By Salvatore Salamone
June 17, 2005 | The daily data backup challenge faced by pharma Boehringer Ingelheim is one many large organizations struggle with. Put bluntly, now is never the right time for someone in its far-flung operations around the world.
“We work around the clock. No matter what time you picked, you would [disrupt] someone’s work,” says Jörg Werner, system analyst at Boehringer Ingelheim Pharma GmbH & Co. KG. Nevertheless, backing up the company’s 1.3TB, 1.5-million object Documentum repository is critical. It holds files associated with many aspects of its worldwide clinical trials and drug submission activities.
In an effort to minimize disruption, Boehringer recently began using a new class of backup tool known as “application aware.” Specifically, it chose HOTBackup from CYA Technologies, a product that ties directly into Documentum. Backups are now performed without taking the database offline.
The new approach is also soothing bruised egos, Werner says. Roughly 3,500 people use the worldwide system on a regular basis. Before implementing HOTBackup, Boehringer experimented with compromise solutions, conducting backup at times when it affected the fewest number of employees.
Asia-Pacific had the fewest number of system users, so every day from noon to 2 p.m. users in that region were booted off the system while it was taken offline for the daily backup. “People felt slighted,” Werner says.
Going Back in Time
Working with Phoenix Systems Integration GmbH -- an integrator with expertise in developing Documentum-based content management systems for regulated enterprises -- Boehringer settled on HOTBackup as a solution to its backup problem.
HOTBackup, like other application-aware storage software, works directly with the program generating the data that must be backed up. HOTBackup uses a Documentum application programming interface (API) to access information while the database is still online. It saves any changes in a Documentum repository at user-defined intervals.
The software can be used to roll back an entire Documentum system to a “point-in-time”-style restoration in which everything from the files and the relationships between files is re-created. The user encounters no noticeable impact (i.e., performance degradation) while it is running. The software simultaneously captures content and metadata (e.g., annotations, version trees, and workflows) associated with the content. The starting price for the software is about $50,000.
Potential productivity gains are also substantial. By eliminating the need to take the Documentum system offline, previously affected Boehringer facilities gained back two hours of system access per day, or about 500 hours per year per person, Werner says.
Now, backups are typically done every three hours without disrupting employees’ work. In addition to performing live backups, the HOTBackup software also offers improved restoration services. Due to its knowledge of Documentum, “you can restore one document with all its relationships without having to do an entire backup,” says Ernst Lange, Phoenix’s director of European operations.
This type of software is relatively new and not very common. But that may not be the case for long. “The trend today is toward more intelligent storage software,” says David Wilson, an independent IT consultant who specializes in the backup and availability of enterprise applications and corporate data. “We’re already seeing a tighter coupling between the applications and storage hardware systems with the advent of content-addressable storage. There is also a move toward policy-based data handling.”
Other experts agree. “There is a new records-management paradigm today,” said Mike Casey, principal analyst at the IT consultancy Contoural Inc., at the recent Storage Decisions conference in Manhattan.
Casey noted that with policy-based management, rather than using a simple creation date or last-accessed date, information about the actual content (its value to the company, its importance in meeting compliance requirements) is used to decide how data are handled long-term and on which medium they are stored. “You end up saving more data, more intelligently,” Casey said.