Roman Sakhno - Fotolia


Are you stuck in a legacy data center state of mind?

Legacy data centers with proprietary code and outdated business practices are familiar, labyrinthine and difficult to leave behind.

Mountains of COBOL code, 30-year-old mainframes and long waits for process changes that take months to implement are the reality for a surprisingly large part of the IT industry.

Wal-Mart has hammered small, locally owned stores across the U.S. Business models aside, the retailer's real secret is an agile IT system that speeds up restocking and analyzes demand trends overnight. But even Wal-Mart struggles against the massive flexibility of online retailer Amazon, which has an IT platform that allows small retailers to sell through its website, using its online services and software.

These extreme success stories point to the power of underpinning a good business strategy with the best IT infrastructure. That means legacy data centers, servers and applications must go.

Legacy data center gear is slow and expensive to run and maintain. And many legacy programming languages are dead; COBOL programming language is no longer taught in schools, and most programmers want nothing to do with it. In fact, some consider COBOL skills as a job-killer on resumes, indicating someone who is far from leading edge.

When asked why they stick with legacy technology, CIOs usually talk about constrained budgets for IT and the cost and risk of transitioning to modern systems. They've been talking this way for at least 10 years, and their predecessor probably said the same thing. Yet every other part of the company must be overhauled regularly or the company would go out of business.

The cost of sticking with a legacy-heavy data center increases each year, as experts on particular code modules retire from the company and cannot help transition to newer, standardized applications and systems. But this is only the cost of executing the transition. The more painful cost is what's experienced from lost business opportunities that derive from a rigid and inflexible system. The cost of lost opportunity typically far exceeds IT costs, but the numbers are much harder to quantify. And this makes a business case more difficult to prove.

What to expect when you leave legacy servers behind

Start your transition off with a scope, then a plan. Legacy apps are like medieval London in 1666: patched and sprawling, many abandoned and all ready for the Great Fire.

Most transitions to modern IT platforms fall down when they start to challenge business practices. Legacy systems typically enshrine the way the business ran 30 years ago -- or more. Business processes must inevitably change. The legacy system creates a pent-up need for rethinking, and yet it's tempting to keep things the way they are.

Look for prepackaged products that can replace legacy systems. For example, commercial-off-the-shelf (COTS) enterprise resource planning systems demonstrate high agility and best practices engrained in the program.

Consider deploying a modular business apps configuration. Traditional legacy approaches tightly intertwined various departments, but the modern modular concept concentrates on the interfaces between the modules and their ability to interchange data -- allowing for a great deal of independence.

This transformation away from the legacy data center of old now comes to the part of the business suite that is so proprietary that it had to be coded. Separate your beliefs from reality when you assess the need for legacy code. A study shows that approximately 40% of legacy code is unused and won't need any transition actions, so you need to discover dead code before making unnecessary transition plans. Substantial amounts of code may be present to deliver on old, unnecessary business practices.

Proprietary code often glues things together. It's likely some code will need to be written, but definitely not in COBOL.

The public sector is one of the worst offenders for clinging to legacy IT systems and data centers. Efforts to change that, from data center consolidation to a cut-and-reinvest trial program around IT spending, have met with mixed results.

The future IT shop: Are you in it?

Transition is never easy -- making a move requires a concerted determination to change.

Many IT staff look at changes, to the COTS environment and then into the cloud, as a way of replacing the legacy data center and as a job threat. COBOL programmers see no place for themselves in this future. Admins know that commodity and cloud hardware need fewer administrators and support staff. The CIO and immediate staff fear moving away from a base that served them throughout their careers.

The move from legacy servers and platforms to modern, modular and flexible ones won't happen overnight. And the effects these changes have on personnel actually take longer to occur than people expect. Legacy skills are not immediately obsolete. There is a shortage of COBOL experts across the industry, for example, and this will grow as transitioning increases, while skills built around legacy link-ups to new COTS systems are likely to be in high demand.

Resistance to change is not confined to IT. The rest of the business sees disruption looming and will resist too. Any transition program needs the explicit buy-in of all the executive team, with a strong downward enforcement of the will to leave legacy IT thinking behind.

About the author:
Jim O'Reilly is a consultant focused on storage and cloud computing. He previously held top positions at Germane Systems -- creating ruggedized servers and storage for the U.S. submarine fleet -- as well as SGI/Rackable and Verari, startups Scalant and CDS, and PC Brand, Metalithic, Memorex-Telex and NCR.

Next Steps

How to stretch a data center's lifetime by 5-10 years

Understand commodity servers

Hardware to float your cloud upon

Survive in the orchestration era

Dig Deeper on Server hardware strategy