Forum Moderators: bakedjake
[infoworld.com...]
From litterbugs in NYC to:
MS busted with fake Apple "switch" ads [aroundcny.com]
The credibility is always striking.
GUI doesn't mean easy, saying that is the equivalent of saying MS technicians are a bunch of monkeys. Could any idiot be an MCSE?
Windows 2000 Server
(With 10 Client Access Licenses) = $1199 US
Linux server license = $0
IIS = $670 - $800
Apache = $0
windows virus protection = $39.99 per computer
Office XP = $469.99
Open Office = $0
The response I got back from CEOs and CIOs is "BARF".
Senior executives no longer believe the IDC/Gartner/Mega/Giga/etc. research and think tank organizations. Everything is with a boulder of salt. Double and tripple checking is not unusual when using such documents. Several of the members indicated that they got burned in board rooms because they relied too much on a single research firm, and weeks later there was an other research firm with opposing views, or "sponsorship" of the research was too blatant...
The IDC report suggest that the reason Microsoft OS is cheaper because the administrative cost, and the equipment is more expensive over time for Linux.
The synapsis of my members' responses to the IDC document:
* salary.com national median pay - NT Admin $76,319, Unix Admin $67,781.
* A Linux box needs rebooting when you change to kernel(unless you run a cluster, then you never have to reboot). An NT box needs rebooting once a month, just to resolve locked processes and memory leaks.
* A Unix Administrator spends most of her time developing new automation solutions. An NT administrator spends most of the time trying to keep the server up and running
* Any NT hardware can run Linux hardware without modification or special hardware equipment, faster and more efficiently.
In essence - The IDC report is a farce.
As a consultant, I am always wary of writing off studies that don't agree with my opinion, especially if I haven't really disected the whole study.
I think you are missing the strong pro-linux point of this report. I see the study as saying that the Total Cost of Ownership of Linux is reasonably close to windows. Why anyone should doubt that is beyond me. Linux is just a unix variant and it would be very difficult to convince anyone that unix servers are not competative to servers with windows.
I suspect the major difference is the amount of quiet comfort you get with windows out of the box. As the market leader, windows drivers are easily available and the major manufactures of hardware work very hard to make sure their hardware works well in the Windows world.
This hasn't always been true for linux. So the big story that I read is that even with the extra effort needed to run linux in production, the TCO is reasonalby close. (And dropping I suspect.)
In the marketplace, most of the effort I see is not in switching to or from unix servers. It is in consolidating unix into a single variant. I think linux will end up as the unix of choice.
IT departments have years and years of legacy system to support with people trained in windows or unix. You don't just fire everyone and dump your old systems and start over again.
When new applications emerge, there is a chance to be put a new OS into production. So innovation is the key to server switching. Can open source develop new innovative applications is the key for linux taking marketshare from windows. This study does not address that issue.
Great :)
So windozish.
There is no way for you to prove that a study sponsored by Microsoft is unbiased.
"Banana growers sponsored research proves bananas are a life saving necessity".
"Milk, it does a body good" sponsored by the Milk producers.
"Research finds that using keyboards repetitevly will cause permanent damage to hands" sponsored by Trial Lawers of America.
"New development research shows that buying cars from manufacturers would confuse and be financially damaging to general public", sponsored by US Car Dealers Association.
You get my point. It is highly possible that the IDC report (which I HAVE "really disected") contains some truths, but when you spit in a glass of good wine, it is spoiled. Irrelevant how good the wine is... and the first thing I notice is the foaming loogy on the top, not the brilliant bouqet!
I think that this is a poor argument. You disected the IDC study but didn't give a single example of untruthfulness or bias in the study.
That makes me believe there is no bias in IDC study but would be interested to know what lead you to your conclusion that it was biased. IDC is a well-respected company and I along with others, value their reports.
My people are all well trained and know how to do their jobs. We act pro-actively to problems and stay up-to-patch, and train continually.
The result: NO downtime. We never crash. Never bluescreen. Admin overhead is minimal, as we use the proper tools to test and roll-out changes as necessary.
The proof: your typical IT shop is considered doing "very well" if it can maintain a budget ratio of .75 to 1% of the gross (in other words, a company which makes a billion a year should spend between 7.5 and 10 million on IT). Our ratio is less than .4% - and unlike 90% of the big spending IT departments, we have a fully online, ready-to-go hot-site for disasters which completely duplicates our server room.
Complaints from users? Other than application issues, virtually non-existent. Complaints about the OS, network, the office suite, and so on simply are unheard of. Complaints about our applications (SAP for example) extremely low. And that is maintained even though we are in the middle of our Windows 2000 and Xp ugrades.
Security issues? I demand s full outside security audit every year (including penetration tests). we are one of the few IT departments that seems tbe actually confident of our security! Virus out breaks? None - because we understand viruses and how to keep them out.
My experience with Unix and Linux? If properly managed, these too are fine operating systems. If not properly managed, they are terrible operating systems. That's simply true of anything related to IT.
