) and directories as placeholders for applications, and made sure that nobody EVER installed NT using a FAT32 filesystem for C:, then I think that NT would have developed into a potent multi-tasking, multi-user OS. That did not happen and it's not where we are now.
BTW I am, and have been for over 30 years, a strong UNIX and latterly Linux advocate, but I do not let that blind me to the merits of other OS's, even if they are only potential.
This is not a title
There seems to be some conflict going on over at CentOS, what with no CentOS 6 out even 6 months after RHEL. Dag Wieers has resigned from CentOS development work and is pointing at Scientific Linux as the distro to move to.
I disagree on many levels
Interesting analysis, but I think you are completely wrong.
Re "Microsoft has grown up, and grown out of its fetish with control of a narrow computing stack"
No it hasn't! That's why it's been so demonstrably bad in the whole mobile and tablet market. The ONLY reason MS occasionally shows interest in anything non-Windows is if it has a longer plan to "embrace and extend" and screw-over the existing users.
Re "My question? When is Red Hat going to do the same?"
Why would Red Hat want to add Microsoft interoperability crud to Linux? There are two main reasons why people are moving away from MS and Windows. Cost and reliability. Lots of people are finally waking up to how bloody unreliable and crappy Windows is. Why would you want to import that crappiness into some other product? You only have to look at the oh-so-successful (and in my mind, pointless) projects like Mono to see how much non-Windows people love this stuff.
On a more general note, would YOU buy Linux from MS? I know I bloody wouldn't! I'd buy it from almost anyone else, but definitely not MS.
Redhat won't add anything non OSS into its product. That's why they won't play interop with MS.
They do support samba for windows servers though.
What a joke!
Typical MS, they choose to chummy up to the OS that is struggling the most at the moment. They very nearly fell completely apart last year (or was the year before?) and their handling of the centOS 6 release has been a shambles: they don't think they'll get centOS 6 final until the end of the year.
Bye, bye centOS I fear. Just look at what MS did to Novell...
No thanks Microsoft. I know what your support is like. You have a problem, call Microsoft and they either tell you to reboot or backup, format and reinstall.
Microsoft are just desperate!
Microsoft are moving from annoying (very annoying) to dangerous now.
Are there no orbital weapons that need testing?
This is the same Sandy Gupta of SCO infamy. The man who claimed, incorrectly, that the ELF file format was copyrighted by SCO and no one could use it without infringing SCO's rights. The man who's copyright analysis was called "indefensible" by Dr. Brian Kernighan. The man who fled to Micosoft when the first set of weals came of the SCO band wagon. And we should believe this guy because of what, exactly.
I wouldn't believe it if he told me the sky was blue, without a signed adaptivity from a noted independent meteorologist.
"Set of weals"?
Floggings will continue until market share improves...
It's a rant
You expect bad spelling. Just be glad I kept my fingers of the CAPS LOCK key and avoided extra punctuation!!!!
Good link. I downloaded SL6 last month but haven't quite persuaded myself to install it yet...
Of course, MS may just have come to the conclusion (along with everybody else who has been waiting 6 months for Centos6) that Centos is dead. That makes this move a cheap way to get brownie points without actually exposing themselves.
Besides, VirtualBox already does this job, for any Linux distro, for free. I've got clients who run Ubuntu on VirtualBox (on XP, I think). Looks great, and is good enough for everyday production work. MS would have to have a pretty damn good value proposition to beat that.
SL vs Centos 6
Lots of debate about Scientific Linux vs centos and release schedules however I find that Centos put a better effort in to maintaining existing releases (for example I don't see a SL 5.6 yet). Yes centos could do things better - I think the aformentioned release schedules and possible ability to donate £ to the project will help.
I think this is not about MS vs RH but rather MS just making sure they get a chunk of the virtual machine hosting business.
Did Sandy Gupta say...
"We have been told repeatedly by our customers that we have to work with other vendors to interoperate with different data center solutions."
"We have been told repeatedly by our customers that we have to work with other vendors to interoperate with different data center solutions. ***Solutions like CentOS.***"
In other words, did *his* words include "CentOS" - the article is unclear about this, as "Solutions like CentOS" appear outside of the quote.
There's a BIG difference.
Terrible article which smacks of a LARGE amount of BS.
As others have stated in the comments, this scenario is massively unlikely, coupled with the fact that Red Hat and CentOS get along just fine thank you very much.
I wish El Reg would put an "Opinion:" prefix in the title of articles like this.
Why does it have to be Windows host + CentOS guest?
It is somewhat bemusing to see Microsoft announce support for CentOS guests, but since CentOS is basically a 100% binary compatible clone of RHEL, couldn't RHEL be a guest too (or will Microsoft look for CentOS-specific stuff (contents of /etc/issue?) and block RHEL)?
Also, Microsoft is basically admitting that CentOS is a good enough OS to be worth providing VM support for. Won't this make people curious about CentOS (and maybe RHEL if they have the money)? Couldn't it even make people realise they can "flip" their guest vs. host setups (i.e. CentOS as host, Windows and CentOS as guests) since kvm is quite a decent free virtualisation system on CentOS that could potentially save you having to buy a Windows Server licence?
It seems to me that this announcement isn't really a "snub" to RHEL, but actually a rare case of Microsoft admitting that Linux is at least as worthy as Windows to be in the data centre.