Businesses Slow to Adopt Linux

milenko81 and several others submitted this CNET story about corporate spending on information technology. The reporter seems to interpret it negatively because Fortune 1000 companies aren't dumping Microsoft 100% and going for Linux. But interpret it as you will.

Comments

Not commercial = bad?

[RagManX]

I'm having this problem with my site. They don't want to go Linux, because they can't blame someone if there is a problem since it isn't a commercial application. I don't understand this, because a)it isn't true (that's why you buy a distro and support package) and b)Microsoft sells commercial applications that have tons of problems, but their license agreement is such that you can't blame them if something goes wrong.

This brilliant company thinking has even extended to the security tools I use here. I can't use freely downloadable tools because someone might have trojaned them. Only if we can pay someone to ship us an install CD can we use it, because if we pay for it, *WE KNOW IT IS SECURE* or something like that. I mean, come on, Microsoft NT 4.0 is super secure, because we paid lots, right?

Let's just face it, techies don't run things, and non-techies are mostly idiots when dealing with this kind of stuff.

RagManX

Re:Not commercial = bad?

[ChazeFroy]

It takes X amount of years to integrate a system into your infrastructure. It also takes X amount of years to remove a system from your infrastructure.

Say a company has spent 5 years integrating NT systems into their department. That usually means it will take another 5 years to get rid of it.

Linux will not be an overnight success...it will take time to supplant Microsoft.

Re:Not commercial = bad?

[Bobzibub]

Ask them this: "If all the resources of the US DOJ cannot beat Microsoft in court, how can our company's legal council possibly expect to win should they sell us faulty software?"

Re:Not commercial = bad?

[pmc]

Say a company has spent 5 years integrating NT systems into their department. That usually means it will take another 5 years to get rid of it.

Mod up the parent, somebody, because this is the crux of the matter.

Where I work there are approx 4000 staff in about 40 offices worldwide. Office numbers range from ~10 to ~1000. We have about 250 production servers all running NT (or 2000). The cost per head for MS is about $20 per person per year in support (this is the money we pay MS, not how much support costs), and about $400 per person one off payment for licencing (OS + office + CALS) - this does us for any related product XP (office, workstation and - I think - server).

Our goal is to cut that spend. But the licence money is dead money - we don't get a refund if we don't use the products. We also don't get a reduction in direct MS support costs if we have, say, 100 of the 250 servers running Linux.

Where we can make a start is nibbling round the edges - for example proxy server is not included, so we can use Linux and Squid. Except (SFAIK) Squid does not integrate with things like Websense which we need to block sites (nothing draconian - mainly web e-mail to stop viruses and web porn to stop lawsuits).

And we could use Apache, but that is a direct increase in support costs. IIS is free (or rather, we're paying for it whether we use it or not, and we'd have to pay more to use something else) however crap it is - and as our servers don't accept anonymous connections the Code Reds had little effect (so no convenient lever there either).

And then we come to the real killer - our business systems (which are extremely good - easily the best of any company I've worked for) are built about MS products - IIS, index server, Exchange, MS SQL held together with NT authentication. It would (will) be a huge undertaking to move these to anything else.

There are chinks of light in this - the MS product line is changing, and a lot of the code is being rewritten anyway, so rewritting in the direction of vanilla or standards will help now (so redoing the MSSQL app to use only vanilla SQL will take an important step towards portability -it may not perform as well, but hell - this is a time when hardware can and should be thrown at a problem).

Another killer is Exchange - there is nothing I can go to management and say "We should look to replace Exchange with XXXXXX" (and, before there is a flurry of "sendmails", Exchange is not just e-mail - in fact think of exchange as an enterprise PDA that also does e-mail).

Anyway, a few thoughts about something that has been exercising my mind a lot recently - any other ideas to get out of the MS lock-in?

Strange negative spin

[Ami Ganguli]

You have to wonder about the negative spin on this. 24% of the largest, most conservative companies in the country are planning to use Linux next year. This is good news. Considering that two years ago nobody believed Linux could make it into enterprise datacenters, this is amazing progress.

When interpreting these figures you need to remember that:

• Large companies move very slowly. Some of these people are still using OS/2.
• Many of these companies don't think about which OS they're using so much as which vendor they're using. In particular, many of these accounts are controlled by IBM. As IBM puts more and more energy into Linux, these accounts will (very) slowly follow.

