[mythtv-users] Mooting architecture for a DataDirect replacement

Kevin Hulse jedi at mishnet.org
Thu Jun 21 18:47:25 UTC 2007


On Thu, Jun 21, 2007 at 12:34:15PM -0400, Jay R. Ashworth wrote:
> On Thu, Jun 21, 2007 at 11:23:42AM -0500, jedi at mishnet.org wrote:
> > > On Wed, Jun 20, 2007 at 03:42:47PM -0700, Yeechang Lee wrote:
> > >> Jay R. Ashworth <jra at baylink.com> says:
> > >> > One of the *most* important characteristics of AIRING data for the
> > >> > PVR market is that you want to [continue to] propagate changes to
> > >> > the audience just as *late* as you possibly can -- 5 minutes before
> > >> > airtime is not unreasonable, nor impossible, if you pick the right
> > >> > transmission architecture.
> > >>
> > >> I inserted a key omitted phrase in the above because, of course,
> > >> lineup changes ought to be propagated ASAP *and on an ongoing basis*.
> > 
> > Changes can be presented by themselves and picked up by any interested
> > parties. It could just be a "differential" from the last full dump of
> > the schedule data.
> > 
> > Take the master copy plus the latest changes.
> 
> Well, the *real* question is "how far back/forwards" do you need to
> go... but remember how Usenet servers *work*.  You can pull, from the

	Go as far ahead as the data will allow.

        Go as far back as storage and processing make prudent.

	Applying n+1 differentials to a base copy is just annoying. 
There has to be some sane smaller limit that will be set by human
convenience more than technical considerations.

	In either case, chunking the data in one week increments would probably
make some sense. "full backups" need to be frequent enough that people don't 
want to throw sharp objects at you.

> server, as much cached older sked data as you need, in posting order,
> and process it.  The subject lines can be machineable, so you can
> figure out which data you don't already have, too.
> 
> That way there *isn't* really a "full dump" or a "master" copy.  All
> there is is "new items"... which could expire 12 hours after their
> airdate, as well.
> 
> Leveraging RFC1036 and NNTP to move blocks of airing data seems to have
> quite a number of advantages, operationally.

	Polices regarding "initial" copies versus "deltas" should be independent
of the transport mechanism.

> 
> Cheers,
> -- jra
> -- 
> Jay R. Ashworth                   Baylink                      jra at baylink.com
> Designer                     The Things I Think                       RFC 2100
> Ashworth & Associates     http://baylink.pitas.com                     '87 e24
> St Petersburg FL USA      http://photo.imageinc.us             +1 727 647 1274
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users


More information about the mythtv-users mailing list