There seems to be at least one UNIX-based OS missing from this list.
In a mature IT market, it becomes hard to make any significant changes in hardware architecture or software design without upsetting the installed base of legacy users. This, of course, makes the evolution of a product somewhat troublesome. Change must fit within the strict confines of compatibility, ensuring both hardware and …
What are you kidding? HP's never late! Just ask Matt the HP Sales Grunt. He'll tell ya'. HP is never later than what they intended to be in the first place. You were just expecting the wrong date.
As far as progressing at Glacial speed, that's not necessarily a bad thing in the enterprise. Of course, that's where HPUX and AIX are focusing (as are VMS and the like) all their relative effort. Of course it's much easier in the Enterprise market to wait for the little guys (read Sun) to do the actual OS advancements and then just copy them (read WPAR's and Probevue). Actually I should be a bit more clear. HPUX does not seem to copy the little guys (read Sun), it seems to actually copy AIX, which copied the little guys (read Sun) in the first place.
Linux copied everyone, but then because of their development process they started coming out with some very interesting features. The big guys ignored Linux too long and now Linux is very close if not already having surpassed the big guys in features. As far as reliability, stability, and scalability Linux leaves a LOT to be desired, but it will more than likely get there. Of course everyone said that Windows would get there as well, and look at Windows today.
Well, OK on the desktop market share vs Linux. The article focuses IMHO largely on operating system features that do not really show up at the GUI and matter most to sysadmins: what percentage of the server market runs on Mac OS 10?
Apple concentrates on GUI and app 'improvements' far more than on basic operating system-level improvements. Knowing something about Linux, Mac OS 10, OSF/1 and Solaris, I think I would choose Mac OS 10 last on the basis purely operating system features.
Just my $0.02.
I'm not sure where you got the idea that HP-UX 11.31 shipped in November 2007.
The first shipments were in February 2007, the 1st update was September 2007, the second update (not first) was March 2008 when you got the new packaging. The fourth update was September 2008.
The nPar functionality is a feature of the hardware platform, you can't strip it out of the OS, you got cell based hardware, you got nPars. The vPar functionality wasn't previously included in any of the bundles, you had to buy it separately.
The vPars version released in February 2007 allowed the support of 11.23 vPars mixed in with 11.31 vPars. The new function this year was allowing PA-Risc servers to mix 11.31, 11.23 and 11.11 all within a single hardware partition. Itanium servers were never supported under HP-UX 11.11, but they have been able to support mixed 11.31, 11.23 environments since the initial release of 11.31.
December 2007 saw a release of vPars that allowed PA-Risc servers to mix HP-UX 11.23 and 11.11.
Talking about IBM, mainframes and the slow pace of change. The article implies this is a bad thing - it's not. One of the biggest selling points for mainframes is that you can shut down your system on a friday night, after the close of business, rip all the old hardware out, roll in a new box and be up and running on monday morning. Better still, no-one will know anything's changed.
Given the mess of versioning conflicts, incompatibilities, unexpected problems, licensing issues and sub-year obsolescences, it amazes me that so many organisations put up with the hell that is windows server upgrades. Maybe the best thing that could happen for them is that the big players don't release anything new for a year or two. I have a feeling that most organisations, for whom IT is merely a means to an end (i.e. to help make money), would collectively breathe a sigh or relief at the possibility that they could just get on with running their businesses., And that they'd love a year off, without having to spend millions just to keep their IT systems at the "minimum acceptable" levels that the constant changes from vendors, patches, forced-upgrades, shortened life-cycles and bug-fixes force on them.
Did I blink and miss the MacosX coverage?
Personally, since I only ran Vista on test boxes, and wiped them after, oeprating systems have been fine for me this year. I've been using a mix of Ubuntu desktop 32 and 64 bit, 64 bit server, Solaris, XP and MacosX. Things have mostly been working nicely, Umbongo 8.10 is surprisinngly lovely- and I speak as both a Slackware bigot of many years standing, and someone who owns shiny Macs. Amazing that they're satisfying my needs. Linux on my eeePC901 is lovely too.
So yeah, pretty good. Using windows mostly for games, and doing all the other stuff on better platforms has paid off nicely. Being able to scrape Vista off boxes and put a proper OS on there was a revelation, as modern machines no longer seemed arthritic and stressed afterwards.
Still though, what is with the lack of a quick "another LOLcat came out for Macs, they still work pretty well but aren't perfect- like everything else" section? Silly rabbit!
Being slightly mean one could suggest that there have been almost no advances in operating systems since Multics. Certainly no new ideas. Indeed in many ways modern OS designs are simply filling in the gaps in their functionality to finaly get up to where Multics was in the late 60's. Not entirely fair, but not too short of the truth either.
In research the biggest impediment to working on advancing operating systems has been Linux. There is feeling that there is really little point. Anything one does will simply be ignored. Linux is a very old kernel design at heart, with its roots in the 70's. Not that Windows (aka NT, aka son of VMS) is all that more advanced. Longhorn was really just trying to put back the features missing out of VMS. (OK, not really fair either, but the ideas certainly came from that long ago.)
The continued dominance of x86 based hosts makes almost any interesting advance impossible. It too stultifies advances by esentially killing off research into architectures that might support new OS designs. When you look back to the 60's and 70's it is astounding how barren the modern landscape is. The accidents of history that got us here are well known. It is such a pity that the current state of the market still makes it almost impossible to undo them.
Looks like a big idea to me:
"VMware's flagship product, VMware Infrastructure, coupled with VMware's comprehensive roadmap of groundbreaking new products provide a virtual datacenter OS for IT environments of all sizes."
Just no battle I suppose... other than making it a reality...