It's the price, stupid!

[ryanwright]

Corporate execs don't understand how something that is free can be worth a damn. I know; I've tried to get Linux implemented in our enterprise as a basic web proxy. (Instead, we spend almost $10k on licenses for Microsoft software and third party filtering applications.)

Here's the deal: When you pay a cool million bucks for the software to run your enterprise, you have someone to bitch at (Microsoft) should something go horribly wrong. With Linux, the only person you can bitch at is that uber-geek you're paying $50k a year. When millions of dollars are at stake every day, you just can't trust a free piece of software.

Obviously, most of us here know this is bullshit, but it's the excuse given by every exec I've talked to. They won't trust their business to free software and a couple of geeks no matter how compelling the evidence. Even a mention of IRC as a help resource elicits manical laughter. If someone setup a high priced licensing & support system for Linux and gave it a different name, businesses might sign on. Sad but true.

One last issue: MCSEs are a dime a dozen. Any moron can administer a Windows network. I'm the only one in my group that knows enough about Linux to properly setup and maintain an enterprise server. If we implemented Linux and I left, they'd be SOL. Nobody wants to put their faith in one or two employees, especially when those employees have knowledge that is (let's face it) hard to come by. The proportion of people that can adminster a Linux server vs. those that can admin MS is huge. Probably thousands to one. It's just not easy to find a good Linux guy, let alone the 5 to 10 of them it would take to run a medium sized network.

Re:It's the price, stupid!

[SPiKe]

>One last issue: MCSEs are a dime a dozen. Any >moron can administer a Windows network

>The proportion of people that can adminster a >Linux server vs. those that can admin MS is >huge. Probably thousands to one.

A lot of the guys that can do Windows correctly are guys that do Unix correctly.

I've met the exceptions, but they are rare.

A good admin is a good admin is a good admin. All one has to do is force yourself to think outside of just one particular mindset.

Re:It's the price, stupid!

[tclark]

True story about a company where I worked:

One day in a meeting, the president of the company said he was concerned because we were using MySQL. He said something to the effect of "How good can it be if it's free?". I could barely keep from laughing at him as I thought about the free software (Linux, Apache, Perl, PHP, CVS, sendmail, etc) that our mission critical apps used. In fact, none of our critical software was proprietary.

Naturally, I didn't think it was necessary to tell him this.

Duh

[chuckw]

Of course it's low on the budgetary radar scope. They're not paying for Linux and they already have the Unix expertise in house. Since Linux runs on darn near anything, they probably already have the hardware there too.

Not surprised

[Anton Anatopopov]

Linux is a great OS, and it is a miracle that it has come this far. But to penetrate the business market takes more than good technology, it needs suits.

Most computer purchasing decisions are not made by tech-savvy developers with their finger on the pulse of modern developments. They are made by golf-playing middle management who are being bribed left right and center by their suppliers with free trips to Hawaii and other inducements.

Linux will make it in the end, but it will be because one of the pre-existing corporations or management consultancies starts pushing it, because it improves their own bottom line.

Sad but true.

This got me!

[gosand]

The very last sentence of the article: " Ultimately, technology managers don't want to hear about the operating system, Robinson believes. 'All you care about is wanting a stable, scalable platform for applications to run on.' "

And the answer to this question is not Linux because....?

Windows, Security

[JoeWalsh]

The research cited in the article purports that interest in Microsoft Windows and computer security are both strong among Fortune 1000 companies, while interest in Linux is weak.

Does anyone else find that position absurd? If folks were interested in computer security, you'd think they would have no interest at all in Windows.

Ah, well. So it goes. Just another sign that most people are idiots.

Security?

[whjwhj]

Two comments on this: 1. Although at least some companies are switching to Linux and open source software, when was the last time you heard about a company dumping their open source software for Windows? Just doesn't happen. 2. The article mentions how "security" is in the mind of IT professionals these days. Yet Linux is phenomenally more secure than Windows will ever be. Seems like a lot of IT folks need to be educated.

