Clean up your data act

Companies need to make their data items more accessible, which is where information lifecycle management can help, writes Helen...

Information lifecycle management (ILM) is promising to do for information-rich organisations what television’s “How clean is your house?” clutter busters do for unruly households. Namely spring clean, reorganise and store data items to make them accessible according to need.

The IT department is one of the worst hoarders, simply because storage is so cheap that it has been possible to stack up mountains of data. But the exponential increase of information flowing around companies, compounded by tougher rules on regulatory compliance, means that is set to change.

The term ILM has been coined by storage suppliers keen to create excitement about their products and with an eye for rebadging their offerings. However, ILM’s biggest contribution is to introduce a shift in thinking, says Tim Jennings, research director at Butler Group. “A business perspective requires information to be made available and managed according to its criticality and its ability to meet auditing obligations,” he says.

Ovum analyst Graham Titterington says traditional storage system management software and networks deployed by most organisations will continue to play a part in ILM. “From an auditing point of view, it is possible to get storage management systems to keep a record of the activities they perform,” he says. Likewise, from a productivity perspective, it is possible to automate the archiving, copying and deletion of files and to move data into write-once, read-many times storage, or to nearline, cheaper, disc-based storage.

However, he says, “Computer storage does not meet the new requirements in a number of ways.” In particular, the vanilla way most office software and systems are set up and stored means that anyone can set up or delete a file. “Nor do they provide the audit trail which is required by the new rules, where you have to do what the act says and prove it.”

Marcus Hill, director of storage at BT, agrees that many of the underlying technologies required for ILM remain the same. Hill, who has conducted a five-year storage rethink and implementation at BT, says what ILM is really about is “a people, process and technology discussion”.

For the IT manager, this adds up to a server and storage consolidation exercise, a possible technology refresh and consultation with data owners in order to set up policies for categorising data and assigning appropriate storage. This is a spring clean on a huge scale for any type of organisation and the size of the task may be the key impediment to organisations taking the first step.

The cost savings should be a sufficient inducement for most organisations to carry out the server and storage consolidation part of ILM. Butler Group estimates that an IT manager can expect operational cost savings of, on average, between 15% and 25%, depending on the scale and scope of the consolidation. By reducing its 36 datacentres to six, BT has reduced its IT budget by £40m over the past two years, says Hill.

 Hill says the five-year task was “not without its complexities” for BT and Titterington believes the task is so difficult because information and data are fundamentally different. “Most information can be indexed in 40 or 50 different dimensions,” he says. He cites the example of a medical x-ray image with case notes attached which relate to numerous associated conversations about the case, but which might fall into different medical categories.

“This level of categorisation calls for a lot of discipline. Every time a file is created, you have to understand what the key words are that should be attached to it for it to be useful and found.” The technology that comes closest to placing information in its context is content management, but these rules are rarely applied universally across an organisation, says Titterington.

A key problem that affects all companies is how to merge structured and unstructured data. Structured data, which tends to be organised within neat rows and tables within databases, and unstructured data, which is contained in documents and applications, are stored separately but need to be accessed as a holistic information source for business purposes.

Tony Christie, principal of imaging services at Xerox Global Services, says that although summarisation and categorisation tools have been available from document management suppliers over the past six months, their use is not yet pervasive. “It is common experience for someone to have access to all the billing details about a customer, but to lack the actual billing document,” he says.

E-mail is perhaps the single most pressing unstructured data issue to solve. Butler Group says 60% of critical business data is stored in e-mails and yet this is a resource unavailable to the rest of the organisation and with poor control. “If you reach your designated storage limit, it is a case of ‘delete from the bottom’. This is symptomatic of a technology approach to storage,” says Jennings.

BT took the plunge to re-engineer its data storage because of commercial necessity, but it is now prepared for legal compliance. However, many organisations look likely to embrace ILM out of fear of the regulatory stick rather than a return on investment, and may prove to be vulnerable to supplier hype.

Storage case study: BT

BT started its server and storage consolidation exercise five years ago, motivated by the need to reduce its IT bill. To date, it has freed up £130m operational costs for research and development and over the past two years has slashed £40m off IT costs.

As more regulations have come into place, the storage redesign exercise has gone “hand in hand with our compliance efforts”, says director of storage, Marcus Hill.

Like other blue chips, BT used to have one tier of storage for data, irrespective of its value to the business, which was fully replicated and available 24x7. This was applied to the 21 million online transactions it conducts a month, the CRM Siebel systems, product design data as well as file and print material.

A priority was to get a handle on billing data, which is massive in volume with a distinct lifecycle. “The data is most important just prior to issuing statements, and then has to be kept for the six months when most queries occur. Thereafter it does not need to have the same level of availability,” says Hill.

A key part of the exercise is to do an audit of all data and to assign it a commercial value. For this Hill recommends using a third party. “You do not want to get into a war with data owners who say their data is more important than someone else’s. Winning the negotiation with data owners relies on getting a cost structure for different storage options and then all parties can have a profit and loss discussion.

Hill counsels IT directors to approach their finance director with a cost effective plan for making ILM an infrastructure issue. Typically, a finance director-led initiative is engineered at application level, but in his experience this fails to leverage the full value of business data at each part of the lifecycle.

This article is part of Computer Weekly's special report on storage, produced in association with Hitachi Data Systems.

Read more on IT risk management