If you just "install it" and hope it runs without proper study, training, understanding and management, well, guess what, no OS is going to stand up long to that. the scary part, is MOST (90%+) IT shops do not understand what they are doing. As evidence - many systems are still not patched against NIMDA, and many DNS's are still running the old, full-of-holes BIND, and may sendmails are still ancient, open-relay spam dreams.
The main difference (security wise) between Apache and IIS is: in IIS you have to turn everything OFF and in Apache you have to turn everything ON. Thus, if you do nothing IIS not secure, and by default Apache is secure, simply by default.
Oh yes, and it's not easy to get an MCSE these days. It requires much study, much understanding and hard work. I'm getting very tired of hearing IT wanna-be's who don't understand claiming that MCSE's are easy to get. They are not and someone who has an MCSE in Windows 2000 does have a basic understanding of a very complex networking environment. These people (assuming they did not just memeorize brain-dumps) are to be respected. Period.
Okay, that's the end of my rant for today.
Richard Lowe
These people (assuming they did not just memeorize brain-dumps) are to be respected. Period.
Everyone I've ever encountered who claimed any MS certificate + experience as their sole qualifications was an unmitigated disaster for their IT department. Not one of them has been able to do more than read the canned solutions from Tech Net. They don't know how things work and show no sign that they'd know how to troubleshoot even if they did know how thing worked.
Mind you, I've met perfectly competent, skilled people who hold MS certs and have their primary skill set in MS software. They usually also have degrees and always have a theoretical understanding that evidently isn't covered in MS cert exams. I assume Richlowe is such a person.
I also wish to note that I'm generally skeptical of people who put too much emphasis on their "certificates" of any kind, be they MS or other. (This doesn't keep me from wondering if maybe I should get a few, since people in hiring positions don't seem to give a **** about my BA.)
I'm curious how much time your group spends adding patches and service packs.
We've taken advantage of the System Update Server technology, so updates and patches are not time consuming at all. Generally, a couple of days total per month is all that's needed.
Everyone I've ever encountered who claimed any MS certificate + experience as their sole qualifications was an unmitigated disaster for their IT department. Not one of them has been able to do more than read the canned solutions from Tech Net. They don't know how things work and show no sign that they'd know how to troubleshoot even if they did know how thing worked.Mind you, I've met perfectly competent, skilled people who hold MS certs and have their primary skill set in MS software. They usually also have degrees and always have a theoretical understanding that evidently isn't covered in MS cert exams. I assume Richlowe is such a person.
Yep. 25 years of experience. Started on RSTS/E on the PDP-11's, moved to VMS, RSX, RT11, Mac OS, Linux, Unix and finally Windows (very reluctantly). Lots of courses to stay current, practical experience and so on.
MS certificate + experience as their sole qualifications was an unmitigated disaster for their IT department
Agreed. But that's true of any certification or even schooling in general. Much more is needed, esp a desire to stay current and a willingness to understand and some intelligence.
Richard Lowe
This sounds like an invitation to flame. I expect you really mean just book knowledge and no practical experience.
Don't we all start that way, though?
I have many long years of experience with computers. I think by and large the people with unix (including Linux) and/or windows experience are bright and talented and I've found few people in my life that are unmitigated disasters.
This sounds like an invitation to flame
That came out a bit more inflamatory than I intended. However, no, I don't just mean book knowledge without practical experience. I mean just certification exam answers without any understanding deeper than the configuration options of particular apps and operating systems. This could be just as bad with similarly flawed education in any system, I just tend to see practical examples of it with people who have been trained to work on M$ stuff.
In general, I think that a deeper understanding of the way things work comes from either (1) good formal education and continued study or (2) Nearly as much industry experience as I have life experience, plus the continued study.
MS network (150+- computers) doing the usual office things, mail, file sharing, print server etc. It is definitely a full time job.
I knew a guy who did that. When I met him, he was a gumpy, pessimistic stormcloud hated by every user on that network. He was probably working at least double time, and not keeping up with things at that. Several new hires later, he was like a new man, because there was enough manpower to keep the place going, finally.
"FWIW, a good fiend of mine in in charge of a medium scale MS network (150+- computers) doing the usual office things, mail, file sharing, print server etc. It is definitely a full time job. Listening to him talk is proof to me that a MS solution is not simple or easy in practice."
150 computers running Windows would be mostly desktops. I've never seen 150 computers running Linux desktops. I doubt that many people have.
I do believe that your friend was having some difficulty, but I can't see how this translates into a valid comparison with Linux. Why I should believe that Linux desktop applications would require any less handholding?
Because the process of using Linux educates the user. Linux users become smarter and more efficient than they would in an operating system that tries to keep blinders on it's users.
MS administration promotes a mentality that the end user is to stupid to understand the underlying process. Keeping the workers in the dark is more efficient only in the short term.
FWIW, a good fiend of mine in in charge of a medium scale MS network (150+- computers) doing the usual office things, mail, file sharing, print server etc. It is definitely a full time job. Listening to him talk is proof to me that a MS solution is not simple or easy in practice."
