Tech Perspective: IBM's Linux investment

Linux is about 10 years old now, which means that it's been around for as long as Windows NT and Netware 4.

We are increasingly using Linux in our shop, and we've proceeded as follows.

First, give the Techs a copy of Red Hat, and have them install it for use as their own server. Either for Non Pirated MP3s, or more useful storage, give them run of the box, and allow them to work together on it just as something to play with. The end result of this will be that they have some basis for using Linux itself, and how it works, as well as learning Samba. Once this is working, they will generally feel comfortable goofing around with Linux. Try to give them a bit of time to do that, even if it's just 15-20 Mins a day. Once they are comfortable with that, they will have spent a month or so supporting file and print sharing through Samba (for themselves.)

Then, when a new application comes up for you to install, start from day one with it on Linux. Or, invent a new application for you to use internally. An IT intranet, or something. Make the managers aware of the project, and give them reports regularly (bi-weekly, or whatever). Then build up to more important things. Linux or not, it doesn't make sence to replace a solution that works just because you can...

IBM's support doesn't particularly matter in the short term, because beople aren't going to go out and buy an S/390 or an AS/400 and junk all their current intel boxes just so they can move to a Linux based network. First, there needs to be a comfort factor. Management needs to see Linux as more than a geek toy, and Tech needs to see Linux as something they can work with. It's like dating. You don't get married and then ask about a person's likes and dislikes. You date, build trust, get comfortable, and THEN commit to marriage. Same with Linux. Start Small, and build up. When you need to replace your ERP package, then worr

Concerns about using Linux in Mission Critical? Basically none. Linux is at least as secure, and reliable as Windows, Solaris, Netware, etc. The concerns are more focused on finding and retaining competent administrators. Buy a large Dirtribution, Probablly Red Hat (You've mentioned IBM, and they generally work with Red Hat), Pay for support if you need it.

In the short term, AIX has a bright future. As Linux gets more popular, it does 2 things. *nix in general is more familiar to people since Linux knowledge will carry over to AIX much easier than Win2K will. Secondly, As you seem to be doing, people consitering Linux will also consiter other alternatives to Windows. As alternatives increase in frequency, they will raise expectations of Windows, which so far haven't been met. AIX allows the transition to a much more stable and scalable platform than windows, but doesn't require the 'leap of faith' that Linux seems to. AIX is familiar, and well supported. Until Red Hat(or Caldera, Storm, Suse, whatever) reach IBM's level of brand-reassurance, especially individually, but even as a combined team, IBM's AIX will be far more likely to be the solution of choice.

Kev.

Maybe I live in a different world than the Gartner folks, but I have yet to see "hype" emerge as a special risk to the adoption of Linux. Every OS has it's evangelists and cheerleaders but within the data center environment they don't seem to hold much sway.

I, and most of my colleagues, choose their platforms based on a careful analysis of capabilities, risks, benefits and total cost of acquiring, installing and supporting the package.

Linux is welcome in my data center for those applications for which it is demonstrably appropriate. So are Oracle, Solaris, MVS and Windows2000 on a similar basis. IBM's investment in Linux will undoubtedly broaden the scope of applications for which I will consider Linux (for example, to support aDB2 database), but it doesn't radically skew my assessment one way or the other.

paul

I'm afraid that Gartner has missed the central point of IBM's Linux strategy. It is classic bait-and-switch. IBM's argument will go like this: "Mr/Ms Customer, go ahead develop on your Linux servers, but when it's time to deploy, use our industrial strength AIX software platform. No changes, no conversion, but we'll give you a mission critical operating system". I think IBM has seriously harmed Linux's future by relegating Linux to a development-only platform. IBM has co-opted the Linux production platform and will now position itself as the developing platform by creating a lower cost version of AIX with an enhanced development environment.
As for IBM's top-to-bottom Linux strategy, running Linux on IBM's mainframe hardware is laughable. Has anyone bothered to find out how well(if) it runs? This isn't the first time IBM has had a Unix variant running on their mainframe hardware. Price/performance will be unbelievably high.

As an Applications Manager, I am gun shy about using Linux for any of my applications. Application Software vendors have not come out in droves behind it as a platform for their systems. Sure the Network and hard core technical folks will love it. It's another thing for them to learn and play with. But from my perspective, I'm not going to be on the bleeding edge of application software development and maintenance on a Linux platform. Too much risk, not enough benefit.

I agree with a previous post, hype is not a significant factor in the process competent shops will use to deploy Linux. Most shops will choose Linux for sound reasons, I doubt too many will make any major deployments because Linux is cool. One thingthat would help reduce the threshold of Linux acceptance is a deeper knowledge base. IBM has significant electronic learning capabilities. IBM should develop and deploy, free to anyone, a series of in-depth web-based training modules addressing Linux. These modules could begin by addressing Linux on a server and progress to the desktop if necessary. This plan would no doubt anger other companies with Linux training programs but unless and until Linux becomes more mainstream companies, who are on the fence regarging Linux, will be reluctant to spend scarce training resources on Linux.

Linux is just another unix, except that it
runs on Intel (and other) hardware and it is
free. I like it, and use it in my datacenter.
It fits in well.

I back up my Linux boxes with my enterprise
backup solution, Tivoli Storage Manager. It
gets backed up just like my Solaris boxes.

I would love to run some of my SAP R/3
applications on it, but the SAP apps that I
think are ripe for Linux, SAP hasn't quite
finished porting them to Linux yet. These
are the Internet connectvity related apps,
such as SAP's Internet Transaction Server.

When the applications are there, I will
definitely run them on Linux. That goes for
my ERP apps, as well as any other
Server-based apps. And, BTW, I run Linux on
IBM's Intel based serversand yes, the fact
that they have excellent support for Linux
has helped them win my business.

Gartner hit the nail on the head with their analysis of IBM's Linux strategy. IBM is master of the bait and switch tactic--they have never been a good bed partner... not with Novell, and not in this situation. Look for IBM to covertly hustle proprietary AIX in preference to open Linux. Probability: 1.0

This question was closed by the author