Is it possible to apply the principles of holography in distributed data storage?

Question

Data storage

The fundamental difference between holography and all other methods of image registration is the distribution of information about all the captured objects over the entire surface (c) from wiki

Is sharding possible in such a way that

• The data on each storage unit/server was sufficient for autonomous operation and at the same time gave an idea of ​​the system as a whole.
  
• Adding any unit to the available pool gave consistent data and clarified the overall view of the system.
  

That is, on the fingers, so that after downloading part of the data from one storage, I would already be able to watch a movie, only in poor quality. And pumping it up from another could improve the quality to HD.)) What to read, maybe research papers? Or correct me if I blunted something obvious.))

Answer 1

read codec formats - it works for torrents), but, of course, the initial distribution will be at seams - practically scrolling / rendering the entire video
, you can still have Yandex lectures or speeches where they are about their Yandex - disk, but in general there is not about "download quality", how to store recurring

Answer 2

I'll give you a hint:
Take an N-gon and rotate it in X-dimensional space.
Each corner of an N-gon is a unit of information (like a pixel).
All angles is the minimum "resolution" of all data (e.g. frame).
Each angle during rotation is an increase in "resolution" by a minimum value.
Your task is to segment the complete data into an X-dimensional space by inscribing the full minimum of this data into an N-gon.

Answer 3

All major video codecs use the following principles for compression:
1. We try to predict what the current frame will be from the previous (sometimes also from the next) frames.
2. We look at what data can be thrown out so that the picture does not visually deteriorate.
3. We compress with the usual methods.
As a rule, within one scene (2-6 sec) the data is indivisible.
We can potentially jump to different scenes, but within a scene it's either all or nothing. If there is a beginning and end of the scene and the middle is skipped, then only the beginning will be displayed on the screen, then there will be porridge.
Therefore, within the framework of the current codecs (h264, h265, etc.), sharding is not possible. Potentially, yes, you can come up with a codec. Only the compression ratio will be terrible (the bitrate will be prohibitive).
Again, there is another solution: many hosts store several versions of the video - from very small files to full quality. There are players that automatically pick up the desired file ( https://ru.wikipedia.org/wiki/HLS).