Moving to Linux
A Small Business Introduction to Linux

By Mark Rais, author of the Linux book Linux for the Rest of Us.  This article published by ReallyLinux.com.

A radical change is resulting in ever strengthening momentum to Opensource software use.  Most importantly, this beneficial change is happening in significant numbers not only here in the United States, but through out the world.

Why?

Ultimately, cost savings is one key factor.  The other primary factor is freedom from licensing dominance that occurs in any non-Opensource environment.  Linux is enabling large and small corporations world wide to end one of their most significant overhead costs with relative ease.  Although a Microsoft sponsored study suggested that Linux is still too immature to realize true cost savings, reality doesn't support this.  More companies today are migrating to Linux from other Operating Systems than ever before in history.  No matter what some statistics try to convey, few are persuaded by the strange notion that Linux costs more.

All this having been said, this article is a brief summary of many dialogues and migrations I've personally been part of in order to help those who may be evaluating Linux for their small business.
 

Step One:  First Things First
Before any business or venture should ever consider moving away from their current OS to Linux, a clear focus must be defined.

There are three principal categories of Linux use as applied to business environments.  These three categories should be evaluated separately and distinctly rather than all together since any wholesale movement to Linux is bound to fail initially.

All business environments tend to include these three distinct uses for computers: Servers such as email servers; Workstations including graphic intensive PCs or number crunchers; and Desktops for basic PC use including checking emails, writing documents etc.

Any movement to Linux should begin by first defining which systems fall into which categories and doing a bottom line cost assessment for the category itself.  In many cases the most expensive category will be Servers.  However, I've consulted a number of businesses and non-profit organizations where the most significant cost is in Desktops as a result of upgrade costs, licensing fees, and new staffing overhead.

"A business that tries to tackle all three categories simultaneously in a move to Linux  will fail at both cost cutting  and streamlining."

Fundamental to any move to Linux is the absolute requirement to define which category has the highest current costs, and then to focus on that single category.  A business that tries to tackle all three categories simultaneously in a move to Linux will fail at both cost cutting and streamlining.  Biting off more than we can chew has become a common step in the process of business, but only because leadership consistently refuses to acknowledge that slow, meticulous change is often more successful than hurried, broad change.

Of course everyone hates change, especially the employees who are directly affected.  Therefore, leaders should avoid adding gasoline to this fire by throttling up a sudden, wholesale change to Linux.  Linux succeeds undeniably when it is part of focused and specific planning, not when it's some executive's bright idea that will end up creating corporate pressure cookers.
 

Step Two: Clean Up Terminology
I keep bumping into people who talk about "migrating" to Linux.  In fact, even a few of the editors at reallylinux.com wanted me to use the term in my title so it would convey what most people are thinking.  However, it's a misnomer at best.

Before we overuse or misuse the term "migration" any more, I suggest that in almost all instances, most small businesses want to perform a move to Linux rather than a migration to Linux.

Any true migration deals with data transfer, which in most cases is data in a database or data stored in certain data structures.  Therefore, there is no direct correlation between changing to the Linux Operating System and migrating data from one database to another.  From a strict standpoint, since the layer in-between is a database or the interpreter for the particular structure, migration is usually a more complex task than an operating system change.  Keeping this in mind then a business running on NT Servers with an Oracle 8i database will have a relatively easy change to using Linux running an Oracle 8i.  In fact, in some cases, almost no migration work needs to be done.  Instead it is a transfer of data that can occur with relative ease if database design was standardized to begin with.

"Don't get moving to Linux mixed up with migrating to new databases.  They may be sequential steps along an overall plan, but they are not directly linked."

Take the other extreme and we see that real serious and painful migration occurs when data is stored in several Microsoft Access databases that then need to migrate to an Oracle 8i DB running on Linux.  Although, for obvious reasons such a change would offer significant benefits, it must also be considered a real migration rather than simple OS change.

I've been part of six major data migrations in my life.  One migration project included the conversion of 650 databases at the Terabyte level and made tortures like fingernail ripping and electroshock appear more enjoyable.  It took my team the better part of a year just to ensure the current data structures made sense to us!  On the other hand, we also had a project to migrate electronic commerce data (some 100,000 records).  Although the stress of error was still high (we were moving records for brand name partners) the number of records and the actual data base changes were very basic.

The main point is that NONE of this had much to do with Linux.  Certainly we would be moving to the Linux platform, and inevitably the operating system nuances played some role, but little impact to the actual migration occurred.  Therefore, it is vital we stop using the word migration when we are talking about OS change, and vice-versa, leaving the term to the serious transport of data from one data base to another.

As I mentioned earlier, lest we seriously misuse the term "migration," I suggest that in almost all instances, most businesses want rather to perform a move to Linux.

Migrations at the data and database level are substantially more significant and need to be evaluated beyond the scope of operating system changes. 

