Redundant backup is an add-on feature which is integrated into BountySite storage node, for backup data reliability. Storage nodes are placed near hosting servers and Redundant Backup helps in storing a copy of data in remote location for Geographical data redundancy and disaster recovery.
The idea was simple. Keep storage node near hosting server for fast backup and restore. What if we loose the storage node? Redundant backup feature was then added to ensure a copy of data is kept in a remote Geo location.
Since, we support customized storage node, we let the customer decide on whether Redundant Backup is to be integrated or not. End site owner can always use Backup Sync to upload a copy of data to own public cloud account, irrespective of whether Redundant Backup is configured on storage node.
N copies to Remote Dedicated Server
We have built our own Redundant Backup receiver API for storing a copy of backup data on remote dedicated server. We expect some customers to have higher level of trust on dedicated servers than some cloud platforms. Also, this gives an opportunity to upsell services. To add more reliability, we allow storing multiple copies on different Geo locations.
Let's say our hosting server is based out in Utah region datacenter. Our storage node can be located in Utah region datacenter itself or Los Angeles or San Francisco. We can configure our redundant backup node within US, say New York, Dallas and Atlanta, or can choose global locations New York, London and Frankfurt across globe.
What if the primary data on storage node is corrupted? Will redundant backup overwrite with corrupt data?
Redundant backup API accepts redundant backup copy of data and stores in a temporary file. If this a first copy, the temporary file is renamed as redundant data copy. If there is already an existing copy, which was taken on previous backup, the file sizes are compared. If the new temporary file is smaller than the previous copy file, a new file is created. The previous copy is not touched, and BountySite notifies internal system of the event. In most cases, we expect the new site backup to only be bigger or same as previous copy.
Storage node has several checks in place to catch I/O related issues, besides the fact that redundant backup also checks file integrity before overwriting previous redundant copy.
Storage node invokes redundant backup only when the backup has downloaded some bytes, which implies that there were changes in site, be it files or database.
Redundant backup nodes stores data on industry standard AES 256 bit encryption.
Redundant Backup to S3
BountySite storage node can now keep one redundant copy to any S3 compliant storage, irrespective of n copies of redundant backup on servers. AWS S3 is the most reliable storage service in the industry. The data is encrypted in transit(using SSL API) and at rest using AES 256 bit encryption.
The other benefit of storing redundant backup on S3, is that you only need to fetch data in case storage node crashes irrecoverably. Using AWS S3 directly for backup and restore is expensive, as download(in restores) bandwidth charges are steep.
Site owner can also keep track of redundant backup logs, by checking Backup Logs.
Site owners can also use Backup Sync to keep a copy of data in their own public cloud backup accounts. Backup Sync password protects the archive with FTP password, besides ensuring that data transit is over SSL.
Redundant backup adds reliability to website backup data, despite having storage node in same region as hosting server. By having storage node near hosting server, this helps us to achieve faster backup and restore.
Many good hosting providers have their own backup, in which case BountySite storage node does not need redundant backup and can be disabled on a custom storage. I would like to highlight a point that BountySite stores snapshot since the site was first backed up, but most backup service have retention period beyond which data cannot be fetched.
Does that mean BountySite replaces server backup software?
No! BountySite focuses only on website contents(/public_html) and is concerned about website security only. Server backups are meant for the whole server, which not only includes website contents, but also applications and config files needed for running a hosting server. Popular server backup software like Acronis and JetBackup, helps in recovering a whole server during disaster recovery. BountySite is designed for website security. The reason BountySite backups website is to be able to guarantee restore back to pristine version and make malware removals simple and easy. So, we cannot be compared with server backup software.