Re:Cost of linux administration cost of windows

[MacGabhain]

False.
The cost of adminning Windows servers is considerably higher than the cost of adminning Unix servers (of any flavor). You can cover far fewer servers with a single admin, and you need at least 3 times the number of physical boxen than you do with Unix systems because a: Windows scales horribly and only runs on hardware designed to be workstations b: Windows requires at least two redundant servers for each primary server to maintain the uptimes of any Unix c: Windows is only able to perform properly if each box only runs one particular server function. Put a print server, a web server and a file server on the same box and none of them will work well (well by Windows standards).

Linux does not always have a professional face.

[Junks Jerzey]

I sincerely do not mean this as a troll. I am both a user of Windows and Linux, and I don't lean zealously in either direction.

If nothing else, at least with Windows there's a large company with financial interest behind it all. Sure, Windows sucks in lots of ways, but at least you won't find them generally working toward what customers want.

With Linux, it's a bit scarier. Not so much with the kernel as with desktop environments and applications. With WordPerfect for Linux, I felt like I was just being used as a pawn by Corel to get a foothold in a new market, and the quality of the software was secondary. Miguel, of Gnome fame, often sounds an overly idealistic college student. It makes me stop and think "Should I really be letting this guy determine the direction of the software my company uses?" Sure, you can pick and choose different products, but with Windows you don't have to. If you go with Windows 2000 or XP and Microsoft Office (or just Word) then you don't have to worry about making the wrong choice. There's often too much personal agenda behind open source software for Linux.

We all can offer Linux solutions

[korpiq]

This is what I'm doing:

• practice nice, low-tone, clear-speech advocacy from the clients' point of view rather than technical (total costs,reliability,security)
  
• find a small/medium-size company in need of firewall/file+printer-sharing services
  
• offer a box that does it all, guaranteed, with remote administration when needed, with unbeatable price
  
• check out their needs and environment (SMB password encryption, for instance) and find technical solutions (usually someone has done it already)
  
• install debian, samba, netatalk, apache, lprng
  
• set up netfilter accompanied with squid and postfix to drop dangerous attachments/scripts if sold as a firewall to secure windowses
  
• set up a SIMPLE internal webpage for user account management (ask me)
  
• offer enhancements like RDBMS, extranet (ftp/http-download) - what can we come up with?
  
• repeat until world.domination() == TOTAL
  

Coming up next year or so:

• test out the Linux office packets, make up a desktop solution for office use
  
• promote a solution with 100MB switched LAN, diskless workstations booting from server, centrally, remotely administrated for low cost
  
• remember to spread FUD about viruses ;)
  

I'd do it more given time and customer contacts (best advertisement you can have is a happy customer talking about you to its clients.)

Share administration burden (what? doing something wrong?) with trusted friends.
Take a fair price for your work, but avoid greed.

This can and should be done as a side-job, unless you get very successful in the long term.

Only fix what's broken, security hole, or a client-requested enhancement or new service. Never say "can't do", say "I'll look into it" and go for the web; Never say "you can't afford it", say "I'd be forced to hire people for approximately $this much money, would you like to try something else instead?"

I could go on for hours, but you'll find it all out once you start thinking about it.

Make difference where you can.

Reasons why...

[truesaer]

1) customers have heard of windows/microsoft

2) large customers get benefits, real or imagined, from being a good customer of a company like microsoft

2) Bill Gates

working at a certain large company, there was a new project that the software development folks were working on planning. The business customer demanded .NET be used. This was before it was even released. At this point, the tech architect was willing to use windows, but just wanted to use regular microsoft products that had been out and that the developers were already familiar with. Finally, it looked like they were going to win and the customer would just have to go with it. Then, Bill Gates had the CIO fly out to meet with him, and within a few hours, .NET was back.

One of Linux's weak points is that in the world of big business, there aren't people that can leverage a new product like Microsoft can. I guess most people here would say thats a good thing, but it isn't helping fortune 500 companies choose to ditch microsoft.

Stop claiming that Linux is free idiots!

[alexhmit01]

Its really easy, you try to keep the execs out of the decision process. Get your requirements and develop a strategy.

Go buy the high-end Red Hat distribution. We did this for a client that knows we were using OpenBSD off a $30 CD. We also buy the nice Red Hat distribution.

Call up Red Hat and set up a support arrangement. They are a reasonable sized company. Alternatively, call IBM.

The advantage with Microsoft is that if you are a big company (say, the Fortune 1000 in the article) you get the special phone line for support, etc. Microsoft supports the Fortune 1000 a little differently than your pirated copies of Win95.

