5 ms·

Why do we need smart contracts and a blockchain in there, do distributed databases not offer this at either better security or performance?

by a3w 4y ago

deweller 4y ago
There needs to be an incentive for users to host other people's data. Centralized servers and advertising is the current model. Blockchains with tokenized value is another model.
- a3w 4y ago
  Two point of views: Server offering person Alice: If I want to use the service, I want to use the service, not get cryptocoin shares. And I usually want stuff on the internet for free. Server or service using person Bob: If I want to make money by hosting, I need someone to pay for my hosting-mined coins. I don't understand as to who in this overview is willing to pay for a place on the blockchain with fiat currency? Someone needs to pay Alice for the efforts she has put into hosting, but who transfers something worth real money to here? Even the author stated that he wants to use a blockchain for user identification, not monetary incentive. But perhaps he means identification is what is granted to you, no money involved. But that can't be it either: So you can host a server, to gain the right to use the server? No, that is the status quo. Lets take a look what happens if you do not run your own server: So you host a more or less public server for others as the scenario proposed here. A client connects, and sends that that shall be stored in a blockchain (i.e. append only, publicly distributed database) to upload [1] and later confirm their identity. When not running the server, who says that the server is not malicious and uses a different identity for the federated network than it displays to the client that connected to it? [1] Something like the the Web of Trust exists between existing servers, or the server will have to spend a resource like having coins or a history of already signed up users to gain the right to add user identities to the blockchain? The Proof of Stake/Work paragraph above probably now again applies for true "trustless" systems, that require full client (i.e. technically server node) equals user identity.
- blamestross 4y ago
  The whole bad assumption of Bitcoin is that you need to bribe people with a scam to host things. BitTorrent's DHT never had this problem. BitTorrent existing and working IS the value. "We have a working social network" is likely sufficient value to motivate users to host things.
- cowtools 4y ago
  I don't think you need to incentivise people to host each other's data. People will host their own data, and they may host others' data for free like in BitTorrent. I am skeptical of the need for filecoin-like systems in the first place. If all you have is a hammer, then every problem looks like a nail.
  - k__ 4y ago
    The "hammer" were centralized systems, and now we have an alternative with Filecoin, Arweave, and the likes.
    - cowtools 4y ago
      The hammer here is both. Cryptocurrrency is solving an incentive problem here that doesn't really exist. If anything, File hosters should not be the ones that pay for the hosting, it should be whoever is retriving the file, and those funds should go towards relay nodes and hosting nodes. filecoin's incentive structure essentially replicates that of the existing internet, where file hosters will have to find some other form of income like advertisements or tracking to cover their costs. Here is a comment I made yesterday on the matter: >Instead of "liking" or "upvoting" a post on a centralized forum, why not "rehost" or "forward" a post on a decentralized forum: essentially seeding it like in BitTorrent or "pinning" it in IPFS. "Followers" of a user donate their storage and bandwidth to them, combating bureaucratic attacks like delisting and DDoS against popular users.
      - k__ 4y ago
        I think, this is the main problem. Nobody really wants to host other peoples stuff, especially not on their desktop or smartphone. That's where Filecoin comes in. But yes, I think the Filecoin model is to heavy weight right now.
        cowtools 4y ago
        people have extra storage and bandwidth that they don't need. Anything that the public decides not to rehost is not worth being published.
- groffee 4y ago
  > There needs to be an incentive for users to host other people's data. And then you inadvertently host 'terrorist' materials and get 25 years in jail. People are trying to redesign 'social networks' with no fundamental understanding of what they are, how to run them or the very real risks involved.
- nanomonkey 4y ago
  The incentive for me is that the information becomes off-line first. I've been on Scuttlebutt for 5 years now, and all of my discussions are available to me to go back to and expand upon whenever I want, no matter if I'm connected to the internet or not. Necro-posting (bringing up old discussions) becomes a viable option, as does scouring through old information. What is more interesting to me, is utilizing this functionality for something beyond a social network, something more productive like an ERP system, yet allowing the social aspect of collaboration improve such a platform.
hrdwdmrbl 4y ago
Yes. A blockchain is one form of a distributed database. A blockchain allows writes from multiple parties without permission. AFAIK, that's the key
- cowtools 4y ago
  Blockchains are only useful for low-bandwidth tasks that require complete consensus.
- halfmatthalfcat 4y ago
  "The key" to whom? Most people wouldn't know nor care about a blockchain backed social network - most just want it to work and be without all the current baggage. This being "the key" is only to those forcing blockchain into things where it doesn't belong to justify it's existence.
  - k__ 4y ago
    Well, people who get their content removed probably care about the network being decentralized.
  - douglaswlance 4y ago
    It's not hard to understand the difference between something accruing wealth to the owner or not, being able to be taken down or not, or being able run your own node or not. People might not care about the technical implementation details. But they do care about what the implication of those details would be.
douglaswlance 4y ago
Which distributed database systems would serve this function?
varunsrin 4y ago
author here. what distributed database would let you own your identity without any third party party being able to take it away from you? that's a necessary condition for sufficient decentralization, and the only distributed databases capable of achieving such guarantees are L1 blockchains.
- a3w 4y ago
  > What distributed database would let you own your identity Owning a private key file when signing transactions allows to prove your identity to other parties. That is, what a blockchain uses to map identities to accounts/wallets, too. Having a consensus protocol does not provide identity proving, but the eponymous censensus is based on a share of either state or computational power. And state is just a database. State that decides what goes into the database is a recursion, voila an example would be Proof of Stake. Or as censensus protocols go, other names exist for the concept of proof of previously-done-arbitrary-things, which allowed you to mention a share with you cryptographic signature (i.e. hash derived from your private key) in the append-only database. Cases like a "plant a tree" eco-coins. Also, to see the bigger picture on > What distributed database would let you own your identity: from a technically more complex angle: Tim Berners-Lee proposes a DRM system for users to own their content, instead of using this technology against user liberties. That would include owning your identity and enforcing your will on you intellectual property. I found these plans not feasible, yet, but at least people who know about cryptography still research the topics of user rights and P2P networking.
pcthrowaway 4y ago
> Some believe that decentralization requires the entire social network to be on a blockchain. This is unnecessary and even undesirable. Social networks generate petabytes of data every year, which can be very expensive to store on-chain. Blockchains also make it difficult to delete data forever, which is a desirable feature for users. A network design that leverages on-chain systems to decentralize ownership while using off-chain systems for a better user experience is a better path to building social networks. The article does indicate a blockchain is undesirable For another decentralized social network (that runs entirely without a blockchain) check out Scuttlebutt (not affiliated)

