Posted on 12/03/2002 9:40:52 AM PST by chilepepper
Windows 2000 servers are cheaper to run than Linux ones, sometimes, says an IDC study which was by strange coincidence sponsored by Microsoft. The study will come as welcome relief to Microsoft salespeople who have been parroting the 'cheaper than Linux' line to general disbelief, but whether anyone else will believe it is another matter.
Nor indeed should we take the study at anything like face value. IDC set itself the task of measuring total cost of ownership of the two server operating systems over a five year period, segmenting this into five areas of server operation. Over a period of years one would naturally expect differences in the purchase price of software and hardware to account for a declining proportion of TCO, with support and staffing costs accounting for an increasing proportion. And lo! This indeed is what IDC found.
However, erm, correct us if we're wrong but we were under the impression that obstinate corporate customers who hung onto their Windows servers for a whole five years without upgrading were more or less open source loving commies in the eyes of Redmond. We haven't as yet seen the full study, but suspect software upgrade costs, and the associated cost of new hardware during the period, may not have been entirely factored in to the Windows 2000 server tab. Linux can have similar upgrade cycles if you want it to, but there is really no similar hardware upgrade imperative if you do decide to move up a version.
Aside from that, the study looks maybe a little stacked in other areas. It finds the support and staffing costs for Linux are greater, largely because Linux systems are more difficult to configure, manage and support than Windows ones, this itself being because Windows has more mature, easy to use management tools.
Which from a certain perspective, i.e. a Windows network manager's perspective, is true.Large-scale, properly set up Windows networks with a ton of hardware and GUI management tools all over the shop needn't cost a lot in terms of machine minders, whereas an open source network without these tools will need the requisite number of skilled geeks making incantations over bash prompts. But this is comparing apples and pears, the geeks will serve you better than the deskilled machine minders when something goes badly wrong (which it will). In any event we doubt the smooth-running easily-managed Windows network actually exists anywhere outside of slideware.
The differences in cost IDC identifies are relatively small, and vary depending on the tasks involved. For example, supporting 100 users on a networking server would cost $13,263 for Linux, and $11,787 for Windows; obviously, the 'difference' here could easily be wiped out by a Windows server upgrade, or by the network in question being run by a company with a background in the Unix, rather than the Windows, space.
Windows also comes out better, according to IDC, in file, print and security. The first two are scarcely surprising, given that a chimp can drive them under Windows while under Linux you need a slight understanding of what you're doing, but it's not rocket science, and would be even less so if Microsoft were a little more helpful to the Samba team. And the third, security? IDC seems to be having a crack at conceiving the inconceivable, and we'd just love more detailed evidence.
There is however one area where the study reveals just the teensiest problem for Microsoft's sales people - Linux it finds (confirming the general received wisdom) is cheaper as a web server. Now, given that plugging computers together on a LAN, sharing files and printing is the stone age trivial stuff, while web serving is more in the 'next big thing for businesses' department, do we not foresee an impending catastrophe, given which it is that Windows is allegedly good at? If Microsoft believes this stuff at all, it should surely be deeply worried by this particular bottom line of the study.
For the record, IDC doesn't identify a particularly large gulf between cost here, it's less than 10 per cent. But as a corrective we offer a counter-study prepared earlier this year for IBM by the Robert Francis Group. This put the total cost of a Linux system over three years at less than half that of a Windows equivalent and, significantly, noted "some initial costs [for Linux] were higher at points."
Windows (also) = job security for computer geeks
Windows vs Linux = endless holy wars for people who have too much time
Network administration isn't cheap. With Linux, you need a geek to keep the network intact. With Windows, you have to pay through the nose for forced upgrades every couple of years (I won't even get into the security issue with MS networks). There was a time when doing business with MS (at least from the consumer side of things) was an ok experience. Those days are just about over. I'd rather deal with a geek who acts cocky because he knows Linux inside and out than deal with a geek who believes that it's his destiny to control every facet of computing.
The problem is the Microsoft generation that follows these severs. It is a new operating system, rewritten from scratch all in the object oriented language of C#.
