However you will be hard-pressed to find a robust data governance program at most organizations because most organizations go about it the wrong way. Many start with a big-bang approach which can be summarized as coercing a dozen key stakeholders to take half a day a month to talk about standards, road maps and maturity models.  It also doesn’t help that business leaders often struggle with quantifying the value of “data management” or “data quality”.

Build your data governance program one issue at a time.  Find a data problem, find the associated process issue, measure the breadth of the issue, work with the stakeholders to get it fixed, measure the results.  Rinse and repeat.  As you work through each issue, start presenting the results cumulatively.  Try to focus on low hanging fruit initially (few stakeholders, easy fixes) and as you build measurable results, expand to include larger issues with more stakeholders that require broader decisions.  Keep up the regularity of the issues you find, create a tracking list or queue, and start having regular meetings with the stakeholders including IT.  All of a sudden, you have yourself an actionable group of stakeholders making data governance decisions. They see the value, they understand the importance, and you have shown quantifiable value early and often.

Happy New Year!

Julie Lockner

2010 - Then Founder & President, CentricInfo

2011 – Now Sr Analyst & VP, Data Management, ESG

An oversimplified example might be: database A has a table with a million rows containing five years of data. After a thorough analysis, it is revealed that only three months of data are really used and the rest can be removed, effectively cutting the table down to 50k rows. Now, when implementing a tiered storage strategy, we only need to store 50k rows (perhaps on a higher tier based on its usage) as opposed to splitting hot and cold data and overstoring unneeded information. We’ve effectively reduced the hardware, facilities and support cost by not storing the other 950k rows of data and at the same time aligned the business process (data usage) with the correct storage tier. Expect to release somewhere between 50-75%* of the now unused storage back to the procurement pool to be repurposed.

With a full-cycle ILM implementation, we can show not only that the right storage mix for our business (business performance) has been delivered, but that we have reduced the overall usage (therefore cost) of our IT assets (business efficiency). In other words, true organizational value will stem from ROA: the more effective and efficient use of assets. To deliver ROA with ILM, intiatives must address everything from the business process to the disk.

*Estimated range. 95% of the written data was removed, however, you need to account for database overhead (reorganization, partitioning, etc) and storage overhead (LUN sizes, virtualization, etc).

One of the key goals of the capacity planning discipline is to achieve a balance in infrastructure assets. Essentially, this involves balancing CPU – your raw horsepower – with I/O delivery or throughput. An example of an unbalanced architecture would be where you have lots of CPU to crank through data, but your throughput is too constrictive to allow full use of your CPU. 

In this situation, either you have the correct CPU configuration, and you need to get faster disks, or your disks are of sufficient speed and capacity and you have over purchased CPUs. This is essentially an exercise in balancing good old supply and demand. The exception in this case is that the invisible hand† that naturally achieves balance is replaced by conscience decisions made by IT professionals.

But how do capacity managers know when to pursue supply-side or demand-side manipulation? Oftentimes this is done by taking a look at historical utilization, performance, waits and other measures.  If I/O waits are high, then focus on increasing your throughput (supply).  If your CPU is pegged, then you need more horsepower. Unfortunately, most capacity planners stop there and are missing a bigger piece of the puzzle…the context. Even if you reach the holy-grail of balanced infrastructure, who is to say that the point of IT economic equilibrium you have achieved is in the right equilibrium? It is little wonder that many business teams view IT as the invisible hand.

–begin Dennis Miller-ish rant here–

Attempting to plan a tactical infrastructure that meets future business plans by reading historical throughput rates is like trying to plan your retirement strategy by only looking at pay stubs from last year. Pay stubs do not tell you how many college tuitions you need to pay for or that your 1989 Honda Accord is nearing its final demise. The incorporation of business planning is absolutely crucial. Processes are intimately ingrained in both tactical infrastructure design and the ever-changing landscape of business strategy. Without thoroughly integrating capacity planning with business planning, you are missing the point.

Tom Kornegay, Freelance Contributor, Data Architect Extraordinaire

†From Wikipedia.org: “In general, the term “invisible hand” can apply to any individual action that has unplanned, unintended consequences, particularly those that arise from actions not orchestrated by a central command, and that have an observable, patterned effect on the community.”

Unless the company has a super star management team (which we all think we have), progress halts and money is wasted.

This Blog – Bridge The Gap – focuses on real issues facing companies, small and large, that are dealing with conversation gaps between the Business and IT. With a lack of business requirements, IT can’t build supporting infrastructure. Without rockin’ and appropriate technology, the business can’t compete, excel and conquer.

We will cover topics from cool technologies, stupid processes, and things that seem to work. We are a fresh voice in the mundane din of vendor presentations and don’t have time for a bad sales pitch. Let us know if there is something specific you want us to talk about or just want to rant about a lost day at a vendor EBC – because you will never get that day back.

Welcome to Bridge the Gap blog – and no we are not talking about the Tube.

Julie Lockner, President & Founder, www.CentricInfo.com