By Kevin Komiega
When InfoStor examined the pros and cons of content-addressed storage (CAS) technology one year ago, users were singing the praises of CAS platforms and their ability to securely store and retrieve unaltered data for long periods of time. But, upon recently revisiting the topic with some users, one can see that scalability, performance, pricing, and vendor lock-in issues are starting to crop up in the field-issues that may give some smaller CAS vendors a chance to make inroads into some customer accounts.
Part of Hal Weiss’ job as a systems engineer for Baptist Memorial Health Care is to plan the direction of his organization’s server and storage architectures. Weiss and his team manage IT resources for 15 hospitals, 28 clinics, and two home healthcare providers across western Tennessee and northern Mississippi, with data centers in each hospital.
“Healthcare is growing like crazy and healthcare organizations, as a rule, don’t throw anything away, but there’s a very good reason for that,” says Weiss.
Patient information, medical images, and scanned documents are just some of the types of data Baptist Memorial needs to keep handy. “You always have to have some place to store it,” says Weiss.
As a result, Baptist Memorial has a wide array of storage systems from a multitude of different vendors, including a pair of EMC Centera CAS systems.
But recently, Weiss and his team had been experiencing the pains associated with scaling their systems and supporting new applications.
“The problem with some platforms is that you’re locked into their hardware and the performance [is not good],” says Weiss.
Weiss has also run into problems with getting the CAS systems to support more than one application. His Centera systems support the organization’s PACS imaging system and serve as the archive for patient records.
“We ran out of object space before we ran out of physical space,” says Weiss.
The cost of buying disk capacity had also become a concern, which prompted Weiss to begin evaluating a new CAS software solution from Caringo.
Caringo’s fixed-content storage product, CAStor FSG, is a software-based platform that is hardware-agnostic. CAStor FSG enables administrators to set policies, permissions, and constraints for fixed content to meet internal goals and regulatory compliance requirements. Once the policies are configured, CAStor transparently monitors data and enforces the policies, ensuring the correct number of replicas for a particular file, verifying file life-cycle information, and managing secure deletion. The software is installed from a bootable flash drive and supports industry-standard hardware.
Eventually, Weiss hopes to transition to the Caringo system for all of his archiving needs in an effort to achieve greater flexibility in scaling and managing his data. “Some applications have to be specifically written to take advantage of Centera or RISS. That requires going back to the vendor to see if they can help you write the scripts,” says Weiss. No so with Caringo, he says. “The scalability and performance of the Caringo approach is very attractive, and the cost is much less than everyone else’s offerings.”
The CAStor software installs from a bootable USB flash drive that can be plugged into any computer. Nodes can be added to form a cluster. Other options include a network-boot capability and remote replication.
Deciding between a tier-one vendor and a start-up is one of the oldest dilemmas of the technology age. Big companies are established and tend to have more proven products, while new companies can sometimes offer lower prices, but often operate under a cloud of uncertainty.
“It’s not a secret that CAS is usually expensive and proprietary,” says Mike Karp, a senior analyst with the Enterprise Management Associates consulting firm, “and a vendor like Caringo makes it easy for users to get competitive pricing on supporting hardware.”
CAS and tiered storage
EMC customer and Centera user Pat Zinno, director of infrastructure services and support for New Jersey-based non-profit Atlantic Health, has more than 100TB of managed storage on a variety of EMC systems.
In a given year, Zinno handles more than 68,000 admissions and 623,000 outpatient and 151,000 emergency department visits. “If we kept all of our data on a high-availability platform, the cost of managing it all would quickly overwhelm our budget,” he says.
Atlantic Health teamed up with EMC to adopt an information lifecycle management (ILM) strategy with tiered storage-including CAS-to automate the data archiving and removal process. “We’ve achieved dramatic costs savings [more than $600,000],” says Zinno.
Atlantic Health stores radiology images and patient records on Clariion systems at each of its three main hospitals. Using EMC’s DiskXtender software, images are automatically archived after 90 days to EMC’s Centera CAS storage at the main data center. Atlantic Health also uses EmailXtender to archive Exchange e-mail older than 90 days from Symmetrix to Centera platforms, where it stays for three years before being deleted.
“We looked at solutions from NetApp, HP, and Hitachi,” says Zinno, “but we wanted to standardize on a single platform. Centera fit into the realm of all of the HIPAA and compliance regulations we have to deal with for everything from clinical image retention to e-mail archiving.”
But Zinno says some of the same archiving issues he had two years ago still linger. “There is no rock-solid software management, but it’s not so much the Centera. It’s partly due to the application vendors. They don’t like to commingle.”
The sheer growth of Atlantic Health’s data has prompted the organization to buy more capacity for its dual Centera systems, each of which has about 31TB.
Atlantic Health uses EMC’s SRDF/S software for real-time synchronous data replication between Symmetrix systems located at the two data centers 20 miles apart. The healthcare network also mirrors the Centera archives between the data centers using Centera replication software.
CAS means compliance
Healthcare is an obvious sweet spot for CAS technology, but financial services is also dealing with stringent regulatory compliance issues and is turning to CAS to meet its archiving needs.
For example, Leesport Financial Corp. was searching for a storage solution to help retain its e-mail data for a federally mandated term of seven years. “We fall under a lot of different regulatory statutes that we have to comply with, and that prompted us to make a decision on an e-mail archiving solution,” says Jack McLaine, MIS manager at Leesport.
Leesport must comply with numerous regulations, including the Check 21 Act, Sarbanes-Oxley, the Gramm Leach Bliley Act, SEC 17a-4, and NASD 3110.
When planning for capacity requirements, McLaine built in a 20% to 30% allowance for overall capacity growth, but those numbers went out the window when Leesport acquired another banking company, bringing in hundreds of new employees and an influx of e-mail that overwhelmed his archiving infrastructure.
McLaine evaluated EMC’s Centera and Permabit’s Dynamic Information Services platform. “Both solutions would have met my needs with regards to growth of the e-mail archive, but I was looking to expand my storage solution to support other applications like document imaging,” he says. “I was concerned that Permabit was a fairly young company, but every company starts out young,” explains McLaine. “They were getting a lot of backing from other companies and that made me comfortable.”
McLaine opted to go with the Permabit product, which turns high-volume disk-based hardware into a storage system offering non-rewritable and non-erasable (WORM) capacity for long-term retention. Permabit’s Dynamic Information Services also supports standard NFS/CIFS interfaces for broad application coverage, a key consideration for Leesport Financial. And cost was not the primary concern because both the EMC and Permabit solutions were in the same ballpark, says McLaine.