UNIX, BSD-UNIX, LINUX, and Windows 200 are all bound by the old programming models created by Kerrigan, Ritchee, and Wirth over 30 years ago. Structured function based operating systems are not going to hack it in the oject oriented world. And C++ is not going to hack it either.
If we compare operating systems to air planes all of todays operating systems are prop planes. UNIX is by far the better plane. But MS is working on a Jet plane... and the free software people are improving the world's best prop plane. If UNIX does not address the programming model and the programming tools, they will be stuck in an old outdated technology.
Look at the last 20 years. UNIX was a fantasic operating system and DOS was a tinker toy. UNIX was a fantasic operating system and Windows 3.1 was complex and buggy operating system. UNIX was a fantasic operating system and Windows NT was a sorta good operating sytem. UNIX was a fantasic operating system and Windows 2000 is its equal except for cost. NOTE the debate is not over better, it is over which costs less. The "mines better than yours" arguments are subjective arguments. We are no at who costs less.
See a pattern there? Every few years MS gets a little better and closes more of the gap. It now for all practical purposes closed. At one time the new developements like UNIX were in non commercial hands .... Bell labs and Berkley. The information was public domain. The open source guys need to get with the new program... get with the next generation. The open source guys are behind the development curve, largely because Gates' billions have bought the best operating system talent in the world.
The problem for UNIX is were do they get the super talent required to create the millions of lines of code for the next generation. They need to get it for free or they will have to charge a Microsoft price. The problem is every time some guy shows up with the talent and starts doing stuff for free, Gates has someone contact the talent. Microsoft asks how would you like to be paid millions for doing what you are now doing for free. It is amazing how many prefer being paid millions rather than doing it for free.
That is the Microsoft advantage and there is no way free ware can compete.
However, I also believe that Micro$oft licensing is becoming such a royal pain that the door is left open for alternatives, and it is CERTAINLY the case that Micro$oft did not invent software development tools... these have been around for a long, long time.
A number of these environments (third party of course) already exist which run on BOTH Micro$oft and Linux platforms. This presents a very real danger to Micro$oft, their idea to present developers with a development cocoon means that they will be able to switch over to the Linux version of the cocoon very easily -- so unless the company is really masochistic or there is something else that the M$ version has to offer, then they would be just as happy running this stuff from Linux, PARTICULARY if the development environment can CROSS develop, which is quite likely...(National's LabView is an immediate example that comes to mind)
as for people moving on to real paying jobs, sure many do, many many do open source as a labor of love, or for professional prestige and many already have daytime jobs. in some cases, the open source project they are working on is sanctioned and even partially funded by their employer (in particular certain government/DoD software shops: much of the cool stuff w/ BSD came from there, as did Mosaic of course - numerous examples abound)
Did you read the article? It was about "PC" servers.
"PC" microsoft servers vs. "PC" Linux servers to be specific.
And OBTW even when you pay for the study, you don't get the results you asked for.
Baruch HaShem Adonai Yeshua HaMashiach
Praise the Holy Name of the L-rd Jesus the Christ
chuck <truth@YeshuaHaMashiach>
NOT!
Try to write a 500,000 line program using the UNIX tools and then try it with Visual Studio.net. The cost of writing and debugging is millions as opposed to thousands. MS tools are orders of magnitudes better than anything for UNIX. I know something about operating systems and tools I wrote part of IBM's PC DOS 6.0 and portions Borland Delphi 6 and 7. I also wrote some small portions of LINUX. UNIX tools for serious programming (C and C++) suck compared to windows tools. Using Visual Basic and Visual C++ on a project can turn out a finished program in months that would take years in LINUX.
Henry Ford had a monopoly on cars for several years. In 1915 over 90 percent of the cars that existed in the world were Fords. That is when Genearl Motors was founded by Durant. He named his first car after a famous race driver named Chevrolet. The Chevie was very much like the Ford. If you could drive a Ford you could drive a Chevie. The ability to run a Windows 2000 server does not mean you can run a LINUX or BSD UNIX Server or even a desk top for that matter.