Therefore, please don't get moving to Linux mixed up with migrating to new databases.  They may be sequential steps along an overall plan, but they are not directly linked.  Instead, a preliminary and foundational step to future improvements is an initial move to Linux.
 

Step Three: Keep Focus
I had a business leader who became extremely excited about changing his company to Linux as a result of a simple evaluation that showed cost savings.  He was enthusiastic to the point of insanity.  His desire was for all of his staff to get retrained, for his infrastructure servers to be wholesale moved to Linux, and for "every machine in this place including the coffee maker" to run Linux.

"I persuaded this leader  to slice his grand Linux visions into reasonable chunks  and apply Linux initially to the most expensive or needy areas."

Now, I don't doubt that Linux is up to the job, and indeed every reasonable application and function for computers in this particular corporate setting could gain from Linux use either in terms of performance or cost savings.  However, the concept that all of this change should come overnight with beneficial results was ignorant at best.  I persuaded this particular leader to slice his grand Linux visions into reasonable chunks and apply Linux initially to the most expensive or needy areas.

In a few short weeks (it was actually four days, but his staff were working over time to make the change quickly) they had converted all of their desktops and several intranet servers including print servers to Linux.  Training had been minimal and was supported by the exceptional training and support benefits of SUN Microsystems which kindly provided for the great and popular OpenOffice.og program they chose to use.  Now they were running all of their PCs with an operating system that was free to own, free to manipulate, and free to upgrade combined with running office productivity software that was a quarter the cost with far lower future upgrade prices.  All of this AND they also gained the benefit of better performance from their intranet server, mail server, and print server.
 

Reality Check Time
Okay, what's the catch you ask?  Can it really be this good, and if so why isn't every small business converting to Linux?

The answer to this is a combination of FEAR and FAD.

Fear represents not only the emotional concerns about change and I mean ANY type of corporate change.  Even when the type of coffee that is brewed changes it is bound to cause someone some kind of negative response to the change.  But fear also includes the fear of the unknown which results in passive resistance.  How much is your company or institution paying in dollars for this fear?  In many cases we find that people are not willing to change software or the OS out of personal fear of change that will inevitably cost the company thousands of dollars... and I'm referring to a small company not a mid-size venture.

Fad represents the hooks that other companies get into us when they embed their applications so deeply into our lives that to disconnect ourselves from their product will cause pain and discomfort at the least.  Let me give two personal examples that may help convey this idea:  Microsoft Exchange Server is a very complex system... and if you've ever had a crash and tried to restore from tape or other backup you know full and well what I'm talking about here.  It is complex and ultimately costly to maintain.  Yet many corporations use Exchange Servers all over the place.  This is a result of what we can call ingrown apps. Another example is the popular Outlook that has it's own demands of what it prefers to run under and with.  I've had more than four companies in the past few weeks contact me with an "emergency" related to their Outlook email handling.  I usually try to help and then ask why they are using Outlook?  The answer almost always is "because that's what we've been using forever."

Is that a GOOD reason?  Is it a reasonable enough excuse to explain thousands of dollars in upgrade costs, maintenance costs, and ultimately daily inefficiencies?  How often I've tried to help someone work out a solution to another Trojan or Spam hack that used Outlook as the door.  It's remarkable what some people decide they are willing to live with regardless of the pains and costs.  Yet, this strange FAD to stay with a certain product line and company also results from many small business IT departments simply existing day to day in fire fight mode and not giving planning a thought.  This is not the result of bad IT people so much as bad leadership decisions.

I mean to be serious here.  I have personally witnessed one IT director and his staff running from server to server, room to room all day long, several days a week just trying to keep their stuff alive.  There I sat in the same office but next to a perfectly good Linux server handling about the same number of users with an up time of over 120 days.  I assure you I was not doing daily maintenance, worrying about failures, or running around like this.

What can possibly be the reason some do their jobs running around like chickens with their heads cut off?  Job security?  No.  It's related to the inability to come up for enough air to actually be able to reevaluate infrastructure, corporate policy, and for that matter longer term IT goals. 

To plan requires stability.  If they did have the chance, inevitably almost all of them would choose to do things significantly different and Linux would absolutely be popping up on the list of best available options. 

Today, well over a decade since the Linux revolution began, many small-size institutions are turning dramatically and with beneficial results to Linux solutions.

For those that are not, neither FEAR nor FADS are reasonable excuses for clinging to ideals and not applying realistic plans for incorporating Linux.  People, whether with entrenched fears or not, will inevitably turn to what works. 

Linux works.

I'm convinced that in less than a decade, more computers will run Linux than any other operating system in the world.  We can choose to embrace this or cling to our fears and fads.  Choosing to move to Linux now is an important business decision -- both for the short term and long term.

More articles on Linux and the Small Business from ReallyLinux.com:

• Opportune Time to Switch to Linux
• Getting Linux on to the Desktop