[AG-TECH] Status of Sessions Captured by Voyager at SCGlobal
Rick Stevens
stevens at mcs.anl.gov
Thu Nov 22 00:03:06 CST 2001
Lisa,
Thanks for the hard work and excellent report.. I can't wait to view the
South Pole
session.
--Rick
At 10:33 PM 11/21/2001 -0600, Lisa Childers wrote:
>Dear Folks,
>
>Just wanted to give you an update on the project to archive SCGlobal using
>Voyager.
>
>The truck arrived at Argonne from Denver on Tuesday morning (11/20) and 6
>icy-cold voyager servers were delivered to our door. I was advised by folks
>here that we should wait until the machines reached room temperature to turn
>them on. By waiting we avoided creating condensation inside the machines
>and thus potential electrical mishaps. At the end-of-day on Tuesday the
>boxes were still too cold to power up.
>
>By this morning they had thawed and we began unpacking and bringing up the
>servers. Today we have had a first look at our data. There is 208G of
>SCGlobal content spread across the 6 servers. Bear in mind that because of
>paranoia there is redundancy in some of the sessions. When possible a
>session was recorded simultaneously w/two servers. I estimate the unique
>data size to be roughly 160G.
>
>We have begun the task of reviewing recordings. Some time today was spent
>performing random playbacks in a rough attempt to categorize problems seen
>in the recordings. Current thinking is to code sessions with a grading
>system, where sessions are tagged with the following keywords:
>
>SCG5 = problem-free session
>SCG4 = problems with the session, but interesting nonetheless
>SCG3 = out-of-synch session, potentially interesting
>SCG2 = difficult problems with the session resulting in an uninteresting
>session
>SCG1 = session is unusable
>
>"Problems" so far have included out-of-synch packets, packet loss and
>malformed or unknown packetized data.
>
>We have begun moving the data off of the loaner servers. This is a big
>task. The audio and video data are already heavily compressed, so we can't
>make it much smaller than it is in its raw form. Also it is not as simple
>as a straight ftp, because in addition to disk files, native voyager data
>include metadata which is stored in a db. We are currently transferring
>data by a custom Voyager session exporter and importer.
>
>One exciting session we ran across today was the presentation from the South
>Pole. 17 nodes attended the presentation, contributing 81 video and audio
>streams to the session. There is loss, but it is still way cool.
>
>I'll post an announcement as soon as these data are available on a public
>server.
>
>Thanks,
>Lisa
More information about the ag-tech
mailing list