The University of Wollongong, south of Sydney, has moved from a tape-based backup system to disk-based NAS dramatically reducing restoration times and allowing for inter-site disaster recovery.
Like most enterprises, the university’s data is growing rapidly and, as a result, the tape system was jeopardising the ability to successfully backup the data in a reasonable time frame, according to associate director of IT Joe McIver.
“We’ve been using tape for many years and, in our experience, recovering from tape takes a long time, so we had to start looking at alternatives,” he says.
With about 20TB of tape storage, the university procured NetApp NAS systems and migrated the tape infrastructure, running off Sun Solaris, to be used for permanent archiving.
One of the university’s challenges is a requirement to record research data which cannot be replaced and may need to be archived depending on its reuse needs.
With 10 NetApp filers housing between 25TB and 30TB of data, they mirror the data across to a disaster recovery site and each night get mirrored to a backup device in a third location.
This storage is used for backing up clients, servers, and Oracle databases.
Another project aimed at improving disaster recovery is virtualisation.
“We are in the process of migrating over 150 Windows servers into a virtual environment which should improve our disaster recovery capability,” McIver says.
“It will give us the ability with reasonable efficiency and result in less power, heat generated, and all the environmental matters which are important.”
The new infrastructure is not the first attempt by the university to move to disk-based storage. About five years ago it purchased cheap NAS devices with “devastating results” due to software bugs.
“NetApp would have been our preference at the time but we couldn’t afford it,” McIver says. “Now we could lose two sites and still recover. When it was tape we are talking days but on disk we are talking hours.”
McIver says the ROI for the project is really around the piece of mind that the uni’s data is safe, but the other part is “purely a time thing”.
The existing system is expandable to 100TB and once that fills up IT will look at procuring another device. Its storage requirements are growing at an estimated 10TB per year.
Senior systems administrator Peter Gray says the 10 existing systems will also be consolidated into six and form a clustered configuration.