![Blast from the past
https://web.archive.org/web/20131219030753/http://vitalik.ca/ethereum.html
11
Basic Building Blocks
Network: Ethereum will run on its own network with a memory-hard proof of
work (not yet released) and a 60-second block time using single-level
GHOST (see http://www.cs.huji.ac.il/~avivz/pubs/13/btc_scalability_full.pdf)
to improve security and fairness with fast confirmation times. The restriction
to single-level is done for simplicity and because a very fast block time is
undesirable for other reasons - namely, blocks will potentially take a very
long time to evaluate, so high levels of waste are computationally
undesirable, and block validation time will be potentially very high variance.
Currency: [...]
Issuance model: [...]
Transactions: transactions in Ethereum will be simple, with one sender,
one recipient, a value, a fee and a message consisting of zero or more data
items that are integers in [ 0 ... 2^256 - 1] (ie. 32 byte values). All
transactions are valid; transactions where the recipient has insufficient funds
simply do nothing. Transactions sent to the zero address (ie. whose
hexadecimal representation is all zeroes) are a special type of transaction,
creating a "contract".](https://image.slidesharecdn.com/ethcc-eth2-tests-and-implementation-190312104411/85/Eth-cc-eth2-tests-and-implementation-11-320.jpg)
Eth cc --eth2-tests-and-implementation
- 1. Eth 2.0 Testing and Simulation 1
- 2. Hello!I am Mamy Ratsimbazafy Part of the Nimbus Eth1 and Eth2 implementer team at Status “Adolescent Full Mustang” @mratsim @m_ratsim 2
- 3. Where are we? https://github.com/ethereum/eth2.0-specs Phase 0 ● The beacon chain ● Status: ○ testnets incoming Responsibilities ● Coordination layer ● Sharding ● Block processing ○ Proof-of-Stake ○ Managing validators, shards, committees and attestations. ● Additional features ○ Finality ○ RNG ○ Cross-shard communication ○ Eth1 transition 3
- 4. What’s next? https://github.com/ethereum/eth2.0-specs Phase 1 ● The shard data chains ● Status: ○ Spec written, no implementation started. ● Responsibilities ○ Consensus over the data (e.g. account balance) Phase 2 ● The VM / execution layer ● Status: ○ In discussion (eWASM?) ● Responsibilities ○ Executing transactions (from simple transfers to smart-contracts) 4
- 5. What’s next? https://github.com/ethereum/wiki/wiki/Sharding-roadmap Phase 3 ● Light clients Phase 4 ● Cross-shard transactions 5
- 6. What’s next? https://github.com/ethereum/wiki/wiki/Sharding-roadmap Phase 5 ● Ethereum 2.0 endgame 6
- 7. What’s next? https://github.com/ethereum/wiki/wiki/Sharding-roadmap Phase 6 ● Super quadratic sharding 7
- 8. Diving into the beacon chain Credits: Hsiao Wei Wang, Ethereum Foundation 8 Beacon Chain provides random numbers Shard Chain provides data Shard 100 B1 B2 B3 B4 B5 VM provides state execution result Shard 1 Main Chain provides staking B1 state root B2 state root B3 state root B4 state root B5 state root
- 9. Diving into the beacon chain https://observablehq.com/@cdetrio/shasper-viz-0-4 Credits: Casey Detrio, Ethereum Foundation 9
- 10. Diving into the beacon chain Credits: Diederik Loerakker, @protolamba 10
- 11. Blast from the past https://web.archive.org/web/20131219030753/http://vitalik.ca/ethereum.html 11 Basic Building Blocks Network: Ethereum will run on its own network with a memory-hard proof of work (not yet released) and a 60-second block time using single-level GHOST (see http://www.cs.huji.ac.il/~avivz/pubs/13/btc_scalability_full.pdf) to improve security and fairness with fast confirmation times. The restriction to single-level is done for simplicity and because a very fast block time is undesirable for other reasons - namely, blocks will potentially take a very long time to evaluate, so high levels of waste are computationally undesirable, and block validation time will be potentially very high variance. Currency: [...] Issuance model: [...] Transactions: transactions in Ethereum will be simple, with one sender, one recipient, a value, a fee and a message consisting of zero or more data items that are integers in [ 0 ... 2^256 - 1] (ie. 32 byte values). All transactions are valid; transactions where the recipient has insufficient funds simply do nothing. Transactions sent to the zero address (ie. whose hexadecimal representation is all zeroes) are a special type of transaction, creating a "contract".
