![Page 1: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/1.jpg)
Fritz HengleinUniversity of [email protected]
2nd Symposium on Distributed Ledger TechnologyGold Coast, Australia
Blockchain deconstructed
![Page 2: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/2.jpg)
Professor of Programming Languages and SystemsUniversity of Copenhagen
Head of ResearchDeon Digital AG
Fritz Henglein
Related background
• European Blockchain Consortium (ebcc.eu)
• Steering committee chair, Innovation network for
Finance IT (CFIR.dk)
• Principal investigator, Functional technology for high-performance architectures
(FUTHARK)
Academic background and affiliations
Areas of interest• Programming language technology• Theoretical computer science
(algorithms, semantics, logic)• Blockchain technology• Contract management• Financial technology• Enterprise systems
![Page 3: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/3.jpg)
What is a blockchain/DL system?
3
![Page 4: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/4.jpg)
What is a blockchain/DL system?
Any peer-to-peer decentralized system• storing a single, consistent tamper-proof
record of events• admitting only appending of new events• enforcing a fixed or user-defined protocol
(contract) for appending new events• Guaranteeing unique ownership of
resources (no double spending)• with no trusted or privileged parties• based on cryptographic principles
4
![Page 5: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/5.jpg)
What is a blockchain/DL system?
• organizational and technical decentralization;• tamper-proof recording of events and their evidence; and • guaranteed resource preservation and credit limit enforcement
5
A computer system characterized by
![Page 6: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/6.jpg)
Organizational and technical decentralization
• Technical decentralization: A distributed peer-to-peer system
• Organizational decentralization: No single or select group of organizations controls/has privileged rights to system compared to others
• Governance policy for regulating membership, functionality, conflict resolution, etc.• Group of organizations operating and using system can
be open and self-authenticating (nonpermissioned, ”distributed ledger technology”) or closed and externallyauthenticated (permissioned, ”blockchain”).
6
![Page 7: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/7.jpg)
REA accounting model (extended information and contracts)
• Resource (= asset): Money, licenses, physical objects (trucks),...• Information: Data, invoices,...• Agent: Person, company, institution, autonomous device,...• Contract: Specification of obligations, permissions and prohibitions• Event:• Atomic event: • A transfers R to B • A transforms R to R’• A informs B of I• ...
• Complex event: Set of events that satisfies a given (sub)contract
7
![Page 8: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/8.jpg)
Tamper-proof recording of events and their evidence
• Event recording: Events are recorded• Tamper-proof: They cannot subsequently be altered or deleted• Evidence• for atomic events: signature, plus supporting evidence of event having happened
(pictures, receipts, DNA samples, GPS data, )• for complex events: (mathematical) proof that a set of events is a correct execution of
a contract
8
![Page 9: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/9.jpg)
Digital Twin via physical evidence framework
time
Alternative state 1
Alternative state 2e'
Contract
Physical evidence frameworkPhysical world
History in Blockchain
e
now
Physical assetsand
physical events
![Page 10: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/10.jpg)
Guaranteed resource preservation and credit limit enforcement
• Resource preservation: Transfers keep the sum of all resources invariant:• A transfers 50 ETH to B: The sum of all ETH is the same. Atomically, after event A has
50 ETH less; B has 50 ETH more.• We allow negative numbers
• Credit limit enforcement: A transfer is only valid and effected if thecredit limits of each agent are respected. For above transfer:• If A owns 60 ETH and has credit limit 0: Valid.• If A owns 30 ETH and has credit limit 0: Invalid.• If A owns 30 ETH and has credit limit 20: Valid.
• No-double-spend guarantee = all agents have credit limit 0.
10
![Page 11: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/11.jpg)
Use cases for adaptive credit limits
• Full-reserve monetary system: One agent (the central bank) has nocredit limit (or dynamic credit limit according to some governanceregime), all others have credit limit 0.
• Fractional-reserve monetary system: A designated set of agents(”banks”) have a dynamic credit limit, all other have credit limit 0.
• Demand-driven production of physical assets: A car manufacturer has no credit limit (they produce cars on demand), all others have credit limit 0.
11
![Page 12: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/12.jpg)
Commutativity theorem, part 1
• Theorem: If every agent has an infinite credit limit, then all resourcetransfers are valid and can be executed in arbitrary order. (Each orderresults in the same state of ownership.)
• Corollary: Contract execution involving k agents requires only consensusby the k agents on which events have happened.
• Note, usually k=2. The Internet with TLS (with tamper-proof recording of message sending) is a permissioned blockchain/DL system if there are nocredit limits!
12
![Page 13: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/13.jpg)
Commutative theorem, part 2
• If some agents have finite credit limits, outside validation of theirresource transfers is required. • Point-to-point communication between the two agents only is insufficient.• Some information about resource transfers must be ”leaked” to other nodes for
validation.
13
![Page 14: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/14.jpg)
Canonical distributed ledger architecture (functional view)
SHARED SYSTEM:RESOURCE TRANSFER
VALIDATION ONLY
CLIENT SYSTEM:SINGLE PARTY INFORMATION
CLIENT SYSTEM:SINGLE PARTY INFORMATION
CLIENT SYSTEM:SINGLE PARTY INFORMATION
CLIENT SYSTEM:SINGLE PARTY INFORMATION subscribe
postquery
private messages
14
![Page 15: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/15.jpg)
Canonical distributed ledger architecture(distributed systems view)
CLIENT SYSTEM:SINGLE PARTY INFORMATION
CLIENT SYSTEM:SINGLE PARTY INFORMATION
CLIENT SYSTEM:SINGLE PARTY INFORMATION
CLIENT SYSTEM:SINGLE PARTY INFORMATION subscribe
postquery
private messages
S H A R E D
S Y S T E M :
R E S O U R C E A N D
I N F O R M A T I O N
M A N A G E M E N T
S H A R E D
S Y S T E M :
R E S O U R C E A N D
I N F O R M A T I O N
M A N A G E M E N T
S H A R E D
S Y S T E M :
R E S O U R C E A N D
I N F O R M A T I O N
M A N A G E M E N T
S H A R E D
S Y S T E M :
R E S O U R C E A N D
I N F O R M A T I O N
M A N A G E M E N T
consensusand
replication
15
![Page 16: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/16.jpg)
Conclusions and open problems
• No consensus on globally total order of events is necessary• Current blockchain/DL systems solve an unnecessarily hard problem (transaction
order consensus)
• Not even consensus on partial order of events is strictly necessary.• Consensus on resource transfers needs to ensure that the set of all
eventually validated resource transfers respects all credit limits.• Specialized consensus protocols for resource transfers are conceivable
and needed for scalability:• Hierarchical clearing and settlement (hierarchical ”sharding” by partititioning of
agents)• Time- and resource-sensitive validation (bigger transfers require more time)• Insurance (appyling transaction fees to covering losses due to overdrafts detected
too late)• ...
16
![Page 17: Blockchain deconstructed - Griffith University · 2018-07-13 · Blockchain deconstructed. Professor of Programming Languages and Systems University of Copenhagen Head of Research](https://reader034.vdocument.in/reader034/viewer/2022050412/5f896323426cdd7ab214fc06/html5/thumbnails/17.jpg)
Thank you!
17