It's not that people don't want change. It's that they want to be able to schedule the change on their terms, which means that the old product has to stay available and maintained until it has gone from mainstream to extreme legacy minority interest on the customers' timetables. Sometimes, the new "improved" product will be judged by the market to be a lemon, and then its vendor should eat humble pie, stick with selling and maintaining the old popular product, and try again.
This is what Microsoft won't understand, and one reason why Vista is so un-loved (quite apart from it being slower and smellier than sewage sludge). They don't allow us (Joe Public and Joe's small business) to continue to buy XP. You buy a new computer because an old one breaks down beyond repair, and you have to migrate to Vista, then and there, however inconvenient and expensive that might be. You can't even go out and pay extra for another retail copy of XP to replace the one locked into your dead PC - they have stopped selling it.
This is a company telling you that if you won't buy what they tell you to buy, they don't want your money or your custom! Faced with arrogance on this scale, what to do? The obvious answers are Apple Macs and Linux. If one has to go through an untimely unplanned migration hell anyway, why stick with the company that forced it on you, and that you have to assume will do it to you again if they get away with it this time?
In passing, might the hardware product of the year be the EEE PC? Still better than any of the horde of imitators it has spawned. But the real point is that it has shown up Vista for what it is, and has forced Microsoft to go back to selling XP (so far only on SCCs) for fear that people might make friends with Linux and start asking for the same on their corporate desk-tops.
Would you accept "UNIX with the greatest consumer market share as measured by web page impressions"? I would have thought it was at least relevant in the Solaris part of the article, given how much of Sun's open technology Apple are cribbing nowadays - DTrace, ZFS, etc.
That said, here's a cheer for fewer mentions of the Mac on The Register!
i think what is getting lost in so much of the visualization hype is the small shop w/ 3 or less servers. They are not big enough for visualization. And being that is where all the development money is going there isn't any revolutions that they could really use. I think their is massive potential in file/print sharing that hasn't been explored on a small business aspect. It may include things larger shops take for granted but smaller ones can't afford.
Small shop not big enough for virtualisation? Hardly. VMware workstation (for creating VMs) is quite cheap (much cheaper than buying another PC) and the entry-level VMware server is free. This is all you need to get started. For example, you can then safely do development work in a VM in the same box that is doing production work in different VMs. There's also the advantage that you can carry on running (say) Windows NT for ever in a VM, whereas it's really hard today to buy new hardware that can run NT directly.
And although dynamically moving VMs neds big-iron storage solutions or some considerable Linux savvy, don't knock the cheap approach. Server failed? Unplug its hot-swap disks, plug them into the backup server, boot, and all the Windows VMs come up with their Linux host in the backup box. Takes under an hour. Linux is extremely tolerant of sudden hardware changes, unlike Windows, which is locked onto a particular motherboard by design. Windows in the VMs won't notice any change.
Backup during 24x7 operation is yet another advantage. Suspend the VM, snapshot the host filesystem (Linux LVM operation), resume the VM maybe ten seconds later, copy from the snapshot to backup media at leisure, and remove the snapshot. Simple, and quite easy to automate so it happens lights-out at 3am. Price of the software, precisely zero.
OS changes are driven by hardware changes. I hope to read something about SSDs with a next article and how they enter into this. They differ from HDDs and will need specific OS support to take off. What I am missing is how 3D graphics are working their way into the OSes. It may not be noteworthy but I cannot imagine them playing a side role in the future.
BSD is Unix, not Unix-Like. It's based on the Unix kernel (Berkley Software Distribution or sometimes called Berkley Unix), unlike Linux which is not based on the Unix kernel. Of course Linux borrowed most of the Unix way of doing things so it is very much "Unix-Like". As in "It want's to be LIKE Unix when it grows up."
Probably the two best OSes that were completely different, and yet worked similarly.
"Oh Ess Ecks ?"
I think you'll find that that is pronounced "Oh Ess Ten" ...
"Of course Linux borrowed most of the Unix way of doing things so it is very much "Unix-Like". As in "It want's to be LIKE Unix when it grows up.""
Linux is quite grown up, and works better than any OS that was ever allowed to be labeled "UNIX" ... and I think you'll find that it only aims to be POSIX complient, if you check.
@Chris iverson : nope, OSX is fully certified UNIX, as of 10.5:
@Charlie stross: with Apple developing technologies like Open CL & Grand Central, 10.6 introducing full ZFS support, FSEvent enabling powerful features like Time Machine - I think you're mistaken in not seeing the innovation that's taking place in OS X.
The omission of OS X from this article is pretty glaring. Millions of people will be unwrapping christmas presents this year that are OS X powered (iPod Touches, as well as iPhones and Macs). OS X might not be a big hitter in the enterprise space - yet - but that's the enterprises' loss really.
As mentioned above, large IBM iron has very stable and working operating systems which one can replace the hardware over the weekend and having users none the wiser. The Microsoft problem (Vista) has happened because XP has been around for so long. Since it (for whatever it is worth) is reasonably stable (or has become so) over a few years (released late 2001) people have become accustomed to it being there, and lacking a replacement for over 5 years, expect it to always be there. Microsoft's fault was not having new OS releases like it did before (95, 98, 98SE, Me, etc.) in around 2.5 year intervals. We all expect XP to be around and workable because it has been. Vista didn't have traction because it was late, VERY late.
I will note that a similar problem exists for Debian Linux since it doesn't have releases often enough. The problem is to get something out the door, and get it out quickly. In high-tech terms, 2 years is a LOOONG tine. Those who fail to understand this will fail. A bit of competition helps this along, and those companies that have some seem to be well regarded. Those that don't .....