3 ms·

This line of thought needs to noticed be much more. While I was initially also very excited about distributed/decentralized processing/filesharing systems, this

This line of thought needs to noticed be much more. While I was initially also very excited about distributed/decentralized processing/filesharing systems, this exact type of economic calculation seems to make the existence of many small, personal nodes in such networks unlikely. Even ignoring the electricity and network cost, if the setup and maintenance of such systems takes more than a few minutes per month, it just won't be worth it for the majority of potential providers. They would have to be zero-click installs, daemons in the background, unnoticeable, without requiring any attention from the user, despite issues like moderation. Makes me wonder how the distribution on other filesharing/torrent networks looks like. Another line of questioning I have is about the possible effects of this system should it become popular. Cryptocurrency mining caused huge spikes in graphics card prices - will the same apply to high-capacity harddisks, due to people trying to run mini-datacenters for profit? How much storage will the Filecoin network be able to provide? How would a hypothetical large storage supply interact with pricing models of other providers? And on a more optimistic note, will it enable individuals to develop a new class of programs using lots of storage which aren't possible currently, technically or economically?

gruez 6y ago
>Makes me wonder how the distribution on other filesharing/torrent networks looks like. I think the key difference is in filesharing networks, there are tens/hundreds/thousands copies of data, and each peer can jettison the data at any time without penalty. In contrast, storage networks (I'm using sia as an example here) only have three copies (by default), require upfront commitment (6 months by default), and you stand to lose your collateral in the event you delete the data prematurely. The latter two properties exist to counteract the unreliability associated with only having 3 copies if peers can randomly drop out at will. >will the same apply to high-capacity harddisks, due to people trying to run mini-datacemters for profit? Theoretically no. Presumably because every gigabyte of storage that's being bought for the storage network, is every gigabyte that someone doesn't have to buy to store their data.

gruez 6y ago

>Makes me wonder how the distribution on other filesharing/torrent networks looks like. I think the key difference is in filesharing networks, there are tens/hundreds/thousands copies of data, and each peer can jettison the data at any time without penalty. In contrast, storage networks (I'm using sia as an example here) only have three copies (by default), require upfront commitment (6 months by default), and you stand to lose your collateral in the event you delete the data prematurely. The latter two properties exist to counteract the unreliability associated with only having 3 copies if peers can randomly drop out at will. >will the same apply to high-capacity harddisks, due to people trying to run mini-datacemters for profit? Theoretically no. Presumably because every gigabyte of storage that's being bought for the storage network, is every gigabyte that someone doesn't have to buy to store their data.