This article can also be found in the Premium Editorial Download "Storage magazine: The 15 top storage products of 2004."
Download it now to read this article plus other related content.
ILM is a complex process; it's not just a matter of moving infrequently accessed information to inexpensive storage. It involves the greater challenge of addressing the overall holistic management of information from its creation to its demise. This requires an effective method of classifying data, but this is only part of the picture. Accessibility, archiving, migration, backup and recovery must also be addressed and integrated with corporate security and business-continuity frameworks. In addition, any effective ILM system must provide a mechanism for finding specific pieces of information based on indexed keywords.
The ILM process includes:
- Determining information classes based on the business importance of the information (critical to disposable), as well as security and privacy requirements (public, customer or confidential).
- Understanding when each class is created, accessed and retained, and when it can be deleted.
- Assigning a policy to each class of information that defines backup, recovery, accessibility and storage requirements throughout its lifecycle.
As if the concept of ILM alone didn't
While there's no immediate solution that will satisfy all ILM requirements, there are applications available that solve common information management issues for electronic messages, document repositories and databases. Sorting through them to find the most appropriate solution is the key.
When implementing an ILM system, users should look for three main functions:
- Mobility–moving information between storage platforms or tiers.
- Management–interfacing with the application managing the information.
- Organization–having the ability to track, index and retrieve the information.
There are also several key characteristics users should identify when assessing ILM tools. ILM products that move information around the storage hierarchy should be capable of maintaining metadata information so that the original application or user can locate and retrieve it. Most current ILM tools don't integrate with middleware applications and rely on the application that created the data to move the information to different storage tiers.
Users should note how information is managed to ensure it's protected and integrated with the file owner's application. For example, you should determine if the tool interacts with the information application so that its data continues to be backed up appropriately, or if it provides a pointer to the information for an audit trail or tracking. In addition, an ILM application should provide some structure and organization to the information. Ask how data is classified using the metadata and indexing so it can be searched and retrieved easily. Ideally, the application should let you find information, regardless of its location on different tiers of disk or on tape, using search criteria such as dates, keywords, topics or file type.
Unfortunately, a "magic bullet" ILM tool that correlates, monitors, tracks and manages all structured and unstructured information isn't even on the horizon. But there are many point solutions that are mature and effective at managing certain information pools. The most mature product areas in this arena are electronic messaging, document and database archiving and management. All allow for indexing and categorization of the metadata associated with the storage medium, and all can provide some insight into the information contained within messages and documents.
Tools that monitor, report and manage file-level information in Windows and Unix environments are also beginning to appear. These products are designed to view file systems, find old and infrequently accessed files, and to assist with some level of migration among storage tiers. Some of these products allow you to traverse file systems, review files and produce reports on stale files, orphaned files, files not accessed for a specified period of time, duplicate files, specific file extensions and more.
This was first published in January 2005