The ILM imperative

Smart use of the data storage infrastructure is helping many companies tackle data management and compliance burdens – but an information lifecycle management strategy requires a lot of upfront effort.

By Jessica Twentyman

Storage costs have never been lower. Data volumes have never been higher. For information management professionals, that makes shopping for more hardware to accommodate burgeoning data stores seem like a relatively pain-free option. In a recent survey by market research company Vanson Bourne, for example, four-fifths of respondents admitted that they simply purchase more storage rather than investigate ways of getting more from their existing resources.

They may be straying into dangerous territory, warns Nigel Ghent, UK and Ireland marketing director at data storage supplier EMC. “This kind of approach is not sustainable in the long term – for the simple reason that it’s extremely costly to keep all of your corporate data in its entirety on highperformance primary disk,” he says.

Robust sales of primary storage may be good news for EMC, but the company is more interested, claims Ghent, in helping customers to use their storage resources more “smartly” – eliminating redundant copies of data, moving less critical data to cheaper storage media and recovering over-reserved capacity.

Those that carry out these activities as part of an overall information lifecycle management (ILM) strategy have reaped impressive results, claims Ghent. EMC recently commissioned a survey of 250 UK and Ireland organisations from analyst group Quocirca. Respondents said that they have now moved from the planning and implementation stages to incorporate elements of ILM in their storage strategies as an intrinsic part of the business infrastructure – and that confidence in their organisations’ ability to manage and retrieve all types of data has significantly increased as a result.

That could be dismissed as mere marketing spin, but a host of respected and independent industry voices are equally positive about ILM’s ability to deliver on its exceedingly ambitious promises. Analysts at Gartner, for example, point out that utilisation of typical direct-attached storage rarely exceeds 50 per cent. By contrast, highly efficient networked storage environments, organised according to ILM principles, they claim, can operate at between 70 and 90 per cent capacity utilisation.

For many organisations, however, that kind of efficiency is still a pipe dream. In fact, since the term ILM first achieved industry-buzzword status several years ago, many information-management professionals have found ILM to be intriguing in theory, but complex in reality. They should not feel discouraged, says Gartner analyst Ray Paquet. “ILM is a grand vision,” he says, “and, as with most grand visions, total success is unlikely to be achieved immediately.” ILM, he adds, is a three-to-four-year strategy – at the very least.

But that is not to say that they shouldn’t start that journey as soon as possible. Individual storage management initiatives, he says, can yield substantial return on investment, even in the short term. “If you modify the storage used and the levels of data protection throughout the lifecycle of a piece of information, you can immediately lower the total cost of ownership, achieve some level of compliance and improve availability of information by aligning information with business goals and service levels.”

In fact, many organisations may have already progressed further along the road to ILM than they actually realise, without necessarily having in place a dedicated ILM strategy – much less an ILM budget. For example, data protection is part of the information lifecycle, so re-examining the back-up strategy and introducing new systems there (which may introduce another ‘tier’ to the storage infrastructure) might justifiably be considered to be first steps in the direction of ILM.

“The point of ILM is to make sure you’re not eating up expensive fibre channel drives with stuff you don’t need to save, or that, conversely, you’re not tossing information onto tape you might need to rapidly retrieve later on,” explains Nick Laurence, UK enterprise marketing manager at Dell. “Most IT directors I speak to readily acknowledge that their organisations store far too much information on high-end disk and there is, naturally, a strong impetus to move away from that situation,” he adds.

Step by step

Achieving the goal of ILM is entirely dependent on an organisation having a good understanding of the types of data it holds, the value that data has and where it is stored. ILM works on the principle that all data is not equal and that it is used in different ways at different stages in its lifecycle.

For example, monthly financial reports may combine sales orders, shipments, inventory expense and other data for the month. During the processing cycle, the finance department needs to access this data frequently and rapidly. After the reports have been generated, however, the previous month’s data is referenced less frequently as the focus changes to data for the current month. Previous reports can then be migrated to less-available, lower-cost storage, thus freeing up primary, high-cost storage.

The first step on the journey to ILM, then, is achieving a good understanding of what data is stored and where it is kept. From there, data can be categorised according to its requirements – for example, how often it is accessed, how quickly it may need to be retrieved and how long it needs to be kept for, in accordance with regulatory and legislative mandates.

“This can, and should, include a ranking of applications and their data, relative to one another, based on the value of that information to the business,” says Michael Peterson, an analyst at Strategic Research Corporation. “This value can include intrinsic value, based on how much money is lost due to unavailability of the data or the application, and it can include extrinsic values, such as risk and company image associated with loss of the data.”

The value of the data then leads to the definition of service level objectives (SLOs) – acceptable levels of service for performance, availability, operational recovery, disaster recovery, security and more. For many companies, this is a “huge challenge in itself ”, says Ghent of EMC. “In particular, it relies on a wide-ranging collaboration effort between the business and IT, which is never easy,” he says.

The second step along the ILM road is to tier the organisation’s storage systems. Paquet of Gartner recommends that IT organisations with small environments maintain no more than three tiers of storage, while those managing larger environments use four tiers.

“We generally advocate four levels of storage. First: fast access, high-performance primary disk. Second: low-cost disk such as SATA [serial advanced technology architecture] disk. Third: tape technology where data must be retained but is unlikely to be referenced again. Fourth: offline tape in a secure facility, possibly offsite, which can be manually reintroduced into a tape library in the very unlikely event that it needs to be recalled,” he says.

Except in isolated cases, four tiers should be more than enough. “Managing the tiers and the service-level agreements (SLAs) associated with anything more than four tiers will decrease or even eradicate the return on investment provided by a tiered storage infrastructure,” he warns.