Why do we need smart contracts and a blockchain in there, do distributed databases not offer this at either better security or performance?

by a3w 4y ago

Why do we need smart contracts and a blockchain in there, do distributed databases not offer this at either better security or performance?

deweller 4y ago

There needs to be an incentive for users to host other people's data. Centralized servers and advertising is the current model. Blockchains with tokenized value is another model.

a3w 4y ago
Two point of views: Server offering person Alice: If I want to use the service, I want to use the service, not get cryptocoin shares. And I usually want stuff on the internet for free. Server or service using person Bob: If I want to make money by hosting, I need someone to pay for my hosting-mined coins. I don't understand as to who in this overview is willing to pay for a place on the blockchain with fiat currency? Someone needs to pay Alice for the efforts she has put into hosting, but who transfers something worth real money to here? Even the author stated that he wants to use a blockchain for user identification, not monetary incentive. But perhaps he means identification is what is granted to you, no money involved. But that can't be it either: So you can host a server, to gain the right to use the server? No, that is the status quo. Lets take a look what happens if you do not run your own server: So you host a more or less public server for others as the scenario proposed here. A client connects, and sends that that shall be stored in a blockchain (i.e. append only, publicly distributed database) to upload [1] and later confirm their identity. When not running the server, who says that the server is not malicious and uses a different identity for the federated network than it displays to the client that connected to it? [1] Something like the the Web of Trust exists between existing servers, or the server will have to spend a resource like having coins or a history of already signed up users to gain the right to add user identities to the blockchain? The Proof of Stake/Work paragraph above probably now again applies for true "trustless" systems, that require full client (i.e. technically server node) equals user identity.
blamestross 4y ago
The whole bad assumption of Bitcoin is that you need to bribe people with a scam to host things. BitTorrent's DHT never had this problem. BitTorrent existing and working IS the value. "We have a working social network" is likely sufficient value to motivate users to host things.
cowtools 4y ago
I don't think you need to incentivise people to host each other's data. People will host their own data, and they may host others' data for free like in BitTorrent. I am skeptical of the need for filecoin-like systems in the first place. If all you have is a hammer, then every problem looks like a nail.
- k__ 4y ago
  The "hammer" were centralized systems, and now we have an alternative with Filecoin, Arweave, and the likes.
  - cowtools 4y ago
    The hammer here is both. Cryptocurrrency is solving an incentive problem here that doesn't really exist. If anything, File hosters should not be the ones that pay for the hosting, it should be whoever is retriving the file, and those funds should go towards relay nodes and hosting nodes. filecoin's incentive structure essentially replicates that of the existing internet, where file hosters will have to find some other form of income like advertisements or tracking to cover their costs. Here is a comment I made yesterday on the matter: >Instead of "liking" or "upvoting" a post on a centralized forum, why not "rehost" or "forward" a post on a decentralized forum: essentially seeding it like in BitTorrent or "pinning" it in IPFS. "Followers" of a user donate their storage and bandwidth to them, combating bureaucratic attacks like delisting and DDoS against popular users.
    - k__ 4y ago
      I think, this is the main problem. Nobody really wants to host other peoples stuff, especially not on their desktop or smartphone. That's where Filecoin comes in. But yes, I think the Filecoin model is to heavy weight right now.
      - cowtools 4y ago
        people have extra storage and bandwidth that they don't need. Anything that the public decides not to rehost is not worth being published.
groffee 4y ago
> There needs to be an incentive for users to host other people's data. And then you inadvertently host 'terrorist' materials and get 25 years in jail. People are trying to redesign 'social networks' with no fundamental understanding of what they are, how to run them or the very real risks involved.
nanomonkey 4y ago
The incentive for me is that the information becomes off-line first. I've been on Scuttlebutt for 5 years now, and all of my discussions are available to me to go back to and expand upon whenever I want, no matter if I'm connected to the internet or not. Necro-posting (bringing up old discussions) becomes a viable option, as does scouring through old information. What is more interesting to me, is utilizing this functionality for something beyond a social network, something more productive like an ERP system, yet allowing the social aspect of collaboration improve such a platform.