Don't sell them on free. Tell them that you worked with Red Hat's OS, and you found that it is better suited to this project. Inform them that you can reduce downtime (their real concern) instead of a couple grand on licenses.

More importantly, emphasize that it will save you time.

We don't use Linux or *BSD on desktops, it is too expensive.

Win2K or even WinXP involves a short installation procedure (before lunch) then come back and finish. Setting up a Linux desktop (for a technology guy, not end users) takes a few days of playing. Win2K tweaking with fun apps takes $1000 in software (including a $200 shareware budget) and you're good to go.

Look at your salaries. See what it costs your company to have you putzing around for days.

BTW: with the MS licenses and a point-and-click installer, how much time does it take to get another server up and running. Including your downloading Redhat over the corporate T1 (or whatever), how much of your time is spent putzing around on IRC, etc.

Sure, IRC is nice for REALLY hard problems. However, having a server down for 1-2 days while you troll USENET or IRC for help isn't acceptable.

Next time a MS solution is being proposed, try to get 24 hours to stall. Take the same list of software, and the budget, and CALL Red Hat Sales. Tell them what is going on, and ask them to put in a bid.

Alternative, call a Linux consultant, and work with them to put in a bid for the implementation AND for the Red Hat support contract. If the Red Hat fee is less, show that to management.

You all would get a LOT more credibility with management if:
A) You dress like professionals (I did NT Consulting for 4 years... we all wore a nice shirt and khakis... the Linux guys would often wear jeans, it makes a difference; my BSD shop does it too, it matters)
B) Emphasize solutions, not technology (they are looking for a solution, show that you understand this. Emphasize the savings in downtime, not licensing fees.
C) Focus on REAL cost savings. Don't CONSIDER unsupported downloaded applications. Discuss support agreements, Red Hat Network, etc.

Geeze, this isn't rocket science guys, understand what the executive is trying to accomplish.

Alex

Slow to take up?

[ColaMan]

The reason why companies are slow to take up linux? Because their trial 500GB linux data store is still working though fsck after some idiot MCSE pressed ctrl-alt-del to log in at the console!

It's happened before, and it'll happen again ...

'Gee, I've got this blank text screen thingy and it just says "Username:" ... hmmmm .... I wonder what it's for? I'll just press ctrl-alt-del to log in and ...
OH GOD NO!!! '

Re:Cost of linux administration cost of windows

[Lumpy]

True that legacy windows scales horribly.. Windows Domain model is horrible and requires a BDC at every user collection point. Granted there reallys isnt a decent model available on any side. But Domain really sucks.

Where Linux/Unix blows all microsoft products completely away is that I can set up my office with 40 sub-pc's and a linux terminal server and have one person, one non-guru person completely manage those 100 users and the server with very little effort. Update staroffice? it instantly is updated on every workstation. Lockdown? no problem, configuration changes to migrate to all users (Example new printer... try to instantly install a new laser printer on 40 Windows machines without visiting every machine or confusing users by adding it to the login script.

windows is based on very old computing ideas and models. It's not efficient and has always been in catch up mode when it comes to networking and enterprise computing (Unix was there decades ago.)

98% of the corperate drones do not need the power of a full PC, running off of application servers or a terminal server is plenty and the NT model should have been based onthat from the very beginning. (but wasn't so they could charge $300.00 per workstation for the Operating system and then charge you for server access, etc...

There's big money in ensuring that none of your users are running legally.

1 Unix admin can easily do the same work as 10 MCSE's. and that is the point... the typical MCSE is a clueless lump.. I've been in this business for over 10 years and I have met only 5 good NT admins.

Why? because very few NT admins can build a SMP server from parts, redesign the entire network, Understand TCP/IP understand SMB, understand fully NT permissions,happily hack and romp in the registry with full confidence.

and the reason there are very few of theese good NT admins? companies will not pay for them. They'll just hire another fresh MCSE to replace that uppidty admin.... MCSE's are plentiful.

unix admins? rare to find, you have to pay for them, and you cant fake being a Unix admin, it is really easy to BS your way as a NT admin.

I agree, NT can be coerced or forced to work right, I have a NT server+SQL6.5 that has ran for 12 months and only rebooted for Service packs and critical patches... it's rock solid otherwise...