Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    isabelsavannah
    @isabelsavannah
    Yeah. In the ethereum use case, we're not really attempting to provide security
    Rich Nistuk
    @rnistuk
    and send us that response, with all the correct key-age
    isabelsavannah
    @isabelsavannah
    but because the response includes a transaction ID, it won't be able to pass as the response for a different request
    They can't change the transaction ID without invalidating all the signatures
    Rich Nistuk
    @rnistuk
    Oh, linking response to request
    isabelsavannah
    @isabelsavannah
    yep!
    Rich Nistuk
    @rnistuk
    ok, so every node in the swarm is an oracle, that means that every node in the swarm can write to ethereum?
    isabelsavannah
    @isabelsavannah
    in this hypothetical, yeah
    In practice it doesn't need to be all of them
    Rich Nistuk
    @rnistuk
    But only one would win the "right" to do the write
    isabelsavannah
    @isabelsavannah
    No, they can all write
    Rich Nistuk
    @rnistuk
    the same data?
    wouldn't that mean that I would over pay for that write?
    isabelsavannah
    @isabelsavannah
    Oh yes, this would be way more expensive ;)
    Rich Nistuk
    @rnistuk
    And may end up with duplicates of my data
    in the ethereum block chain...
    isabelsavannah
    @isabelsavannah
    I mean that was the whole idea
    You ask several people instead of asking one
    Rich Nistuk
    @rnistuk
    That way you are a bit more convinced that your data will get there...
    a bit more certain, I mean
    isabelsavannah
    @isabelsavannah
    Since we're almost out of time, I want to go through what I think is the strongest option right now
    Rich Nistuk
    @rnistuk
    ok
    isabelsavannah
    @isabelsavannah
    I don't think we want ot be in the position of writing to ethereum ourselves, becasue that binds us tigheter to ethereum- we'd rather work with any blockchain
    Instead, here's what we do
    The ethereum contract, call it Alice, that wants to talk to bluzelle, creates a second contract, called Bob
    Bob-contract contains, visible for anyone to see, the DB operation Alice-contract wants to perform.
    Rich Nistuk
    @rnistuk
    a meta contract
    isabelsavannah
    @isabelsavannah
    Bob-contract's logic is a single public function, which is "You supply some data. If that data validates as a response to the request, signed by the bluzelle swarm, I'll give you all the money I have"
    And then anyone can carry that out
    Even if I try to interfere with our bet by not delivering the message, then you could just deliver the message yourself
    And the money the contract pays out is your fee for doing so
    John
    @jyclam
    Each request is a new contract?
    isabelsavannah
    @isabelsavannah
    Yeah
    Rich Nistuk
    @rnistuk
    The first correct response wins the money?
    isabelsavannah
    @isabelsavannah
    yep
    Doesn't actually have to be a new contract per request, that's just the simplest way to think abou tit
    Rich Nistuk
    @rnistuk
    and we determine validity by us signing the request, and the responder signing the response...
    Truthiness is determined by the consensus?
    isabelsavannah
    @isabelsavannah
    And the tecnique can actually be chained so that the paying the message-deliverer and paying bluzelle for the db operation together atomically, though w edon't really have time to go into how right now
    yep!
    Rich Nistuk
    @rnistuk
    Cool
    Thanks Isabel!
    isabelsavannah
    @isabelsavannah
    This solution is assuming that the swarm can somehow collectively sign the response
    Alright, thanks for coming out!
    Rich Nistuk
    @rnistuk
    Good night everyone!
    John
    @jyclam
    That's pretty cool
    Thanks Isabel!
    isabelsavannah
    @isabelsavannah
    Next session will be on the subject of using karma/reputation accounting to enforce good behavior in consensus algorithms that don't otherwise tolerate bad actors (raft)
    amastracci
    @amastracci
    Great discussion! Good night all.
    Colin Charles
    @byte
    Today this came on my radar (one needs to register for free to read it) — https://www.infoworld.com/article/3282413/database/sorry-but-blockchain-databases-are-just-not-that-secure.html