Information Lifecycle Management: Being Bought vs. Being Built

Return to Careers Page

What began as a noble idea and an appropriate (if not complex) objective ended up being somewhat denigrated as simply the way for storage vendors to sell more of their wares. ILM itself became a tarnished term and deservedly so since not only were there few points of integration among the various capabilities and methods of information collection, management and distribution, vendors (e.g., EMC, IBM, CA, HP) promulgating the wonders of ILM often discussed little more than their latest storage gear. The good news, however, was a growing recognition of the need to drive the movement and management of information relative to regulatory and business policies that would govern the lifecycle of that information. But while policy administration related to information management certainly entered the spotlight, each capability, from capture, to retention, to archive, to distribution, etc. maintained its own method of establishing and enforcing policy, often because a separate application governed each element. The lifecycle of information from its time of creation through its use typically remained in one world (see Figure 1), while the lifecycle of information from the first time it is physically stored through its long term archive and/or destruction was in another. Despite increased integration through web services architectures, this issue has remained fairly constant.

Additionally, few end users and fewer vendors realized that a key not only to unifying ILM but to succeeding in appreciating the lifecycle of information is to have a clear understanding of the value and/or risk inherent in any piece of information. Without this knowledge, it is impossible to establish an appropriate lifecycle entailing correct policies for retrieval, retention, destruction, or tiers of storage. While there is general acknowledgement that this is difficult to do since a knowledge of information’s metadata and the actual value or importance of the information (which is not typically captured in metadata) are required, several companies have cropped up to address this, and numerous others (i.e., in the search/retrieval arena) could apply their existing classification/categorization capabilities to it as well.

Now, though, as a result of fervent vendor consolidation – within and between the ECM, storage and security worlds – it may finally be time to consider when (vs. “if”) ILM will be a reality. Consider the following recent events:

  • EMC makes acquisitions in systems management and security (with RSA being the latest) after having acquired several ECM and other software companies
  • CA acquired MDY, a records management provider that will likely be paired with iLumin; it has also just acquired XOSoft, a continuous data protection vendor
  • IBM acquires FileNet and continues to integrate capabilities across its Tivoli and WebSphere software assets while making further investments in metadata management, integration, etc.

Concomitantly, Open Text (now acquiring Hummingbird), FileNet (who acquired Yaletown Technology Group), and others in ECM have built and/or bought various elements to drive ILM or ILM-like product suites. Additionally, there are numerous information classification vendors (e.g., Kazeon Systems, StoredIQ, Njini, Abrevity, Arkivio) that, like their search-oriented counterparts (e.g., FAST, Endeca, Autonomy, Convera, IBM), specialize in extracting concepts from content to understand the meaning of information and possibly its value given that meaning. While these companies remain independent at this point, some have already established partnerships with storage vendors to assist their ILM build-outs.

There are several impediments to the creation of a fullblown ILM capability set. These include the following:

  • Not owning all the pieces as listed above
  • Not integrating the pieces that are owned or available through partnerships
  • Inability to extract the nature of the value or risk associated with information and using this data to feed policy rules (and thus lifecycle decisions)
  • Lack of robust (and often consistent) metadata across all information types
  • Lack of an integrated policy engine or ombudsman that enables rules to be created, instantiated and administered across all information management components
  • Inability to transfer appropriate data even among component areas (for example, within storage among backup, disaster recovery, and on-line storage) relative to policies and usage

While the ILM panacea won’t appear any time soon, it’s clear that the vendors realize that it isn’t simply a euphemism for hierarchical storage and have begun to understand more clearly that ILM must address both structured and unstructured data and be able to link the rules that govern information during its active use as well as its long term retention. Sarbanes-Oxley and other regulations have given end users the desire and (often) fiscal capability to manage across information types and platforms more deeply, and ILM can ultimately be a boon to such demands. The good news is that there is enough technology addressing each component part of ILM to enable several enterprising vendors to at least amass a suite of capabilities – even if integrating them cleanly is several years off.