Unix-to-Linux Migration Has Come of Age

The driving force behind every migration is the promise of a better tomorrow. It’s no wonder, then, that a good number of migrations, human, animal or otherwise, have resulted in a positive ROI. Let’s take migrations in IT systems.

Shrinking IT budgets, coupled with the cost advantages that Linux-Intel platforms provide are driving migrations to Linux. I have noticed an increasing number of enterprises seeking information on Linux migrations, as well as opting for the migration — and the numbers are growing.

With migrations come the herd and crowd mentality. It is said that it takes a mere 5 per cent minority to influence a crowd’s direction and that the remaining 95 per cent will follow for sure (findings from a research conducted by the University of Leeds, 2008). Applying these statistics to IT systems migration, it’s tempting to infer that organisations migrating to Linux are being driven by the herd mentality. So, should you follow the herd?

You’ll discover the answer to that shortly. Meanwhile, let me put it in simple and direct terms: “If it can be migrated, it should be migrated.” You’ll learn why in the course of this article.

But before we get there, let us touch upon some important aspects of the migration itself, including when to migrate, what to look for in migration vendors and a few best practices.

When should I migrate?

It’s not a question any more of whether organisations should migrate; it’s merely a question of when. There are no ready-made answers for when you should consider migration. A considerable amount of calculation goes into deciding when you should take the plunge — and it all boils down to a reasonable risk-reward ratio.

A good yardstick is the ROI that you can expect from this migration in a fixed time-frame. Your quest for positive ROI should be pegged at a reasonable time-frame (around 2-3 years) from the start of the migration. This means that you must realise a positive ROI at the end of that period, if not earlier. If your vendor can categorically determine and commit to this time-frame, then you must consider migrating.

Who should I turn to for migration?

A vendor-agnostic vendor

This term isn’t intended to be an oxymoron. It’s important to find a vendor who will provide you standard service, without locking you into proprietary technology. Some vendors may have compelling reasons to make you play their cards, like having their own RISC/UNIX (Reduced Instruction Set Computing) product lines, or their own Unix versions.

The vendor you choose should have a commitment to industry standards like POSIX (Portable Operating System Interface — for Unix) and LSB (Linux Standard Base). Insist that the vendor proves there is no lock-in for hardware, software, operating systems or services.

The breadth and depth of experience

Look for vendors with experience that spans the migration spectrum, from large-scale enterprise-wide migrations involving multiple applications, to individual projects for single applications. This ensures that they have the technical and strategic know-how, and have gained both a breadth and depth of knowledge. Having rolled up their sleeves and gotten down to work, they have the migration process figured out almost to the point of rendering it into a scientific methodology.

Another key factor is to look for a vendor who has the capability to solve support issues in-house. Such vendors should have also migrated to Linux from a variety of source platforms including popular Unix variants: Solaris, AIX, HP-UX, SCO, NCR SVR4, as well as other legacy platforms. Further, they should offer you a choice of Linux distributions as target platforms, like Red Hat (RHEL), Novell (SUSE, SLES) vendor distributions, or community distributions like CentOS and Ubuntu.

Meticulous methodologies

Meticulous planning and strong project governance are key requirements for a smooth migration, and these can stem only from established and well-developed methodologies.

Apart from solid methodologies, make sure the vendor brings to the table a good mix of practice leads, consultants, program managers and architects — preferably people with a good network of industry partners, which includes the likes of Intel, Red Hat, Novell, Accenture and BMC.

The vendor should have sound methodologies that could help you assess, design, implement and maintain migrated environments. The vendor should also be able to maintain servers, storage and software under a single umbrella.

Best practices of migration

Good vendors with established migration practices come with consulting methodologies and best practices that have room for a whole gamut of considerations. They address specific customer pain points, and present a comprehensive solution that covers the entire SDLC (software development life cycle), spanning assessment, planning, design, migration and deployment.

As pointed out earlier, at the heart of the decision to migrate, lies the calculation that determines the time-frame for a positive ROI. There are many elements that affect this calculation that an inexperienced vendor might overlook. A few examples are: cost of reconnecting interfaces broken by migration, the impact of migration on the user community, leverage obtained by selling hardware and converting it into cash, factoring in the green element of Linux-Intel (power saving and reduced energy consumption for cooling), etc.

Your vendor’s repertoire of best practices should account for all migration-induced changes, and their effects.

‘If it can be migrated, it should be migrated’

Now, coming back to my earlier quote, here are some compelling points that work in favour of Linux. In relation to crowd or herd mentality, there is another opposing school of thought called the convergence theory. It says something radically different about crowd behaviour: that people who wish to act in a certain way come together to form crowds. That’s how migrations really happen, driven by a common goal and the promise of a better tomorrow.

Similarly, people who stand for Linux migrations come together to convince the industry that it is indeed something that CIOs, business managers, IT directors and Unix administrators should seriously consider. And that it is indeed the way forward for IT enterprises that intend to survive and thrive.