How is my data protected from loss?
Hardware fails, gets stolen or lost. Users accidentally delete files, introduce malware, overwrite important documents. Systems get hacked.
We believe that our storage services are much safer than using internal storage, external USB drives, or NAS devices on your desk!
1. We store your data in a secure location
Systems are stored in UPS protected, environmentally controlled data centres on the University Campus. Our data centres are physically secure and have advanced fire and intrusion detection and prevention systems. Access to these areas is restricted by card access to both the data centre and individual rack. External staff are always accompanied by IT Staff.
2. Servers do not have direct access to the internet.
Systems are accessible directly on-campus, however off campus, files must be access via our MyFiles gateway. This provides WebDAV or HTML access to your Windows SMB filestores.
Administrators use HTTPS and SSH interfaces to administer the system.
3. Protection from user error & malware
The storage controller takes regular snapshots of your data, so that in the event a file becomes overwritten, virus-infected or deleted, we can recover from the last snapshot (or further back if required).
4. Protection from disk failure, corruption.
We use Open ZFS software to store your data. Open ZFS actively detects and repairs errors and ensures data is secure even when disks completely fail.
Depending on system size, disks are grouped into 10-15 devices. For each of these, multiple disks have to fail simultaneously in a single system before the primary array fails. If this were to happen, your data is still safe on the secondary 'mirror' system.
5. We design to reduce the risk and impact of hardware failure
Whilst we include dual network connections and power supplies, other hardware failures can occur. These larger failures can take days to resolve. In this situation, we can promote the secondary mirror, so that you can continue to work whilst we resolve the problem.
Around 5% of disks are predicted to fail in a 12 month period and the likelihood of a disk failing increases as the system becomes older.
The overall risk of hardware failure increases as the system ages in a 'bathtub curve', where failure rates accelerate in years four and five. We therefore peform a hardware refresh after 5 years.
6. We replicate your data to a secondary mirror system
Every evening, we replicate (copy) your data to a second system located in a different building. If all the protections we put in place on the primary system fail, and your data is inaccessable, we can promote the secondary. Secondary systems are built to the same specification as the primary.