Avoiding Microsoft Lock-ins?

bayduv1n asks: "My company relies on WinNT primarily for workstations, file servers, and print servers. In my opinion, Linux based solutions are close to offering a viable alternative. I'm sure that management, however, would like to see mainstream acceptance in the corporate world before considering a migration from the current environment. With the Linux drum beating louder and louder, I would like to make recommendations that would leave our future options *open* and avoid locking into MS technology. I consider a lock-in anything that makes it difficult to migrate. One example of a current lock-in is the macros written for Excel. These would require significant effort to migrate to Staroffice Calc. Implementing MS Metadirectory Service could also be classified as a lock-in. My question is: 'What recommendations can I make now to make it easier to migrate to Linux in a couple of years?'"

Comments

Recognize cost

[AT]

The best thing to do is to make sure all decisions made regarding infrastructure recognize there is a cost. Specifically, there is a cost to going with any technology that locks in, and there is a risk going with a technology owned by and controlled by a single company.

By adding this into your decision-making process, you might not meet your personal goal of going linux, but you will make a better business decision.

Give Em a Bird in the Hand (It's Free)

[4of12]

I sympathize with your recommendation. It is absolutely rational and, in theory, should work.

In practice, however, most corporate IT decision makers live in an MS World, where even the ability to perceive lock-in is severely diminished.

There are constantly new MS Market Extending products being offered to fix holes in their existing MS all-encompassing fabric. Needless to say, these new products typically introduce further lock-ins in the guise of features. I don't have to tell you how many times I have seen products sold that have a functionality that was either:

• already in UNIX years ago, or
• not needed in UNIX because of its better design (virus scanners, anyone?)

No, I think there are 2 important strategies you should follow, in addition to emphasizing total costs.

At every opportunity, promote adherence to openly published standards for interfaces between all parts of products in your enterprise.

Ask about RFCs, about protocols for data exchange, about file formats, about interoperability with competing products, about published APIs and the ability to get to your data ten years from now when you're afraid that Win 2012 won't run OfficeXP because of the upgrade treadmill you've been climbing.

Secondly, and more importantly, deploy a few Linux boxes as test bed machines for a new, vital service.

Yes, you can also run a few for various little tasks like print service, file service, web service.

But what I'm advocating here is you thinking up of what your business could really find useful from a computer. I'm postive that a intelligent sysadmin can see many vital needs, such as for a cheap reliable 3-tier (browser/Apache/PHP/MySQL) web based system that can be used to keep track of <insert dynamic items getting away right now> for your company. After you get it working for a few weeks, show it off to your management and let them point their browser to a dynamic updating portal view of their business. PHBs love that kind of control view.

You can get yourself kudoes for something you put together in your off-hours with free software and using that cast-off piece-of-junk computer sitting out by the dumpster. They'll be impressed.

I think the best testimony to the utility of free software in any business situation is an actual working demonstration. When people see it performing a useful, vital function, day-in and day-out, on obsolete hardware, with no big checks to cut for license renewals, it speaks volumes.

Just do it!