Hello potential research collaborator!
Rumour has it that you, a researcher or manager of researchers, are interested in joint research with the Ethereum Foundation. Below are the primary topics the Foundation will be thinking about for the next 2-3 years. If you, like us, enjoy the prospect of thinking about one or more of these topics for the majority of your waking hours, do get in touch. The Foundation does have money to pay the salaries/stipend of those undertaking high-value research.
We have topics in both pure research and applied research. The Foundation as well as the larger Ethereum community seek help on both. Typical outputs from researchers are: peer-reviewed academic papers, technical reports, and/or implementations (prototypes as well as production-ready).
Questions in Fundamental Research
Q1: Can we create a theory of cryptoeconomic mechanisms?
- There are certain patterns that are often used in cryptoeconomic mechanisms. These can be studied in the abstract independently of any specific use case.
- Security deposits (see also proof of stake)
- How do we model capital lockup costs?
- Dual-use of security deposits
- Challenge-response games (one group of actors is given the opportunity to submit evidence that fact X is false, and if no one submits evidence within some period of time, then X is assumed to be true). See also challenge-response authentication.
- Channels
- State channels
- How do we minimize the vulnerability of challenge-response games and channels to liveness or censorship faults of the underlying blockchain?
- Escalation games
- Cross-chain interoperability (see the R3 interoperability paper <todo: add link>)
- Relays
- Hash timelock atomic swaps
Q2: What is the role of cryptoeconomics in distributed systems? What is the role of economics in cryptography?
-
Can we formalize how algorithmic incentives (“cryptoeconomics”) can enhance information security?
-
Modeling behavior of participants in mechanisms
-
Complex game-theoretic interactions
Q3: How do distributed systems influence current economics?
- On net, when and how much does decentralization lower transaction costs?
- No obvious answer. Decentralization _decreases_transaction costs because of: Reduced number of counterparties and reduced need for building trust
- Yet decentralization increases transaction costs because of: increased technical overhead, Decreased usability, increased responsibility.
- Are Transaction costs = transaction fees + coordination costs?
Q4: Within game-theory, can we quantify coordination costs?
- for players running a particular protocol
- for players executing a particular strategy
Q5: What are ways we can manipulate (e.g., guarantee/minimize) coordination costs?
- For example, we can reduce risk by increasing coordination costs.
- Coordination costs are costs from multiple-agents coordinating.
- For example: Discovering potential peers, agreeing on computing coalition strategies, synchronization required for execution, costs of proving to the coalition that players followed coalition strategies, cost of getting rid of individual incentives to deviate
Q6: What protocols have better fault attribution?
-
A protocol fault is uniquely attributable if there is evidence that could be used to umambiguously convince any observer which actor caused the protocol fault. If a fault is non-uniquely-attributable, the blame for the fault can often at least be narrowed down to within N specific actors.
-
Fault attributability in various consensus algorithms
- Chain-based (synchronous) consensus
- Partially synchronous consensus (see minimal slashing conditions)
- Common coins in asynchronous consensus
-
Attributability of censorship or liveness faults.
-
Translating fault attributions into penalties
Q7: What are decentralization’s fundamental limits?
Building on hundreds of impossibility results. E.g., 1 and 2, or even fundamental limits from other areas of computer science.
- What centralized protocols can be decentralized (while preserving guarantees)?
- At what cost in protocol overhead?
- Are there limits to scalability?
- At what cost in incentivization?
- What are the limits to incentivization?
- Limits to attribution
- Limits to mechanism budgets
- With how much security (against coordinated choice, trusted majority required)?
- Limits to fault tolerance
- e.g. in objective protocols and subjective protocols
Objectives in Applied Research
Also knows as Pasteur’s Quadrant.
Right now our primary topics in applied research are: plasma, sharding, and Casper.
