Can Backup be a System of Record?

By: In: Data Management On: Mar 23, 2011

It is often wiser to adapt our systems to what people do in practice, than the other way ‘round.

A big challenge with implementing a centralized authoritative repository representing an organization’s System of Record is that this is often not practical. Many traditional Enterprise Content Management (ECM) implementations failed because of the enormity of the project and the complexity of the implementation, not unlike big ERP and Enterprise Management projects. They put an excessive burden on IT and have required IT intervention for minor changes. ECM systems were often intrusive and required users to change the way they work, causing widespread user dissatisfaction. ECM systems as a result are best suited for the most critical business documents, often in specific departments (Finance, HR, etc.) of the enterprise with well-structured processes or in highly regulated industries (e.g. pharmaceutical).

The need for flexibility and to tap widespread user support across the global organization has moved organizations away from centralized repositories and towards decentralized distributed repositories unified by a holistic policy management approach.

The growth of unstructured information, particularly email, has resulted in the emergence of online archiving solutions as a System of Record. Archives are particularly effective in handling billions of small unstructured items such as email.  As archiving solutions do not include collaboration and document management capabilities and integrate with email and file systems at the back-end, they do not impact end-users. The main focus of archiving has been for compliance and eDiscovery. Archiving solutions typically provide a rich set of capabilities for retention management, monitoring and supervision, search, legal-hold and early case assessment. With the growth of unstructured information and the use of expensive storage for production email and file systems, a secondary focus for archiving is to offload production systems. This is done by moving data that is inactive to the online archive (and deleting the production copy) or by stubbing the data (leaving a pointer in the production system to the data in the archive).

Online archiving of email has become a widespread practice, but the use of archiving systems for files and documents is still limited. The pervasive use of email for communication makes it a frequent target of internal investigations and external eDiscovery requests. This often results in the implementation of email archiving for all users in the organization. In many organizations however, the cost of moving large numbers of files and documents into an archive and then storing another copy is considered prohibitive relative to the perceived gains. File systems also scale a lot better than email systems do, making file archiving less of an urgent problem.

This brings us to backup. Any data that is important for the business has to be and is backed up. Backup has been established as a standard practice for all applications and file systems. Why not leverage this existing and comprehensive repository of information, which is kept safely and securely, for information governance? Why not bring the information governance capability to where the data resides, rather than moving the data to this capability?

Guess what? We know from experience that organizations are already using backup as an archive. Despite the admonishments of industry experts, tape has always been and continues to be used to meet archiving needs. The configuration of long term retention periods for a fraction of data on disk and cloud backup is an indication that it’s being used for archival purposes. eDiscovery requests often result in many days or weeks of manual effort by IT searching backup for relevant data.

Data protection solutions can be enhanced to make a backup repository a System of Record and address information governance needs. Next generation data protection solutions can incorporate archiving by indexing and classifying data as it is being backed up. This will support search for litigation and compliance. Adding sophisticated policy management, specifically granular policies for retention and disposition of data and policies for deletion of inactive source data will help address archiving needs.

Backup is changing in very significant ways. Will backup be busy adapting to the new virtualization and cloud environments? Or will it have a chance to incorporate archiving features and become a System of Record? Is this an idea whose time has finally come? What do you think?

← B2B Cloud – it’s about Security, Compliance, SLA, and Information Advantage You’re in the minority if you’re backing up data just for recovery →


  1. josephmartins
    March 23, 2011 at 4:15 pm

    Many thoughts come to mind on this topic. I’ll share a couple here.

    Storage technology must evolve. It must become significantly more context-aware. IT is inefficient because today’s IT hardware, software and processes rely on guesswork and relatively superficial context. One of the many reasons backups cannot serve as systems of record is the utter lack of context. A backup can’t tell you much about its protected assets or how those assets relate to any other assets in the organization, or why they’re important. Today’s backups are, for lack of a better term, ignorant.

    Today’s archives aren’t significantly better. In fact, most of the tools used by IT today are ignorant about information and treat it as collections of bytes, files, folders and volumes.

    What needs to change? Just as we’ve told hundreds of companies over the past ten years, business applications of all types–from BI, CRM, DAM, DIS, DSS, ECM, ERP, GIS, HALDB, LMS, MIS, PLM, OLTP, RDBMS, RM, SCM, WCM and more–must begin sharing their rich contextual information with IT’s systems to improve the efficiency with which customers deduplicate, discover, migrate, protect, preserve and destroy information assets.

    Concepts such as backup and archive will very likely converge as information is increasingly managed in-context, throughout its lifecycle, all the way down the technology stack from user apps to media. Add the right security and controls and just about any resource can become a viable “system of record”.

  2. TM Ravi
    T. M. Ravi
    March 25, 2011 at 2:26 pm

    Joe – Thank you for your detailed comment. I totally agree with you on the importance of capturing data, metadata and rich application context for backup and archives to be considered viable “systems of record”.

    I do think that some email archives do a good job of “full capture” and maintain data and context. This is a barrier that backup has to cross.

    I agree with you that storage software must evolve to become more context aware. I feel that storage hardware should continue to work at the byte, file, folder and volume level – managing governance policies at the device level would be difficult and expensive.


  3. josephmartins
    March 25, 2011 at 3:37 pm

    Email archives are an excellent example of what could be achieved elsewhere.

    Though I believe eventually email archives will cease to exist as islands and perhaps live within a greater context as well. For example, as part of a larger initiative that preserves email/text/IM conversations as they are related to other information assets that live outside email/SMS/IM (i.e. not as attachments). You and I spoke about this several years ago and the industry still isn’t there yet.

    I agree about the hardware. Though there are instances where additional context might prove beneficial if it can be accomplished in a way that isn’t counterproductive (e.g. intelligent switching, purpose-built blades, etc). At this point, DMG would be thrilled just to see improvements in the contextual awareness of storage applications.

Leave A Comment


About the author

TM Ravi

T. M. Ravi is chief marketing officer for Iron Mountain responsible for marketing and strategy for the company’s cloud, on-premises and hybrid information management solutions that span data protection, archiving, eDiscovery and compliance. Ravi joined Iron Mountain through the acquisition of Mimosa Systems, the leader in enterprise content archiving, where he was founder, president and chief executive officer. Before Mimosa, Ravi was founder and CEO of Peakstone Corporation that provided performance management solutions for Fortune 500 companies. Prior to his role at Peakstone, Ravi was vice president of marketing at Computer Associates (CA), Ravi he was responsible for the core line of CA enterprise management products, including CA Unicenter as well as the areas of application, systems and network management, software distribution, help desk, security, and storage management. Ravi joined CA through the $1.2 billion acquisition of Cheyenne Software, the market leader in storage management and antivirus solutions. At Cheyenne Software, he was the vice president responsible for the company's successful Windows NT business with products, such as ARCserve backup and InocuLAN antivirus. Earlier in his career, Ravi worked in Hewlett-Packard's Information Architecture Group, where he did product planning for client/server and storage solutions. Ravi earned a MS and PhD from UCLA and a Bachelors of Technology from Indian Institute of Technology (IIT), Kanpur, India.