In terms of back ups of the media:
    1. How often are backups taken?
    2. Where are the backups stored?
We do not take backups of the S3 data. 
We rely on AWS to keep our S3 data safe. AWS replicates the content of an S3 bucket in 3 or more availability zones.

The greater danger is that we at Poodll might damage/destroy data. As a safeguard against this we use S3 bucket object versioning, with a lifecycle policy so that non current versions of S3 objects are deleted after 30 days.

A loss of service would occur should our Cloud Poodll server go down. This would not prevent access to recorded media, these are accessed directly on S3. This would prevent recording sessions from taking place and access to some Poodll provided AWS services.

The Cloud Poodll server handles requests for recorders and for some web services such as text to speech synthesis. There is no user authentication at this level, the user's access to Cloud Poodll services is via a site token generated in a Moodle server - Cloud Poodll server session. The Cloud Poodll server generally speaking just serves pre-signed urls to users and audio/video recorders in iframes.

The Cloud Poodll server is backed up daily and exists in a server cluster located in AWS US East 1 (North Virginia). No user data is stored on the Cloud Poodll server. 

c. Should there be a service interruption or loss, what is RTO?

We aim to return services to a fully operational state in the event of an outage, within 6 hours of becoming aware of a problem.