Analysts report that the shift toward both Linux and Windows dominance in the data center has begun and traditional UNIX is on the way out.
Analysts report that the shift toward both Linux and Windows dominance in the data center has begun and traditional UNIX is on the way out.
The relative percentage of machines in data centers running various versions of UNIX has been declining for quite some time now.
And I think it will continue to die for at least the next 20 years. In the meanwhile, Linux has been modelled after Unix but don’t tell anyone it is Unix…
We have begun to implent a few Linux servers at the datacenter I work for, but HP-UX is by far more prevelant.
Wasn’t Linux developed for the desktop in the first place?
“modelled after Unix but don’t tell anyone it is Unix…”
its a unix like operating system and cannot be called as unix
No, it was just meant to be a Minux clone, which intern is a unix clone/workalike…
BTW also expect to start seeing a lot more FBSD machines as linux paves the path to getting more ppl interested in unix.
“Wasn’t Linux developed for the desktop in the first place?”
who cares. it has come a long way
I’m afraid Windows isn’t going to win in the data centre arena. It just isn’t trusted there, but I suppose it depends on what these people have defined as a data centre .
I’m sorry, but this is an article that praises Linux (no option there really) but it is just to try and hype up some non-existant gains for Windows. You always notice that they mention the revenue share Windows has, which is always going to be greater since it is more expensive. That’s a dead giveaway that something has been presented to make Windows look good.
However, having said all that there is still a lot of proprietary UNIX in data centres – it’s just there.
Platforms, such as the mainframe, are just too expensive, said Greiner, and data center managers are finding it increasingly more difficult to justify the cost of some platforms and the operating systems.
The problem is that you still need to seriously invest and maintain core systems, and consolidate them. You can’t get around that. IBM have a more reasonable idea here, and this certainly hasn’t debunked that:
The message debunks IBM’s theory that data center managers are tired of server sprawl and will increase spending on mainframes and other servers, such as the iSeries, in an effort to consolidate systems.
“The mainframe has a maturity that’s superior to Windows and Linux,” he said. “If that someday changes, that would be great.”
No it won’t. It’s because people are just not going to move mainframe systems – they’re needed every day.
“Wasn’t Linux developed for the desktop in the first place?”
Nope it was designed for the x86 processor. In otherwords is was designed for the personal computer. All the press you see about Linux on the desktop now is pretty new.
freebsd has been growing extremely rapidly, the largest hositng companies in the world have recently started to adopt it more. since solaris is going to be open sourced, linux will have some HUGE competetion because it has the ability to scale past linux and freebsd. (anyone who doesn’t believe this should research it)
One fact that everyone seems to forget is that Mainframe (OS/390) can run for months without the typical ‘server lockups’ or reboots. The only time a mainframe needs to be IPL would is during routine maintenance or it can be done with a hot start on JES2. In addition, security in the mainframe world is way beyond anything Linux or Windows can even begin to comprehend. Mainframe applications like CICS, batch-processing can outperform servers by a mile and provide uptime 24/7 for customers without a hitch.
One very important word on servers, they may be cheaper on the front end but lets not forget the thousand points of failure with a server meaning; Linux-Windows environment. One of these operating systems can neither provide the uptime of an S/390 mainframe nor provide simultaneous logins and transactions with ease. I work in the Enterprise with all of the operating systems, and the mainframe is still a solid workhorse with very few problems. Especially, with the introduction of virtual tape systems and other new storage systems, integrating Linux and Windows is done with ease, including large Oracle or MS-SQL databases.
The day they can build a server that can attain the uptime of a mainframe will be a cold day. Because it will not happen because the operating systems Linux and Windows can not handle the load or system processes to do so. With Linux, it is a ‘tinker’ operating system always requiring tuning and Windows is always having a patch to install. As I said earlier, servers have too many points of failure and problems with dead processes, lockups, and so on. Mainframe OS/390 will be around for a long time into the future and then some.
“Windows, Linux to win war over data center; UNIX Loses”
Yes, X86 wins.
Who listens to analysts anyways. Oh, every body that lost their shirts during the dotcom boom, Enron and Worldcom come to mind as well. Remember when WindowsNT was going to take over the datacenter back in the 90’s.
BTW, linux, unix, a rose by any other name … blah, blah, blah.
The following link is the 50 top of longest Uptimes come from http://www.netcraft.com :
http://uptime.netcraft.com/up/today/top.avg.html
If you wanna deal with the quality, here we go..For a small servers at home or office, i think Mitel SME or a Windows is alright.
And look back to Dinosaurs’ extinction, though they disappeared on earth, their skeletons still deserve for all archaelogists to explore..
have a nice weekends
This chicken little assesment assumes that traditional UNIX systems won’t compete with linux/intel solutions in the future. Not to mention as previous posters pointed out Linux is not UNIX by any means and “UNIX Like” does not cut it in all applications. I am sure linux would grow in datacenter use however I see it replacing the functions of Windows boxes sooner than replacing the functions of UNIX Big Iron or Large mainframe installations. Just FYI for those who do not believe, there are things mainframe systems can do that Open Systems just cannot do.
>>Wasn’t Linux developed for the desktop in the first place?
man.
we really need to work on the caliber of osnews visitors…
where are all the multiplatform people at?
perhaps we need some kind of questionnaire that you have to answer before allowed to post.
>> Wasn’t Linux developed for the desktop in the first place?”
> Nope it was designed for the x86 processor. In otherwords is was
> designed for the personal computer. All the press you see about Linux > on the desktop now is pretty new.
First, this is not a troll…
I am actually curious as to what specifi purpose Linus Torvalds had in mind. Clearly it wasn’t for the desktop in the sense meant by the original poster, but neither was it for the server. As it is UNIX-like, it is well suited to the server, but what specific purpose did he have in mind?