A geographically distributed research group has conducted a set of 250 interviews with pupils. Each interview is represented by a video/audio stream, a transcript of the interview, and a set of stills from the video. The length of the interviews varies between 10 and 45 mins. The researchers now analyze the videos and comment on specific behavior of the interviewees. Each comment is linked to a temporal section of the video. The length of a comment varies between single words up to several paragraphs. Comments are seen as research data objects in their own right (with authors, provenance, and additional metadata). 

…Per the preservation policies the research group has negotiated with its Repository Services group: i) the master, archival copy of the video stream will be stored in Amazon Glacier, ii) a lower-resolution, transcoded version of the stream will be distributed across at least three geographically isolated stores and iii) other resources will be subject to the repository’s default storage policies.

..Per the SLA the research group has negotiated with its Repository Services group: i) the video/audio streams will be distributed across no less than independent nodes, ii) other resources will be subject the repository’s default availability policies.