5 ms·
I've been considering Swarm distributed file system because of its closeness with the Ethereum development. It seems to do the same thing and works already but
by caprese 8y ago
I've been considering Swarm distributed file system because of its closeness with the Ethereum development.
It seems to do the same thing and works already but hardly gets any press. IPFS and the Protocol Lab's Filecoin sale seemed to generate a lot of marketing despite it becoming clearer later that Filecoin is for an unrelated incentivized network.
It is hard understand the pros and cons of choosing to use IPFS over Swarm, or where they are in comparative development cycle.
I know many decentralized applications that opt for IPFS for their storage component, and know of the libraries to help different software stacks with that. But I can't tell if it is right for me, versus the state of Swarm.
- nonsens3 8y agoSwarm and IPFS together with Filecoin try to address the same problem - persistent data storage in a decentralised network. Swarm is not at all "working already" - the incentivisation layer for nodes to store data for other users is not implemented and currently mostly theoretical and work-in-progress. IPFS is more mature in comparison to Swarm, but the underlying architecture is rather different.
- caprese 8y agoWhat is Swarm's intended incentivisation layer and where can I read about their plans? It seems like all documentation including plans are outdated, and I was ignored in their gitter chatroom where devs wanted to talk about dev things and outreach people seemed nonexistent. I see things being stored on Swarm without incentives, like plain text
- nonsens3 8y agoSwarm documentation might not be perfect, but it is not outdated - https://swarm-guide.readthedocs.io https://swarm-guide.readthedocs.io I believe the chapters about PSS, Swarm Feeds, ENS, Architecture, among others, are mostly up-to-date. You can read about the incentivisation layer at https://swarm-gateways.net/bzz:/theswarm.eth/ethersphere/orange-papers/1/sw%5E3.pdf https://swarm-gateways.net/bzz:/theswarm.eth/ethersphere/ora... Currently incentivisation is not integrated or implemented in Swarm, so a user has no guarantees about what happens with their uploaded content. If the node hosting it disconnects from the network, it will be gone. The plans to address this are through the sw^3 protocols suite and/or erasure coding. Regarding plain text - it doesn't really matter what bytes you store in Swarm - encryption is implemented and you can store non-encrypted or encrypted bytes, this has nothing to do with incentives for persistent storage. We try to do outreach and answer community questions when possible, but the team is not big and this is currently done on a best-effort base, we could definitely improve on that front, I agree.
- theamk 8y agoDoes Swarm’s closeness with the Ethereum means it will be unsuitable for any tasks which do not rely on financial incentives? For example, if in the future debian is moved to IPFS, then many organizations are likely run local IPFS servers with Debian repos pinned. But if debian is moved to Swarm, I do not think that many organizations will be incentivized - the money are insignificant in the total spending, while engineering effort and organizational overhead (finance) is likely to be very big.
- caprese 8y agoWhy do you think that it would be different? "If Debian is moved to Swarm, then many organizations are likely to run local Swarm servers with Debian repos pinned"
- theamk 8y agoBecause from what I understand, Swarm requires Ethereum and a constant money flow. And the monthly amount of money is hard to predict. In many places I worked at, anything money related has huge overhead from finance department, and something more complex than “$X/mo” will require an immense amount of coordination and permissions.