A data storage strategy called ILM

Information lifecycle management has just about entered India

You can't afford to lose your data these days "� you never know which data will suddenly turn critical. True, a good part of organisational data follows a set pattern and can be gradually retired. But sometimes you may need to retrieve or keep handy even old data.

A telecommunications company, for example, requires call records not only for billing at the end of the month but also to revisit the data if any dispute arises. Still, where do you park all these data? Storing data can be very expensive, if not managed properly.

That's a reality storage experts are fast awakening to. Enter information lifecycle management (ILM). Essentially a storage strategy rather than a technology, ILM tries to classify data according to usage and importance, selects the appropriate storage and evolves a policy for its migration over time.

In doing so, the strategy recognises that the criticality of any data can change either way over time. The attempt is now to automate the migration of data between storage systems not only down the hierarchical order but also up it.

But ILM's prime selling point is that it works with a combination of storage environments to bring the total cost of ownership (TCO) of data down.

Data can be stored in a high performance fibre channel provided by a storage attached network (SAN), in online high capacity low performance or advance technology architecture (ATA), backed up on a disk and on tape.

Storage media costs range from about 15 cents a megabyte(MB) for fibre channels and 5 cents and MB for ATA to 3 cents an MB for disk storage and less than a cent an MB for tape storage.

ILM uses a combination of these different storage media, guided by prudent classification of information and data based on the importance and frequency of usage and the need for archiving.

The three-year-old ILM arrived in India a year ago. ILM vendors have to be not just storage experts but also have the ability to understand different business processes.

That's why only the bigger companies like Hewlett-Packard India, IBM India and EMC Corporation and Hitachi Data Systems (HDS) offer solutions that involve ILM in some form or another.

Still, ILM can operate only with the newer generations of external storage systems such as SAN and network attached storage (NAS) systems, rather than direct attached storage (DAS) systems.

Indian companies are still fairly stuck on internal storage systems (40 per cent of the data) and lean heavily for external storage (60 per cent) on DAS (55 per cent). This makes ILM seem a bit too futuristic for Indian ground realities.

But proponents of ILM here underline forecasts by IDC, the IT consultancy, that the market for DAS will slow down by 23.9 per cent between 2002 and 2007, while the NAS market will grow at a compound annual growth rate of 34 per cent during the period.

But SAN and NAS costs are falling, says Avijit Basu, country sales manager, network storage solutions, Hewlett Packard "� India. "In the past three-four years, the cost per megabyte has gone down by about 40-50 per cent," he estimates.

This apart, the reducing complexities in storage architecture have encouraged more companies to migrate to SAN and NAS. "Today 75-80 per cent of our storage business come from SAN and NAS implementation," Basu adds.

Manoj Chugh, president, India & SAARC, EMC Corporation, backs the point: "In the next 12 months the share of storage between DAS and NAS will move to 50:50 and the current ratio of 55:45 will flip in two years' time, with NAS gaining predominance."

That's why ILM vendors believe in its future prospects. EMC thinks that the biggest buyers of ILM will be telecommunication companies. "Data management is vital to this segment, right from billing requirements and dispute redressals to the deployment of CRM tools," Chugh adds.

Manufacturing enterprises and banks are two other big prospective customers. "With the focus veering to manufacturing efficiencies along global standards to establish international competitiveness, the segment needs all the help it can get in data storage," a storage expert points out.

IBM thinks the media, the pharmaceuticals industry (to store clinical data "� especially companies that sell products in the US and the European Union), insurance companies, hospitals, the government and business process outsourcing companies could be big clients (though not everyone agrees here).

Regulatory requirements in the US and the European Union such as the HIPPA standard (for the healthcare industry) and the Sarbanes-Oxley Act will also push companies with affiliations or or stock market listings there, to pursue a more streamlined storage practice on the lines of ILM.

"The potential for ILM solutions is vast in India as increasing regulatory pressures force customers to look at ILM solutions," insists Shailesh Agarwal, country manager, storage solutions, IBM India.

ILM vendors say that chief information officers of companies and IT managers in India too are waking up to the need to manage data. "There is a good response to our reference information storage system (RISS), which has strong ILM applications, " Basu points out.

With global data traffic growing 100 per cent year-on-year, effective data management is becoming critical to enterprises. EMC's estimation is that ILM can almost halve the total cost of ownership (TCO) by bringing it down to 52-54 per cent of the original cost, when fully implemented.

It eliminates ad hoc storage strategy decisions, underutilisation of installed storage, the use of wrong storage medium and the lack of planning for future growth.

Even so, ILM's success will hinge largely on an organisation's ability to identify its data profile and usage and to evolve sound policies for migrating data to relevant storage space. Are Indian companies ready for ILM?

"A large number of enterprises today do not have a data retention strategy in place. Customers are challenged with having to balance business issues and the technical perspective to maintain a best in class approach to data retention strategy," Agarwal admits.

"In India, where the storage capacities are rapidly increasing and storage management costs are galloping faster, ILM is currently being used in its simplest form, which is hierarchical storage management (HSM)," he further notes.

But ILM is more than just HSM alone, he admits. ILM vendors today are encouraging the piece-meal application of ILM to applications such as e-mail archiving.

Basu sees only four or five large organisations in India showing some comprehension of ILM and interest in it.

Implementing ILM

• Differentiate production data (files used in day-to-day operations) from reference data (files not accessed often but that still need to be maintained and kept available).
• Assess the reference data. While the likelihood of data reuse is largely directly related to age, study their retrieval patterns. SRM tools can be used to track data usage patterns. The reports these exercises yield can then be used to get a fix on storage asset utilisation and the costs involved.
• Next comes the classification of the data based on how critical they are to business operations, their type and age. This will help the organisation's IT department to determine where the data needs to reside through its lifecycle and evolve policies to migrate it to the relevant storage over time. Automated data migration (ADM) tools can then be used to actually migrate the data from one storage class to another, guided by user-defined policies. ADM solutions, unlike the earlier generation linear hierarchical storage management (HSM) systems, allows the retrieval of data to a higher-level storage resource if the data's value changes after it has been migrated to the secondary storage.