Wednesday, June 24, 2015

[hnnnjzmr] Queue to archive

A site wants to distribute content exclusively.  However, someone, a user, is downloading the content and mirroring it in some censorship resistant medium.  The site identifies that user by watermarking the content so each different user is identifiable, then blocks the user from accessing any more content.

Avoid the block by having a very long delay between downloading the content and uploading it to the mirror.  We can do this uniformly using a cache with a FIFO replacement policy: The user downloads content and it is stored in the disk cache, which is rather normal operation.  Ignore the "Valid until" metadata.  (Video is not automatically cached by browsers these days, but that could be easily changed.)  When the cache becomes has become full and an entry is the oldest remaining entry, so about to be overwritten, only then does the it become uploaded into the censorship-resistant mirror.  With a large cache (and reliable hardware, which is a huge assumption), that delay might be years.

A delay of years is acceptable for the purpose of archiving information.

There still remains the problem of the original content provider suing the uploader for copyright violation, or the provider cutting off access years later, which the user might still care about.

No comments :