i've been keeping notes on the analysis i've been performing of cosmo's
disk and memory usage. some of you might find them interesting:
<http://wiki.osafoundation.org/bin/view/Journal/BrianMoseley20051111>
<http://wiki.osafoundation.org/bin/view/Journal/BrianMoseley20051112>
basically, i'm finding that storing nodes even with relatively few
properties causes an absurd amount of disk usage:
not counting blobs, just the workspace's data dir, 2135 nodes are
persisted using 12,265 files and requiring 49,116kb of disk space.
this is using ObjectPersistenceManager. i haven't been able to test with
any other PMs yet.
anyway, the notes at the above urls may not make complete sense - they
assume some amount of familiarity with the cosmo project
(<http://wiki.osafoundation.org/bin/view/Projects/CosmoHome>) and/or
webdav and caldav in general - but i think the data and conclusions are
pretty understandable.
i'd love to hear comments on what i've done. i'm not at all confident in
my methodology and would be happy to hear critiques, as well as
suggestions on how to improve my situation :)
thanks!
