Help the Aged Data
ILM - the new black
ILM, or Information Lifecycle Management as it likes to be known formally, is employed in the areas of storage and storage management. At heart ILM is the management of tiered storage to provide cost effective, robust management of data as it ages. And it is becoming an increasingly popular TLA (three letter acronym). Why? As everybody knows, the volume of data generated in everyday business is mounting at a pace that exceeds even that of the increasing capacity of storage systems. The data itself and the information it holds are of enormous value to businesses; and it is vital that it be held safely.
"Compliance and regulation" represent other factors that now routinely oblige organisations to ensure that considerable proportions of their data be retained for extended periods of time. Indeed, in certain industries, most notably the financial services sector it is common for it to be necessary to keep many records for anywhere between a minimum of six years and, at least in theory, for time without end.
This is where ILM fits in. Few organisations have the financial resources to buy enough front line storage to keep all the data they create indefinitely, or for that matter for a mere six years or so. As any experienced storage manager will tell you, most data is rarely accessed after the first few weeks of its life. When it is new, it is likely that users and their applications will require relatively quick access to the data. However, as it ages, the number of instances when it is viewed tends to fall quite rapidly. Indeed several research studies have shown that more than 90 per cent of data is never accessed more than three months after its creation.
Business needs or regulations may then demand that the data is retained, rather than deleted. Here there may be the opportunity to migrate this middle-aged data away from the high performance, highly available expensive storage hardware on which it was originally created and move it to a cheaper, perhaps commodity-based platform. As the data ages, it may become possible to migrate it once again to an even more cost-effective archive repository.
For those with a mainframe background this may sound very familiar as in many respects this process can be viewed as the modern day equivalent of hierarchical storage management, or HSM.
In order to implement ILM there are clearly a few requirements, namely a storage platforms that is flexible in deployment, that offers a range of performance and price characteristics, and that is capable of supporting multiple servers, operating systems and storage network topologies. Also needed is good storage management software to track, migrate and archive the data. Such software is now becoming available as storage standards mature and as the software and hardware vendors make their systems better at working in heterogeneous environments.
But there is also a need for the storage administrators to understand what data exists in their environments, what applications use it and to understand the business requirements dictating the service levels that data access needs to fulfil. Only then can the storage administrator address the compliance and regulatory requirements for the long-term retention of each class of data. This presents a considerable challenge for almost all organisations.
ILM is feasible today in many organisations, but it will probably not be applied uniformly. Instead it will start in business areas where there are strong drivers, especially compliance and regulation, or where there is exceptional business value to be had. ILM has a value and, with some effort, it can be made to work. It is to be hoped that ILM will not be a fashion for long, but will instead evolve into a routine mainstream operation.
Sponsored: Benefits from the lessons learned in HPC