It would be interesting to see how hard would it be to store the data in the DHT node compressed somehow to save memory (which is currently the limiting resource for the DHT, by far).
Having a simple look at how dump files are compressed (between 1/3 and 1/4 or the original size), one could image something could be gained by compressing the data.
The thing is, though, we'd be compressing individual records, most of which are very small. Still, it could be worth trying.
It would be interesting to see how hard would it be to store the data in the DHT node compressed somehow to save memory (which is currently the limiting resource for the DHT, by far).
Having a simple look at how dump files are compressed (between 1/3 and 1/4 or the original size), one could image something could be gained by compressing the data.
The thing is, though, we'd be compressing individual records, most of which are very small. Still, it could be worth trying.