On Fri, 2016-02-12 at 11:09 +1100, Jim Birch wrote: > It seems to me that it would be possible to improve the satellite > performance significantly via the use of a smart, distributed version of > [a proxy][...] > Squid, something like this: Satellite endpoints run "DSquid" software and > have a large local storage,
What is a "satellite endpoint"? > movies. The satellite continuously broadcasts what goes into the central > DSquid cache using some kind of smart algorithm based frequency of access > across the user base. This kind of assumes that the only data being carried is movies and that it's OK to keep the pipe filled with that data. > might be already pre-downloaded into your local cache when decide you want Where is "your local cache"? And who is (are?) "you"? This all seems to presuppose that there are ground stations that people connect to via fast non-satellite links. Or perhaps that satellite users are networked amongst themselves on fast non-satellite links. If neither of those things is true, I don't see how any amount of caching can avoid sending multiple copies of a given movie at random times over satellite links, once per consumer. I really hope you don't mean caching at the consumer site, and therefore caching lots of movies just in case that particular user wants to watch them! I suppose you could cache the data *in the satellite* to avoid repeating one leg of the trip. Maybe that's what you meant by "satellite endpoint", but it doesn't square with your suggested use of cheap spinning magnetic storage, which would be a disaster in space. The caching idea also sort of presupposes one provider - or are multiple providers supposed to use a shared cache? It's all sounding extremely complex at best, and we haven't even added DRM into the mix. > it. The broadcast stream has zero marginal cost so could of be free to > all the satellites users. I don't understand that at all. Literally - I read the words, but can extract no meaning. Could you explain what "the broadcast stream" is in particular, and why it has no marginal cost? Since you are not usually prone to writing nonsense, I can't help feeling that I have badly missed your point. Regards, K. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Karl Auer (ka...@biplane.com.au) http://www.biplane.com.au/kauer http://twitter.com/kauer389 GPG fingerprint: E00D 64ED 9C6A 8605 21E0 0ED0 EE64 2BEE CBCB C38B Old fingerprint: 3C41 82BE A9E7 99A1 B931 5AE7 7638 0147 2C3C 2AC4 _______________________________________________ Link mailing list Link@mailman.anu.edu.au http://mailman.anu.edu.au/mailman/listinfo/link