1. Base Layer (core protocols)
1.1 Plasma and Sharding [49%]
Goal: Allow Ethereum transaction capacity to scale to better than linear with computational capacity of the n nodes.
-
Sharding FAQ
-
Stateless clients
-
State channels
-
Plasma implementation
-
Data availability proofs [65%]
-
Effective state-space partitioning / Cross-shard communication [15%]
-
High-Level-Languages [20%]
- Topic: Developing a language that knows to send the cross-shard asynchronous messages whenever contracts are located on different shards.
-
Topic: Applying prior theory from multicore CPUs/parallel threading to sharding.
1.2 Proof of Stake [70% complete]
Goal: Fully transition Ethereum from Proof-of-work to Proof-of-stake.
1.3 Protocol Economics [50%]
Goal: Increase economic incentive confluence in all aspects of the Ethereum protocol.
1.4 Stategies for efficaciously hardforking for upgrades [40%]
Goal: Smart-contracts are new territory and the best ideas in the space remain undiscovered. When we discover them, we must be able to roll them out gracefully.
1.5 Ethereum Virtual Machine (EVM) upgrades and optimization [100%]
Goal: Have a fast, efficient virtual machine optimized for processing cryptographic operations and smart-contracts.
Update: Solved! We’re moving to eWASM!
2. Layer 2
2.1 On-chain Random Number Generation [63%]
Goal: This is an important special-case necessary for many applications. We wish to solve it.
2.2 Privacy [40%]
Goal: Allow apps to benefit from the transparency of blockchain-execution while preserving author privacy and the confidentiality of zer data. One solution, among several, is homomorphic encryption.
-
General: Privacy on the Blockchain
-
Mixers [30%]
Bitcoin mixing remains an unsolved problem. As what’s possible in Ethereum is a strict superset of Bitcoin, solving for either case is sufficient. Incentivized Mixing?
-
Voting [10%]
-
Zero knowlege proofs [30%]
-
Other
2.3 Decentralized exchanges [50%]
Goal: We wish to minimize the necessity of trusted third parties in currency exchanges.
2.4 High-level-languages (HLLs) [40%]
Goal: Coding contracts (especially secure ones!) is hard. It should be easier. Please help us.
-
Our packet for recruiting PLT researchers.
-
Languages
-
Formal Verification of HLLs [15%]
-
Other programming language techniques to analyse smart contracts
-
Defensive programming [30%]
2.5 Better Tokens, better token sales
Goal: Understand how to design and manipulate tokens for specific properties, particularly paying attention to better ICOs
3. Ecosystem Analytics
- Create analytics to add to platforms like ethstats.net. Particularly network-level measures quantifying decentralization, resistance to different attacks, etc.
- “Econophsysics” analyses of the ethereum economic ecosystem. Probably using data from etherscan.io.
- Data-science analyses. E.g., how much Ether is unspendable, how often does Ether change hands, etc.
Appendix
Relevant Conferences
Research communities whose interests intersect with Ethereum’s research include (in alphabetical order, non-exhaustive):
- Algorithmic Game Theory. ACM Conference on Economics and Computation, Conference on Web and Internet Economics, Symposium on Algorithmic Game Theory, International Conference on Game Theory
- Blockchain. Annual Blockchain Summit, Coinfest, Consensus, Internet of Things World, Workshop on Bitcoin and Blockchain Research
- Computer Security. ACM Conference on Computer and Communications Security, IEEE Computer Security Foundations Symposium, USENIX Security Symposium
- Cryptography. CRYPTO (International Association for Cryptologic Research), EUROCRYPT (Annual International Conference on the Theory and Applications of Cryptographic Techniques)
- Distributed Computation. ACM Symposium on Principles of Distributed Computing, ACM Symposium on Parallelism in Algorithms and Architectures
- Multi-Agent Systems. International Conference on Autonomous Agents and Multiagent Sytems, AAAI Conference on Artificial Intelligence, International Joint Conference on Artificial Intelligence
