Research Progress Update

15-16/2/2018

Edinburgh University

Blockchain Technology Lab

Orfeas Stefanos Thyfronitis Litos

Part I

Trust Semantics

Motivating problem: user perspective

Trusting a vendor

  • Question: "I found an online store that seems great and cheaper than ebay. Should I buy from there?"
  • Simple answer: "Ask your friends"

Motivating problem: researcher perspective

  • What does "trust" even mean?
  • How to formally argue about trust?
  • Is there a way to reliably measure trust without assuming Trusted Third Parties?

 

  • Today's assumption: ebay is trustworthy
  • Aim: no assumptions, just an amazing reputation system (or why it's impossible)

Solution Approach

  • Model a simple market
  • Model a generic mechanism for answering questions like "which vendor is more trustworthy?"
  • Compare the various mechanisms
  • Show whether a foolproof reputation system without Trusted Third Parties is feasible

Our Model

  • Alice has a desire.
  • Bob can satisfy it for a price
  • But he can take the money and run.
  • So, before paying him, Alice asks the Trust functionality.
  • If it says OK, Alice goes on to Trade with Bob.
  • Question: What is the best (implementable) Trust functionality?

Part II

Payment Channels

Blockchains don't scale

  • Consensus: everybody agrees on what happened
  • Present: every computer learns every transaction
    • Will never scale well

Payment channels to the rescue!

  • As long as interested parties agree, the world doesn't have to know
  • If parties disagree, the world is the arbiter
  • Literature is already big:
    • Lightning (Bitcoin)
    • Perun (Ethereum)
    • Bolt (Zcash)
    • TeeChan (Intel TEE)
    • etc.

Inverting the question

  • What is the minimal global ledger needed for robust payment channels?
  • Does it have to be a blockchain?
  • Is Ripple/Stellar the answer, or can we do better?
  • No answers yet, exploring how trust could be leveraged

What if not a blockchain?

(Vague) idea:

  • The global ledger consists of trust relationships
  • In case of dispute, trust is rearranged
  • On expectation only the cheater loses money

Recap

  • Trust models and semantics
    • Simple market model
    • Ideal Trust functionality
    • Is it realizable?
  • Payment channels (i.e. scaling)
    • Desired properties of payment channels
    • Global ledger designed just for payment channels
    • Could Trust be leveraged?

15-16/2/2018 Progress Update

By orfeas

15-16/2/2018 Progress Update

20' presentation

  • 760