intelligence
Rethinking data management

Does your current strategy meet the business demands placed on your data?

With recent, headline-worthy security breaches providing a stark reminder of the contentious issues surrounding data management, CIOs are being prompted to re-evaluate their current data management strategies.

In August, CEO, Monster.com, Sal Iannuzzi reported that the personal details of 1.3 million customers had been relayed to a rogue server in the Ukraine. Once located, this server was discovered to have collected over 1.6 million personal details via the malicious infostealer.monstres trojan. And then, in November, came the most high-profile data breach to date. The UK's HM Revenues & Customs mislaid two CDs containing the unencrypted details of 25 million British citizens registered for child benefit. Public outrage at the government's potentially catastrophic data management failure soon followed and an enquiry was launched.

As the dust settles, important questions are emerging for CIOs regarding the security and management of their own data assets. These issues extend far beyond big breaches (which the media are quick to pounce upon) and run deep into an archaic data management culture.

“The most important factor influencing the need for enhanced data management is that most businesses worldwide can no longer survive without their IT capabilities,” said president, Horison Information Strategies, Fred Moore. “The value of all types of data is increasing daily and therefore, the requirement to manage that data and make it accessible and secure whenever it is needed is more critical than ever before.”

Across industries, data management strategies are buckling under the sheer weight of the data they process. A recent eMedia survey commissioned by Bridgehead Software revealed that 73 per cent of organisations currently hold over a terabyte of data, up to half of which will not be accessed again. Moore explained that many enterprises hold significantly more, creating a disproportionate ratio between data and resources: “By 2008, it is expected that the average non-mainframe storage administrator will be able to effectively manage nearly 20 terabytes of storage, while the amount of data to be managed will approximate 80 terabytes per server,” he said.

To compound the problem, a 2007 Quocirca report entitled 'Data Management Today' revealed that data is often haphazardly held in “multiple storage management tools in different areas of operation within one organisation, leading to duplication of people skills and effort, and wasted resources in terms of both licensing and management.”

The segregation of these systems poses a threat to the competitiveness of an organisation in the marketplace because it hampers its ability to dynamically deploy relevant information across the organisation as required.

“Older style approaches of large data marts have pretty much had their day – the volumes of data are just too high for anything meaningful to be done with them,” business process analysis service director, Quocirca, Clive Longbottom told CIO Quarterly. “Newer approaches, such as master data management, can provide the best of both worlds – a fast main data mart of referential data, with a federation of less visited data being held in other pre-existing databases.”

Of course, CIOs are constantly looking for new ways to integrate their data into the life-blood of their organisations, and appreciate the business value of the data they collect, store and utilise. However, in order to retain the value, effectiveness and flexibility of this data, CIOs must start heralding change.

Longbottom warns that CIOs that do not embrace newer approaches to data management will find it harder to analyse, utilise and backup their mounting data sets, resulting in spiralling storage costs. “Overall, complete chaos. And it doesn't take much to see that a competitor doing this even half well will not just be a stronger competitor, but will likely start to put you out of business,” he said.

Potentially, non-agile information platforms damage the perception of IT operations at executive level, where a direct correlation between data and value is demanded.
Quocirca's report suggests that a top-down data management strategy is the only way to rectify this problem, whereby all current data is placed within a single storage architecture to facilitate retrieval.

In its simplest form, Quocirca suggests that data from existing systems should be captured at the abstraction layer and assigned a set of policies to distinguish it from other sets of data. However, the technical difficulties that arise from data migration are complex and cannot be overlooked.

In 2007, IT research organisation, Bloor released a white paper entitled 'Data Migration', concluding that this area “has historically been under-valued, under-resourced and not treated with the attention it deserves. As a result, many major migration projects have failed, either partially or completely.” In remedy, Bloor emphasised the need to understand and cleanse the existing data before it is moved into a new architecture. Finally, the new system must be rigorously tested to ensure effectiveness.

Although such a strategy would improve the agility of an organisation in meeting its present and emerging needs, it may be difficult to achieve for established businesses that have acquired incompatible systems during their lifespan. This leads Quocirca to suggest that “starting from scratch has some advantages, as other players who gained much of their technologies through acquisitions are not so well positioned to provide integrated systems.”

Certainly, all enterprises will have to implement such strategies in the long-term, but those that do so in the near future have much to gain. “The demand for a universal policy-based software platform is growing daily and a single platform will enable data management to keep pace with data growth. Expect that this challenge will ultimately be met, though not any time soon, and the company that does it will become a major data and storage management force for many years,” said Moore.

Longbottom also concedes that such a change would be technically difficult because “existing storage assets do need to be sweated.” In reality, he recommends a phased approach. “Use existing storage as necessary, but replace it as it reaches end of life with a more open system that can be fully supported by the software in place. Fully open, manageable storage is a journey, not a one off solution.”

Perhaps one question that needs to be asked is whether the cost of restructuring the data management strategy for an entire organisation will be outweighed by the cost of updating its current capabilities, especially in light of growing public concern regarding the security of personal data, and the changes in government legislation or corporate policy that may result.