2

My understanding, from an answer to previous post is that it's currently cost prohibitive to exchange datasets of significant size over the ETH network.

Is it possible for the ETH network to evolve into a more data-driven (^ size) than BTC PoW model. Where, this model involves a hybridisation of PoW proportional/dynamically adjusted to the amount of SSD hard disk space a user dedicates to the network.

My thinking, is that we would effectively remove the need for future developers to do Full Syncing. So as this type of data driven sharding proliferates, synchronisation/contract deployment would be much faster because the data is safely distributed/replicated amongst a MIN # of global nodes providing pooled SSD storage. Note I propose for Data Storage mechanisms (e.g..IPFS, Swarm, + this theoretical storage sharding model) that a complete file is Never stored on one node. For future rapid "Confirmations" it must follow recovery and sufficient level of distribution. However, equally rapid re-compiling/assembly of the file with private key access.

just a thought, thanks : )

|improve this question

closed as too broad by Richard Horrocks, Badr Bellaj♦, Sebi, Afr Dec 4 '16 at 16:41

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • This is really a proposal rather than a question with a definite answer, so it would perhaps better fit a Reddit discussion. – Richard Horrocks Dec 3 '16 at 9:17
  • Thanks for reading! It was just random thought really : ) Will post this abstract idea on Reddit – SHA256 Dec 4 '16 at 17:37
  • @SHA256 You might be able to rephrase this as a question that's not too broad. I tried editing your other question about Swarm and feel free to improve it. – eth♦ Dec 25 '16 at 10:44
  • The problem with splitting up one related chunk of data into smaller blobs of data over different shards (which you would need to do, you can't have different bits of data on the same shard, there must be consistency, every full node in a shard must process every tx in that node) is that it adds a lot of latency overhead with fetching those blobs. However, sharding does scale storage (as well as computation, bandwidth and I/O) by not requiring full nodes to store the data of all shards, but just one or more shards. – James Ray May 5 '18 at 9:47
  • Storage rent will also help to internalize the cost of storage which is currently externalized, which will inevitably lead to a tragedy of the commons where there is an insufficient supply of storage resources due to no incentive to store data and a larger demand for it. State-minimized executions are yet another innovation, plus layer 2 archival markets. – James Ray May 5 '18 at 9:49

Browse other questions tagged or ask your own question.