The End of Data Silos: Interoperability via Cryptocurrencies

We're stuck in a world of data silos: my photos are spread across dozens of apps and most services try to capture my data instead of working together. Bitcoin and the blockchain's novel form of consensus might inspire us to think about new ways of running databases in a more decentralized manner. With these systems, we're pushed to think about issues like interoperability, transparency, and open access when we start to design our applications, making it easier to work on data across trust domains.

From codemesh.io 2016 in London

Transcript

1. The End of Data Silos:
   Interoperability via
   Cryptocurrencies
   Neha Narula
   Digital Currency Initiative
   

   View Slide

2. +
   

   View Slide

3. interoperability
   

   View Slide

4. View Slide

5. View Slide

6. View Slide

7. DB
   (EMRs)
   Application
   Application
   

   View Slide

8. DB
   (EMRs)
   Application
   Doctor
   DB
   (EMRs)
   Application
   Hospital
   APIs
   

   View Slide

9. Problems:
   Back office reconciliation
   Hard to share – need to have n2 APIs
   Compromises still happen!
   

   View Slide

10. DB
    (EMRs)
    Application
    Doctor
    DB
    (EMRs)
    Application
    Hospital
    APIs
    

    View Slide

11. DB
    Hospital
    Pharmacist
    Doctor
    

    View Slide

12. A new (again)
    database:
    The blockchain
    

    View Slide

13. distributed consensus
    public key cryptography
    common data formats and protocols
    

    View Slide

14. Distributed
    consensus
    Problem of multiple computers
    agreeing on a value
    Used to build a log of operations
    with state machine replication
    Works as a distributed database
    

    View Slide

15. Distributed consensus
    log
    

    View Slide

16. Byzantine fault tolerant distributed
    consensus
    log
    

    View Slide

17. Consensus: a very old problem
    1982
    Byzantine Generals
    Problem formulated
    1999
    Practical Byzantine Fault
    Tolerance paper published
    2009
    Bitcoin introduces
    Nakamoto Consensus
    1990
    Leslie Lamport
    writes about Paxos
    2014
    1980 1985 1990 1995 2000 2005 2010 2015
    Raft paper
    published
    2006
    Google publishes
    Chubby paper
    

    View Slide

18. Public key
    cryptography
    Hash functions
    Digital signatures
    Sharing data selectively
    

    View Slide

19. Public key
    cryptography
    Hash functions
    Digital signatures
    Sharing data selectively
    

    View Slide

20. Common data
    formats
    Rules for what can go in the log
    records
    Agreement on valid operations and
    states
    

    View Slide

21. Components of blockchains
    distributed consensus
    public key cryptography
    common data formats
    +
    +
    

    View Slide

22. using a log to build a
    decentralized digital currency
    Alice Bob
    1MHepPtrqAxZ
    

    View Slide

23. Potential problems with digital token transfer
    1.  Intercept transfer and steal funds
    2.  Spend money without authorization
    3.  Replay attack
    4.  Double spend
    

    View Slide

24. “I, Alice, would like to send
    Bob one coin”
    Alice!
    “I, Alice, would like to send
    Bob 1MHepPtrqAxZ”
    Alice!
    Alice Bob
    Eve
    1.  Intercept transaction and steal funds
    2.  Spend money without authorization
    3.  Replay attack
    4.  Double spend
    Eve
    Alice!
    

    View Slide

25. “I, Alice, would like to send
    Bob 1MHepPtrqAxZ”
    Alice!
    “I, Alice, would like to send
    Carol 1MHepPtrqAxZ”
    Alice!
    Bob
    Carol
    1.  Intercept transaction and steal funds
    2.  Spend money without authorization
    3.  Replay attack
    4.  Double spend
    Alice!
    Uniqueness and ordering
    Alice
    

    View Slide

26. Don’t trust, verify.
    

    View Slide

27. Blocks
    merkle root
    hash of prev
    block
    timestamp
    # txns
    txns
    

    View Slide

28. Merkle tree
    H(HA
    HB
    )
    H(B) H(C) H(D)
    H(A)
    H(HC
    HD
    )
    H(HAB
    HCD
    )
    A B C D
    merkle root
    

    View Slide

29. Merkle tree
    H(HA
    HB
    )
    H(B) H(C) H(D)
    H(A)
    H(HC
    HD
    )
    H(HAB
    HCD
    )
    A B C D
    C in MY
    merkle root?
    merkle root
    

    View Slide

30. log
    Participation
    based on
    identity
    Sybil attack
    

    View Slide

31. Proof of work: Bitcoin mining
    Nakamoto Consensus
    

    View Slide

32. View Slide

33. Blocks created by solving a computational puzzle
    merkle root
    hash of prev
    block
    timestamp
    nonce
    difficulty
    # txns
    txns
    

    View Slide

34. View Slide

35. View Slide

36. Why do people solve this puzzle and
    follow the rules?
    

    View Slide

37. $$$
    

    View Slide

38. distributed consensus
    public key cryptography
    common data formats and protocol
    

    View Slide

39. distributed consensus
    public key cryptography
    common data formats and protocol
    mechanism design
    

    View Slide

40. Rational behavior in Bitcoin
    Try to solve the puzzle for bitcoins
    Build on the longest valid chain
    The one the other
    rational ones are
    building on
    All valid
    transactions
    Why are bitcoins
    worth anything??
    

    View Slide

41. log
    

    View Slide

42. log
    ?
    ?
    ?
    ?
    

    View Slide

43. Downsides of Bitcoin
    Energy usage for proof of work
    Performance
    Concentration of mining power
    Difficulty of use
    Uncertainty of new currency
    

    View Slide

44. What we can learn from Bitcoin
    Open-admissions, rationality-based
    protocols work at large scale!
    Simple transaction formats are useful
    Giving users the power to audit their
    own transactions is powerful
    More open systems
    with less trust →
    -  Choice
    -  Interoperability
    

    View Slide

45. Savings and credit practices in the
    developing world
    Mobile money
    Real time gross financial settlement
    Credentialing
    Property titling and asset transfer
    Global rights management for
    music
    Medical records
    Decentralized publishing
    Many real
    world
    applications
    

    View Slide

46. View Slide

47. A vicious cycle:
    •  High fees
    •  Unproven creditworthiness
    •  Fraud, mistrust in documentation
    •  Unenforceable contracts
    •  Communities without access to basic
    infrastructure
    Our current financial system is not serving everyone
    

    View Slide

48. Blockchain
    The power of peer-to-peer and open systems
    Payments
    Smart
    contracts
    Privacy
    

    View Slide

49. Thank you!
    MIT Digital Currency Initiative
    dci.mit.edu
    ted.com/talks/neha_narula_the_future_of_money
    narula@media.mit.edu
    @neha
    

    View Slide