The third state on the ILM journey is to map categorised data to these data tiers. This should be carried out according to a principle of ‘solution stacks’, according to Petersen of Strategic Research. “A solution stack provides a complete (or nearly complete) ILM environment around a particular application,” he explains.

“For example, based on the SLAs for a specific application and its information, define where active primary storage goes, where long-term online retention is held, and the regulatory and archive repositories,” he says. Once this is done, he adds, an organisation will be ready to begin operating the solution stack in accordance with ILM principles. Automation is the final, essential component for ILM – but it relies on a significant software investment to be successful. Without a rules-based software layer that automatically moves data between storage tiers according to pre-set policies, companies are forced to rely on manual efforts – and that is, at best, a short-term solution.

It is also where many ILM initiatives fall down – and where the storage industry still has work to do. Put simply, the SMI-S industry standard championed by the 300 vendors that make up the Storage Network Industry Association (SNIA) is, as yet, too immature to enable a single storage management software console to enforce sophisticated ILM policies across all devices in the typical multi-vendor, heterogeneous storage environment.

That is a major failing and until the situation changes – and much is promised for 2007 – companies that wish to apply ILM to multi-vendor storage infrastructures must use multiple software products to carry out even the most basic ILM tasks. But by carrying out the first three steps on the roadmap to ILM and by deploying software products that comply with existing, early versions of the SMI-S standard, insist vendors, companies can be sure that their ILM destination is at least in sight, if not yet attainable.

Jessica Twentyman can be contacted by e-mailing, jtwentyman@ark-group.com.

Case study: WHSmith News

WHSmith News is the UK’s leading newspaper and magazine wholesaler, handling more than 50 million newspapers and magazines each week. And thanks to a far-reaching ILM strategy, it is becoming just as adept at handling its own, internal data, according to Robert Wilson, the organisation’s chief technology officer.

“We initially wanted to manage our information more effectively, but we soon realised that we could reduce costs if we managed our information according to its changing value,” he says. “As such, ILM became a critical element of the selection process.”

WHSmith worked with storage supplier EMC and systems integration partner CapGemini to implement a tiered storage environment and map its information and applications to the most appropriate systems. Because WHSmith News wanted to move data away from disk-based systems to tape for long-term storage every day, the new infrastructure was required to automate data movement and back-up to support ILM.

That infrastructure consists of three tiers. High performance disk is used for vital business applications including SAP and Oracle. Mid-tier systems and network-attached storage (NAS) is used for managing quality assurance and the testing and development environment. And a third, Advanced Technology Attachment (ATA)-based tier handles continuous 24/7 archiving.

Now, WHSmith News completes a full back-up of its data every day, which amounts to 21TB. Rather than storing those 21TB on disk, it is moved to tape and sent off-site for more cost-effective long-term storage. “By deploying a tiered infrastructure and automating the movement of data between tiers, our information has a lifecycle of just one day,” Wilson explains. “We have an architecture that provides complete availability and performance of our key applications, while simplifying information management.”

“The results have exceeded all expectations,” he says. “We have seen our storage costs cut by 50 per cent from the deployment of systems and software that meet our exact requirements. Furthermore, we had planned for a four-year return on investment, but will deliver a year earlier than planned.”

For WHSmith, that is very good news indeed.

The perils of primary storage

Companies are incurring unnecessary costs by keeping data on primary storage – and the situation is getting worse, according to an annual ILM survey conducted by storagemanagement software vendor BridgeHead Software. In 2005, only 16 per cent of UK respondents had more than five terabytes (TB) of online data. In 2006, this had grown to 22 per cent, an increase of more than one-third. Similarly, those storing in excess of 10TB also jumped by 30 per cent.

The degree to which respondents believe primary storage is consumed by redundant data also continues to increase. Some 18 per cent more respondents than last year believe they could recoup more than half of their most expensive primary disk space by removing unwanted data. While continued data growth is not surprising in general, the year-to-year comparison captured a clear discrepancy between how different types of data are driving this growth.

Both in the US and UK, the growth of online data is dominated by unstructured data – typically end-user files created outside of an IT-managed application. There has been an 81 per cent increase in respondents who calculated that more than half of their data was unstructured files.

“While on the one hand, this dramatic increase may be a leader of greater corporate productivity, it is a potential nightmare for IT managers,” says BridgeHead CEO Tony Cottrill. “User files tend to be created and stored on network file shares without the benefit of IT’s most structured management and control systems, making them hard to find, manage and correctly store for compliance and data protection,” he adds.

Encouragingly, the ILM survey reveals that businesses have recognised the need to tackle the challenge posed by unstructured files. In answer to the question, “What are your key storage-related areas of interest in the next 12 months?”, file archiving made the biggest leap forward compared to the responses in 2005. An almost 30 per cent increase in UK respondents listed file archiving as a key area of interest, whereas e-mail archiving registered a lower increase of 16 per cent and disaster recovery only increased by four per cent.

However, despite the growing awareness of file archiving, just six per cent of UK respondents said that it was likely that they would be unable to find and restore a three-month-old file and 27 per cent said they did not know how long retrieval of a three-month-old file would take.

“The continued lack of confidence in long-term retention as the apparent adoption of archiving rises leads one to wonder whether the archiving strategies being adopted are effective or if there isn’t a more acute awareness that better tools are needed,” says Cottrill. “Even with almost 85 per cent of companies claiming to archive in some way, the fact that data on primary storage is still growing at 25 per cent minimum indicates that the most effective archiving tools and technologies are not yet as prevalent as they should be,” he says.