Data protection is broken, and IT organizations need new, software-defined solutions, said Wikibon Chief Analyst David Vellante in a recent CUBEconversation (see video below). Data is growing at exponential rates, driven by the Internet of Things (IoT), machine-generated data and new Systems of Insight that combine transactional data with analytics. The 24x7x365 cloud business environment leaves no time for traditional backup windows, and the complexity of cloud business environments with public, private and hybrid clouds creates new issues for data recovery.

On top of this, backup is still usually a one-size-fits-all world, often a daily incremental with possibly a weekly full backup of the database. Because this approach is designed to meet the average needs of all the applications in the organization, “by definition applications are either under-protected or overprotected.”

As a result, backups are failing. The IT organization has no opportunity to test recovery, so it cannot guarantee recovery from a failure or even from lost data. “The cost and complexity is driving IT practitioners crazy,” Vellante says. “By the end of the decade, 70 percent of organizations will have to re-architect their backup methodology.” That is equally true for large enterprises and smaller companies, as well as government entities such as school districts. They all have irreplaceable data that needs protection.

Modernizing DR

A major problem preventing effective data recover (DR) system rearchitecting is that many IT organizations do not know what they have or what they need, said CAS Severn, Inc.,  CTO and VP of Technology Services Joseph King, @jckingmd.

“Most clients cannot tell you how much data they are backing up and how successful those backups are,” King said. His organization “collects the hard data: How long does the client need to store the data? How successful are their backup cycles? What are the regulatory requirements?”

Often IT’s answer is to try to keep everything forever. But data that is kept beyond its useful lire and beyond the period required for compliance exposes companies to legal risks. And it wastes resources that can be in short supply, particularly when tape is no longer adequate.

Case history: Charles County, MD, Schools

These issues aren’t confined to large enterprises – they are just as important for smaller companies and government entities, including public school systems, King said. A typical case history is the Charles County, MD, school system, with 41 schools with 27,000 students and a $300 million budget. Schools, like almost all entities, run on data – student demographics, health records, special needs and test scores, to name just a few. Some of this are irreplaceable, and a data loss can create major problems for the students and for compliance with various state and federal regulations including No Child Left Behind.

One growing issue with its existing tape-based backup systems was lengthy recovery times, said Charles County, MD, Public Schools CIO and Director of Technology Bijaya Devkota on the CUBEconversation with Vellante and King. When recovery was needed, “We had to tell the school principal it would take a few days to get the data back and a week to recover an entire site.” Because backups were not constant, there was also the risk of data loss. Neither the recovery time objective (RTO) nor the recovery point object (RPO) met their needs.

CAS Severn’s answer to the problem was a customized, software-defined distributed backup and recovery system, replacing tape with disk, using existing hardware wherever possible and based on IBM Spectrum Protect. CAS Severn adopted this system because of its flexibility and scalability over previous backup appliances.

“For us this has been a 38-year journey,” King said. ““We wanted to go to market with one solution, not three or four different ones.” Spectrum Protect met those needs. It can run on premise, in multiple locations and, if the customer wants to do their backup to an infrastructure-as-a-service platform such as Amazon Web Services it can easily be loaded onto the platform and run from there. It can run on bare metal or in a virtualized environment. And it can scale to meet the needs of organizations ranging from school systems to large federal government agencies and enterprises.

For Devkota the system provides two primary and interlinked advantages. First, because it operates disk-to-disk, it maintains the original data format of all the school’s data. Second, because the backup is always available over the network from the remote site, recovery is an order-of-magnitude faster. Recovering small amounts of data takes minutes or at most a few hours. Recovering an entire school can be accomplished in a day as opposed to a week with tape backup.

Ease-of-use feature means  normal operations require no manual intervention, King said. Recoveries can often be done by the users themselves rather than requiring expert intervention or calls to CAS Severn. This is particularly important for distributed organizations, such as school systems, that can’t keep IT professionals at every site.

“If something’s easy to use, you’ll probably actually use it,” King said. “Users don’t get that paralyzing feeling when … they’re scared of what they have.”

This is another important advantage for the school system, where educators rather than IT professionals have to manage the backup and recovery of their school’s data. “The system has evolved to the point that our users are happy, and we did not have to add any new resources,” Devkota says.

“When large school systems lose student data, you read about it in the newspaper and see it on CNN,” Devkota said. “You don’t want that.” Spectrum Select gives high level assurance that backups are successful and test restores are possible, as well as the flexibility to size to the user’s need and adjust as those needs change.

photo credit: Peter Kurdulija via photopin cc