Expand my Community achievements bar.

LCCS record/playback questions

Avatar

Former Community Member

Hello, everbody!

I have some doubts regarding LCCS session record/playback:

1) LCCS server starts the playback only when the zip file is completely dowloaded from WebDAV? It will takes forever for large sessions, wont it?

2) If yes, what can I do to improve this upload speed? Is there an, or a set of, datacenter storage service with privileged access to LCCS infrastructure that we can hire to decrease this transfer time?

3) I noted that the LCCS does not download the zip file again per each concurrent playback, it appears to have a kind of internal cache. How does it work? Is there any way to improve it's eficiency?

4) Can I choose which shared model to be recorded or not? It's because there are some PODs that does not makes sense for us in ondemand sessions.

Thanks, guys!

Renato A. Ferreira

6 Replies

Avatar

Former Community Member

Hi Renato,

1) Typically, a recorded session doesn't take too much bandwidth. For gigantic ones, yes, this will take more time to upload than others.

2) Not at this time. So far, testing we've done on moderately long recordings has shown good results from Dreamhost, among others.

3) The cache basically stays on the service while someone is watching a recording (up to 5 minutes after they leave). Note that users watching a recording do incur the $0.01/user-hour usage.

4) You can't choose what's being recorded, but you can choose what's being played back. For example, your playback application could have a totally different layout and choose not to include various pods. You can even go as far as to delete the streams you don't need from the recording .zip.

  hope that helps,

   nigel

Avatar

Former Community Member

Thanks, Nigel.

Is there any known way to force a specifc file name to be used in playback process? It's because I saw the room name within the stored file name, and it made me think that I will need to keep or recreate a room with same name for playback. I agree that multiples rooms are the best way to handle concurrent sessions, but, if it's also needed for playbacks, the number of rooms in my account may grow fastly and make some server-to-server operatios gradually slower.

Or maybe I could also automatically rename the file for something like *********_x002F_PLAYBACKROOM_x002F_sessionID.zip and leave only this room created, as the session ID already is a configurable parameter. Do you see any problem in this idea?

Thanks,

Renato A. Ferreira

Avatar

Former Community Member

Hi Renato,

You shouldn't need to keep the room referenced in the recording, once we

release the 1st production build of recording (hopefully coming very soon!).

nigel

Avatar

Former Community Member

Hmmm... Could we expect something yet for 2011?

Renato A. Ferreira