System managers and administrators that do not do their jobs or do not understand what to do have problems, regardless of operating system or application. Those that do understand what they are doing, and actually do their jobs, don't. It's very simple, really.
150 computers running Windows would be mostly desktops. I've never seen 150 computers running Linux desktops. I doubt that many people have.
That's because Linux is not a very good desktop OS. It is a great web server.
Because the process of using Linux educates the user. Linux users become smarter and more efficient than they would in an operating system that tries to keep blinders on it's users.
Completely and utterly false. Besides, even if it was true, who cares? The user's job is to perform business tasks, not play around or learn about the OS.
MS administration promotes a mentality that the end user is to stupid to understand the underlying process.
Also completely false. Our experiemnce with windows XP, 2000 and NT is it allows the user to perform those business tasks that they must get done to do their job and not worry about the OS.
Each and very OS has it's strengths and weaknesses. It's better to objectively look at those, based upon lab testing done by oneself, ones peers (or those one supervises) or trusted outsiders. The press (which tend to be ignorant and stupid and very easy to deceive) is NOT the place to look for advice.
Linux is a great server OS. Windows Xp is a great workstation OS. Unix is a great application server OS. OpenVMS is probably the best of the non-mainframe OSes. Personally, I like the Mac OS the best for my desktop, but Apple has pretty much worked themselves out of serious consideration for any real business use.
Richard Lowe
I have managed a network of 40 NT workstations + 4 NT servers for 4 years. i am studying for the MCSE W 2000 at the moment. My first Exam is just after christmas.
My question: is it actally likley that someone could pass the MCSE exms with just the book knowledge alone - as seems to have been suggested?
The material i have covered wouldnt have been fully understood without the experience i have had with NT systems. M$ also recommend 2 years experience before attempting the MCSE track.
is it actally likley that someone could pass the MCSE exms with just the book knowledge alone
I've heard that some people hit the braindumps and memorize the answers. Personally, that's kind of misunderstanding the purpose of the tests - to prove that you know.
I believe experience, courses, tests and lab work are essential for understanding anything.
Richard Lowe
I think that this is a poor argument. You disected the IDC study but didn't give a single example of untruthfulness or bias in the study.
Please refer to my previous post which clearly indicates their inability or unwillingness to look up at least compensation.
No ping pong please. You have not proved that the IDC report is not biased either.
IDC is a well-respected company and I along with others, value their reports.
Respect has nothing to do with biased or unbiased.
I stand by my statement that this research paper is a farce, and makes fools out of senior IT executives. However respected IDC is in your eyes, this report is biased. The reason I conclude this because even the most basic research to back up the document was botched. I venture to say they are intelligent enough at IDC to do proper background validation -- which can leave me only one conclusion, that it was intentional.
The problem seems to be the bias of the people reading the report. Dismissing the entire report out of hand seems a bit hasty as does accepting the report verbatim.
I'm quite happy to read an entire report full of bias if I can find one kernel of truth in it. Time for me to go gather some more info, I know someone that really wants to know if the TCO of Windows is less than Linux. That report was just ONE piece of information in the entire data gathering process that I started one month ago. Maybe in another 6 months I'll have some solid data. The folks at the local uni were of some help but the main guy there is a huge Windows advocate. I'm waiting on some more data from him, but he agrees wholeheartedly with the report. ;)
You disected the IDC study but didn't give a single example of untruthfulness or bias in the study.
KM wrote that Linux is poor at SMP; SGI will sell you an Itanium2-based 64-CPU single-image Linux box with breathtaking throughput.
KM wrote that Linux is only useful as an appliance, then the very next point said that it was no good as a compute farm (the ideal appliance); a Linux cluster is the 5th fastest computer in the world, QED
Four lies and a self-contradiction so far, smacks of Alexis de Tocqueville Institute. No doubt Alex is spinning in his grave. Need I continue?
If someone would be so kind as to point me to the IDC study proper, I'll shred that too. They botched their study for RedHat by (amongst other things) scaling small shops to match large shops, and quoting admin salaries which are quite different to the national averages, yet still came up with an answer RedHat liked.
In eWeek Jim Rapoza did suggest that the accounting base of the study should be switched to 3 years rather than five. This would turn some of the numbers more favorable to Linux.
The author's mainpoint is that it is impossible for anyone to produce an unbiased study. This is an interesting observations since eWeek produces the eLab studies.
You may read "Is An Unbiased Study Possible?" by Jim Rapoza at [eweek.com...]
When doing a study for a client, bias can be dealt with. You make a long list of the characteristics of each product being compared. You get the vendors to agree that the list is complete for his product. You work with the client to assign a relative value to each characteristic’s importance from 1.00 to .00.
Next you work with your client to rate each characteristic for each product on a scale of 1 to 10. Then you do the arithmetic and a subjective but unbiased answer emerges.
I expect that the Linux v Microsoft study used some sort of technique like this. It is fair to disagree with the values the author assigns but wrong to call things you disagree with a lie. He may very well be mistaken but in fairness so might you.
Since the Linux press is not pointing out mistakes, I think the author did a creditable job.
