Cybrain - Fotolia

Penn State Health seeks centralized backup, picks Commvault

As Penn State Health's data grew across its vast network of hospitals and practices, it turned to Commvault for centralized backup and a way to archive petabytes of research data.

As Penn State Health, also known as the Milton S. Hershey Medical Center, grew into an expansive health services organization through acquisitions, its IT complexity also grew.

Penn State Health formed in 2014 to consolidate regional health services organizations. Through acquisitions, the Hershey Medical Center has since grown  into a network of hospitals throughout Pennsylvania. This led to the organization's data protection responsibilities extending beyond Hershey, Pa.

Cory Heikel, systems engineer at Penn State Health, estimated between 50 and 100 separate entities joined the organization's network. Those entities were truly separate, from an IT perspective.

Charged with protecting massive amounts of research data and the clinical data from its network of hospitals and practices, Heikel and his team sought centralized backup.

"Everybody was their own little island," Heikel said.

Cory Heikel, systems engineer at Penn State HealthCory Heikel

He said he worried about managing the IT and data protection for all of those offices, each containing important medical information the organization could not afford to lose.

Heikel said some groups in the network were small, with little or no IT resources of their own. It made no sense for these remote offices to hire IT staff, but Penn State Health still needed a way to ensure important medical data was being properly protected.

"These people are doctors and nurses, not IT people. We want to make sure we can handle it from our central location, and we've got enough bandwidth to transfer the data on a nightly basis," Heikel said.

Backing up to a public cloud was one option for centralized backup. However, due to the nature of the data, Heikel said it was too big of a security risk to implement that for Penn State Health. A breach would be outside of Heikel's control, and it would ultimately be his responsibility.

"We're not even really considering public cloud at this point just because of the security implications," Heikel said. "If there's a breach at Amazon, it's not just their breach; it's ours. We're the ones in the news."

After using IBM Tivoli Storage Manager (TSM) -- now IBM Spectrum Protect -- backup software for 20 years, Penn State Health began switching over to Commvault.  Heikel said Commvault allows him to combine secondary storage functions, such as archiving, with his backup system.

Penn State Health began using Commvault to archive research data from the Penn State College of Medicine and now uses it for most of its back up data.

Heikel is testing a 14.5 TB Commvault integrated appliance at one of his organization's remote offices, which is replicating to the data center at Hershey Medical Center. If this proof of concept shows he can monitor and fix things from Hershey, Heikel will deploy it to other practices in the Penn State Health network.

If there's a breach at Amazon, it's not just their breach; it's ours. We're the ones in the news.
Cory HeikelSystems engineer at Penn State Health

"We're going to get one into a site, make sure it works the way we think it will, then start rolling them out to different areas," Heikel said.

Heikel said Penn State Health has about 6 PB to 8 PB of research data to protect, with about 400 TB generated daily. This data includes genomics research data and electron microscope images. On top of this, the organization also protects 1.5 PB of hospital-related data, which consists of medical images and financials. Patient medical records are hosted separately by Cerner Corp.

Heikel said he considered deploying another TSM system and put all the research data there. However, when shopping around, he learned about Commvault's archiving feature, which is built into its backup product. With IBM, he would've needed a separate product to do archiving -- and another round of data ingestion.

"Archiving in Tivoli is another product, runs on its own server and has its own storage pool. We'd have to reingest to an archive," Heikel said.

Heikel has been transitioning Penn State Health off of IBM and estimated he is about halfway there as he moves to a centralized backup process.

Despite the plan to ditch IBM entirely, Heikel said he missed TSM's file-based system, as opposed to Commvault's job-based system. He said he would like the ability to go into backups and selectively delete files he knows he won't need, immediately freeing up disk space. Commvault doesn't delete until the job is finished, causing some delays until that real estate is available again.

Still, Heikel said that wasn't a deal breaker for him, and he's having better results with Commvault.

Dig Deeper on Archiving and tape backup