For this example we are going to make a few large assumptions.  Your company has more than 1500 knowledgeable workers.  You are not an IT consulting company.  You have a mix of existing operating systems.  As an example you may have at least one flavor of Unix, Mainframe or Mid-Range System and Windows Systems.  By at least one flavor of Unix, we mean something like Sun Microsystems Servers, HP’s HP-UX Servers or even a SCO Unix system.  Outside of pure IT companies like Google, I have never seen a company without at least one legacy system that no matter what, can never seem to get upgraded.  Most companies are normally trying to replace these systems with Windows or Linux systems to save money.  So we will assume that your goal is to replace a legacy system with 50 Linux Servers and your horizon to sunset the older system is 5 years.

If your have a flavor of Unix already in house you will likely want to start migrating at least a few of those Adminstrators over to Linux.  The transition is relatively easy with the help of a few Tutorials and a site like the Unix Rosetta Stone.  What if you don’t have any flavors of Unix?  There is a rare group of people that know both Unix and Big Iron stuff like Mainframes or Mid-range.  This type of person on your team would be like finding a pot of gold.  In the real world these people are really hard to find so let us keep moving forward by looking at building a team without the super star.

The first person you need to establish on the team is a Linux Admin.  Notice that we didn’t say a lead admin or a super star.  You have no Linux Admins now though so who can interview them?  Believe it or not, other OS system administrators generally have a lot of transferable knowledge to Linux.  We are not suggesting forcing someone to switch from their favorite OS to Linux.  Rather,wo using their knowledge of how networks and other technologies work, like DNS and DHCP, to filter through applicants and find a Linux Administrator.  Once you have the first one, finding more get’s easier.  We are also a huge fan of contract to hire options in this situation.  This gives both sides the most options in the long term.  If your company does more investigation 6 months into a 5 year sun setting and decides to abandon the project, then you do not need to lay off the administrators you hired.  At the same time if one of your potential hires isn’t up to the task you can replace them easily.

Remeber to hire and staff the team with the end game in mind.  If your evaluation of the migration has shown you that it will take 50 servers to replace the mainframe in the back room, then hiring 2 administrators probably isn’t going to work.  At the start of the project our assumption in this situation would be that you would need 1 Lead Admin, 1 Senior Admin, and 2-3 lower or mid level skilled administrators.  Migrations like this will likely take more administrators to get started, so your Senior Administrator and one of the lower skilled administrators may be a consultant.  However, it’s been our experience that once Linux is in place, Linux servers multiply like rabbits.  Several companies we have seen start with just five or ten servers the first year of the project and ended up with a hundred by the end of the second year.  Then, when people figure out that Linux has arrived they start looking for open source solutions to replace expensive closed source tools.  Seemingly without warning your count can grow exponentially.   So all of this means you should over staff, not under staff and do it with consultants first.

In conclusion, make sure that you plan the project.  Poorly planned projects are the largest threat to Linux in the corporate world.  Plan your team for at least twice the size you think it should be.  Use consultants until you have an idea that your non-functional requirements and system estimates are accurate.  Plan for follow-on projects with your new Linux environment.