To take down Microsoft the competitor will have to do what GM did. Build an operating system that a windows user can run with zero training. Make it totally compatable and huge numbers of companies would buy some just so Bill Gates would not have them in his hip pocket. It would have to work so anyone who can run Windows can run this new operating system with no training. And it would have to cost a lot less. People will not pay more for even equal for a clone. It would have to cost less.
If I were younger I would do a distribution of LINUX that came with Samba installed and configured for a windows 2000/XP/ 98 network. It would configure just like a Windows 2000 sever. I would make it so if you can set up a windows small business server, you could set up the LINUX server. That would do Gates some damage.
The Gates licenseing plan has all the intelligence of Word Perfect and Lotus 123 copy protection. The problem for all software is that we are either at a plateau or it has matured to a much slower growth pattern. If it is at a plateau it may be a long time before those windows 2000 severs are updated. I have the Dot Net beta. I don't see much that is special about it. I see no reason to upgrade.
Here in Columbus Ohio there are 5 stores selling used Computers, including the Microcenter Super Store. When they can sell 3 or 4 year old machines that surf the net, do word processing and spreadsheets just fine that tells you lots of companies that are not leasing hardware are not going to upgrade. The majority of desktops may for many uses become an item like the desk it sits on. It will be replaced when it breaks or wears out. Micorsoft would like to get people on annual payment lease deals before customers figure it out.
It appears MS is trying to force all customers to an expensive lease deal. A company that does a windows clone and sels an old style lease, will eat Gates clock as surely as Chevie ate Ford.
But the KEY is a user interface that requires no training. And a LINUX server that any boob can set up and manage in MS desktop situation.
Gates will be taken down... but by a clone or a system that appears to be a clone... not an alternative operating system.
That was a huge problem for me. I did some LINUX stuff. Another developer after final tests on my suff commented out some of my code to test his code that depended on mine in some manner. It created huge problem that got on 200,000 distribution CD's. Commercial organiztions would not allow that to happen.
That is when I said take your open source and shove it. I wouldn't put up with that for money... let alone for free.
Exacly, people are not going to do that. I run windows 2000. Both servers and desktops. I do not plan to upgrade. I urge my customers not to upgrade too. There are almost no benefits to XP and way too much hassle.
What is needed is a windows clone. You could build it on top of the LINUX or one of the Berkley UNIX kernels. But it needs Drives C D E etc and the same slash for directories. It needs to work just like windows. It needs to use a windows compatable file system too. The KDE user interface is fine if the set up stuff mimiced windows. Setting IP addresses, maping remote directories and drives, and printer set up needs to be very much like windows. The key to acceptancs is a system people can use without aditional training. It needs to handle files just like windows and it needs to be an operating system that can serve or be a desktop. It must coexist and be very similiar to windows in setup and administration. It has to be as close to windows as a Ford is to a Chevie.
How many million copies whould you have to sell for 39 bucks to be rich?
If any Windows IT guy could set it up with zero help or instruction you could eat MS for breakfast and feed Balmer to your pet aligator for lunch.
Exacly, people are not going to do that. I run windows 2000. Both servers and desktops. I do not plan to upgrade. I urge my customers not to upgrade too. There are almost no benefits to XP and way too much hassle.
What is needed is a windows clone. You could build it on top of the LINUX or one of the Berkley UNIX kernels. But it needs Drives C D E etc and the same slash for directories. It needs to work just like windows. It needs to use a windows compatable file system too. The KDE user interface is fine if the set up stuff mimiced windows. Setting IP addresses, maping remote directories and drives, and printer set up needs to be very much like windows. The key to acceptancs is a system people can use without aditional training. It needs to handle files just like windows and it needs to be an operating system that can serve or be a desktop. It must coexist and be very similiar to windows in setup and administration. It has to be as close to windows as a Ford is to a Chevie.
How many million copies whould you have to sell for 39 bucks to be rich?
If any Windows IT guy could set it up with zero help or instruction you could eat MS for breakfast and feed Balmer to your pet aligator for lunch.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.