Press Releases

Why Most Large Enterprise Archiving Projects Fail

ZL Unveils the Hard Questions to Ask Vendors and Avoid Failure in Large Archiving Projects

Published by: Cloud Computing Journal

SAN JOSE, CA – Apr. 7, 2009

The task of electronic records retention and management in large enterprises is still relatively new and the challenges are daunting. The sheer volume of data far exceeds any other application in the large enterprise, with the number of e-mails being generated in, say, one year, typically exceeding the total number of documents in the U.S. Library of Congress.

In addition to the challenge of scaling to handle such staggering data volumes, the archiving system must also be flexible enough to accommodate the fast-changing archiving requirements of the large enterprise, which have evolved rapidly to include storage reduction; regulatory compliance; litigation support; and records management.

Lacking in scalability and flexibility, the archiving products available in the marketplace are mostly unable to deliver the performance required in large records management projects. For these projects, failure rates have reached 90% and over. To minimize the risk of failure, ZL Technologies has collected below a series of key questions which relate to breaking points encountered in many of the failed projects.

Of the many hurdles facing archiving, the most critical issues relate to the search engine. Archiving, in the most fundamental essence, is not about storage, but about finding. Without the ability to find, the storage capability is rendered meaningless. To assess the search capabilities, these are three of the key questions to ask of all vendors:

  • How fast is the search across all mailboxes?
    Due to weak search engines, most archiving solutions try to narrow the search to a small portion of the data, such as 50 mailboxes or “custodians,” because a complete search could take days. However, escalating requirements make this approach of limited searching unacceptable. “The traditional search by selected custodian mailboxes alone often is no longer adequate. eDiscovery has developed to the point where, for many, the ability to search across mailboxes in a quick and timely manner has become a critical requirement,” says George Socha of Socha Consulting, Inc., a leading eDiscovery consultancy. “The shortcomings of a limited custodian search can be substantial. For example, searching custodian mailboxes for messages containing attorney email addresses in the To:, From: and Bcc: fields may completely miss those e-mails involving attorneys which were subsequently forwarded to other people. With forwarded e-mail, client-attorney privileged information could be buried in the body and may well not be found with a custodian search. A search-all-mailboxes capability can highlight these privileged documents and help reduce the chances of inadvertent waiver of privilege.”
  • How accurate is the search?
    Search accuracy is just as important to eDiscovery as search speed, especially in finding the most relevant documents and in reducing the cost of legal time to review search results. Search capabilities such as proximity search (finding a word within n words of another) are critically important to reduce false positives. It should be noted that many well-known archiving products are not able to do proximity search. A related but equally unsettling flaw is inaccurate or inconsistent searches whereby search results can change depending on the order of search words. For example, searching for “promise” within 5 words of “stock” could return different results from “stock” within 5 words of “promise.” With such inaccuracies, the archive runs the risk of non-responsiveness and inadvertent liability exposure such as privileged waivers.
  • Does the archive’s search engine have a future? Or will it trigger punitive costs ahead?
    The life of a search engine plays a critical factor in the success of the archive. Some archiving products use an obsolete or end-of-life search engine, such as AltaVista, where a replacement will soon be necessary. However, when the replacement does happen, users will either be saddled with a massive cost of data migration to the new search engine, or encumbered with the high cost of running two separate archives. These hidden costs should be factored in the purchase decision, as they could exceed even the original acquisition cost.

“The above points represent only a few of the many core elements which impact the chances of success in archiving implementations,” said Kon Leong, ZL’s CEO. “Other key elements include assessing the real cost of ownership and the quality of support for what is a complex application. The archiving market is still in its developing stages and there needs to be a continuing education on these critical issues.”