Fortifying school data backups
From hurricanes to software viruses to accidental keystrokes, many dangers threaten to corrupt school district data or impede access to it. To prevent loss of critical information, districts back up data routinely, on location and off-site.
Districts typically take a tiered approach, moving data between high-cost backup services that offer more immediate access and lower-cost solutions for archival storage. But new devices and lower-priced cloud offerings mean districts no longer have to trade access for cost.
“Solid state storage and bandwidth are so well scaled that data is more accessible whether it’s a local file or stored in the cloud,” says Mike Bronder, group vice president and managing director of Computer Resources, provider of the MMS student information system, and on-site and off-site hosting. “We’re in a sweet spot where there is parity between what the business requires and what the technology does.”
Pros & cons: Tape, disks, clouds
Most districts back up student, human resource and finance records and other essential administrative data every night. The medium of choice has traditionally been magnetic tape that is fairly inexpensive.
But tapes must be labeled and stored in boxes, and attempts to retrieve data that has been stored on the tape fail about 30 percent of the time, says Bill Andrews, president and chief executive officer of Exagrid, a provider of backup storage solutions.
Additionally, retrieving data from magnetic tape is time consuming. (However, published reports state that newer Linear Tape-Open [LTO] cartridges make it possible to retrieve data much quicker.)
Data stored on disks, in contrast, can be accessed more quickly, and disks are much more reliable for backups and restores. Recently, the cost of disks has come down, making faster backup and retrieval more cost-effective for districts, says Andrews, who estimates about 30 percent of districts have switched from tape to disk in the past few years.
Advances in “deduplication” software—which records only the data that has changed since the last backup—reduces the time and storage space required. Typically, only about 2 percent of data at the byte level changes per week, says Andrews. Deduplication ranges from reducing the amount of data required to be backed up from 6:1 to as high as 50:1, depending on the data type and the software.
Districts can also use software to automate the movement of data to on premises storage for quicker retrieval or to the cloud for archiving, based on parameters set by the district. The automation takes some of the burden off district IT staff, says David Mayer, practice director of Microsoft solutions for Insight Enterprises, a provider of hardware, software, and cloud solutions and IT services.
A typical example of hierarchical storage at a large district would be to:
- Back up files accessed day-to-day on a high-performance, on-site server for quick retrieval.
- Store occasionally-accessed files on a hard drive or disk.
- Put archival or rarely accessed data in the cloud or on tape.
A small- to mid-size district likely can store all long-term data—like student records that are no longer needed—on disks, says Bronder of MMS.
Multiple sites offer more security
As an added backup, many districts store data at an off-site location. Each night, Long Beach Public School District on New York’s Long Island replicates data from its finance system to a network at a regional Board of Cooperative Educational Services site less than a mile away. The district also stores tapes with backed up data, like payroll records, at the site.
Data from the student system is stored on a “storage-area network,” or SAN, and on tape. The SAN, located on a server in a district building, is as easy to access as a drive directly attached to the server on the district’s shared network, but does not take up bandwidth.
Long Beach feels confident in the system put in place following Hurricane Sandy in 2012: The district buildings that house the servers withstood the storm and the regional site—a good distance away—would not likely be affected by a power outage or most weather incidents, says Superintendent David Weiss.
Earlier this year, the district had a third party conduct a proactive risk assessment and test the district’s data backup and retrieval methods.
“The key is to regularly put in place protocols that require these nightly backups,” says Weiss. “And we are constantly looking at the whole process for backing up data.”
Putting greater distance between storages sites is an advantage of backing up data with a third party, says Ray Ackerlund, vice president of sales and marketing at Skyward which provides student information systems and archiving, backup, and disaster recovery solutions.
Ackerlund shares the story of a Texas fertilizer plant explosion in 2013 that levelled a district’s data center. The district needed to share information about students who had been relocated to neighboring districts and pay employees. If the district had stored the data in a cloud environment, leaders could have accessed information more quickly. Having a second site far away is a good idea, he says.
Moving finance and human resources data to the cloud alleviated the worry of servers failing at East Maine School District 63 in Illinois. Additionally, the district’s IT staff no longer have the task of backing up data every day, says David Bein, assistant superintendent of business services and chief school business official at the district.
Districts often choose a cloud provider because it frees up district staff from doing the maintenance and monitoring of on-site systems, performing backups, conduct data integrity audits and upgrading software, says Bronder of MMS.
An important consideration for the cloud is bandwidth. Sending or retrieving big chunks of information to or from the cloud can clog district bandwidth when staff and students need to access data and online tools. One option is to schedule the slow transmission of archival data up to the cloud over an extended period of time, like a month, Andrews says.
Expanding the footprint
In addition to data, districts need to back up the operating systems and applications that reside on their servers in the event they are incapacitated. Backing up and then restoring it all is easier when the servers are virtualized, experts say.
“Previously, a ‘server’ was a piece of dedicated hardware running one operating system and one or more applications,” says Doug Hazelman, vice president of product strategy at Veeam Software which develops backup, disaster recovery, and virtualization management software. “With virtualization, that one piece of hardware can now run 10 to 15 ‘virtual servers’, each having its own applications.”
Anne Arundel County Public School System in Maryland has more than 350 virtualized servers at its district, some of which host its finance and human resource systems. Using a software console, the district executes a full backup of all the software that resides on the servers each month.
And each day, it backs up any incremental changes. The backups are stored on a SAN and replicated to a disaster recovery site about 20 miles away, says Manish Patel, the district’s senior systems administrator.
“If a server fails, we can reboot our system and get access to the data much quicker than if we didn’t have virtualized servers,” says Patel. “Otherwise, it would take a long time to reinstall the operating system and all the applications and data.”
Whether districts are backing up virtualized servers or data only, technology options have gotten significantly more sophisticated. And having tape, disk or cloud backups at a second geographical site gives an added layer of protection.
“Each layer of the protection offers an exponential increase in recovery,” says Chris Cadieux, IT manager at MMS and a former district data manager, “with the odds of all recovery methods failing synchronously being less than winning the lottery."
Katie Kilfoyle Remis is a writer based in upstate New York.