hrdwdmrbl 4y ago

Yes. A blockchain is one form of a distributed database. A blockchain allows writes from multiple parties without permission. AFAIK, that's the key

cowtools 4y ago
Blockchains are only useful for low-bandwidth tasks that require complete consensus.
halfmatthalfcat 4y ago
"The key" to whom? Most people wouldn't know nor care about a blockchain backed social network - most just want it to work and be without all the current baggage. This being "the key" is only to those forcing blockchain into things where it doesn't belong to justify it's existence.
- k__ 4y ago
  Well, people who get their content removed probably care about the network being decentralized.
- douglaswlance 4y ago
  It's not hard to understand the difference between something accruing wealth to the owner or not, being able to be taken down or not, or being able run your own node or not. People might not care about the technical implementation details. But they do care about what the implication of those details would be.

douglaswlance 4y ago

Which distributed database systems would serve this function?

varunsrin 4y ago

author here. what distributed database would let you own your identity without any third party party being able to take it away from you? that's a necessary condition for sufficient decentralization, and the only distributed databases capable of achieving such guarantees are L1 blockchains.

a3w 4y ago
> What distributed database would let you own your identity Owning a private key file when signing transactions allows to prove your identity to other parties. That is, what a blockchain uses to map identities to accounts/wallets, too. Having a consensus protocol does not provide identity proving, but the eponymous censensus is based on a share of either state or computational power. And state is just a database. State that decides what goes into the database is a recursion, voila an example would be Proof of Stake. Or as censensus protocols go, other names exist for the concept of proof of previously-done-arbitrary-things, which allowed you to mention a share with you cryptographic signature (i.e. hash derived from your private key) in the append-only database. Cases like a "plant a tree" eco-coins. Also, to see the bigger picture on > What distributed database would let you own your identity: from a technically more complex angle: Tim Berners-Lee proposes a DRM system for users to own their content, instead of using this technology against user liberties. That would include owning your identity and enforcing your will on you intellectual property. I found these plans not feasible, yet, but at least people who know about cryptography still research the topics of user rights and P2P networking.

pcthrowaway 4y ago

> Some believe that decentralization requires the entire social network to be on a blockchain. This is unnecessary and even undesirable. Social networks generate petabytes of data every year, which can be very expensive to store on-chain. Blockchains also make it difficult to delete data forever, which is a desirable feature for users. A network design that leverages on-chain systems to decentralize ownership while using off-chain systems for a better user experience is a better path to building social networks. The article does indicate a blockchain is undesirable For another decentralized social network (that runs entirely without a blockchain) check out Scuttlebutt (not affiliated)