- 12. Beacon chain spec - initial commit https://ethresear.ch/t/convenience-link-to-casper-sharding-chain-v2-1-spec/2332 12
- 13. Life of a beacon chain implementer June-Sept 2018 - The HackMD + ethresear.ch period 13
- 14. Life of a beacon chain implementer Sept 2018-Feb 2019 - Living commit by commit 14
- 15. Life of a beacon chain implementer Feb 2019-Present - Spec releases! 15
- 16. What could go wrong? 16
- 17. What could go wrong? 17 Specs ● “Trivial” bugs (off-by-one) ● Edge cases (genesis, forced exits) ● Vulnerabilities (DOS attacks) ● Underspecification & interoperability (crypto & serialization) ● Slowness (shuffling) ● Incentives / Game Theory Implementation ● Overflow/underflow ● Slowness (using naive spec algorithm) ● Spec miscomprehension ● Vulnerabilities ● Implementation “details”
- 18. What could go wrong? Focus on implementation “details” 18 Cryptography P2P Networking Signature aggregation Serialization Consensus Sync State transition
- 19. What could go wrong? Practical case - shuffling 19
- 20. What could go wrong? Practical case - shuffling 20 - Understanding the shuffling algorithms - Determinism issues - Attack vector concerns - Performance concerns (light clients) - Out-of-bounds bugs (in the specs not even in implementations) And we have 9 teams, each implementing their own clients.
- 21. Why so many clients at launch? Unlike Eth 1.0 21
- 22. Client implementations Prysm (Prysmatic Labs, Go) ● https://github.com/prysmaticlabs/prysm Shasper (Parity Technologies, Rust) ● https://github.com/paritytech/shasper Trinity (Ethereum Foundation, Python) ● https://github.com/ethereum/trinity Yeeth (ZK Labs, Swift) ● https://github.com/yeeth/BeaconChain.swift Artemis (ConsenSys, Java) ● https://github.com/PegaSysEng/artemis Harmony (Harmony, Java) ● https://github.com/harmony-dev/beacon-chain-java Lodestar (ChainSafe System, Typescript / Javascript) ● https://github.com/ethereum/trinity Lighthouse (Sigma Prime, Rust) ● https://github.com/sigp/lighthouse Nimbus (Status, Nim) ● https://github.com/status-im/nim-beacon-chain 22
- 23. What do we do? Common testing repositories - https://github.com/ethereum/eth2.0-tests/ - Handcrafted and generated test vectors - https://github.com/ethereum/eth2.0-test-generators - Generate test vectors using Trinity reference implementation Status: test vectors for “relatively” stable and self-contained part of the spec - Crypto: BLS signatures - Shuffling - Serialization 23
- 24. What did we learn from Eth 1.0? 24 - A test repo that can be submoduled is good - Having no comment (json) in test files is bad - Having an insane amount of lines of code to review is worse
- 25. What did we learn from Eth 1.0? 25
- 26. What did we learn from Eth 1.0? 26
- 27. What did we learn from Eth 1.0? 27
- 28. Coming soon More unit tests - Merkle tree hashing - Fork choice (proof-of-stake) - Beacon state “god object” Client-specific testnets 28
- 29. TODO - cross-client testnet Pending - wire protocol: https://github.com/ethereum/eth2.0-specs/issues/593 Libp2p interop framework: https://github.com/libp2p/interop#interoperability-tests-for-libp2p Sharding P2P POC: https://github.com/ethresearch/sharding-p2p-poc/tree/master/docs 29
- 30. Simulations Kinds of simulations for Eth 2.0 - High-level overview - Full simulation with a real client (coming soon™) - Sharding simulations - Consensus simulations Some are in color ;) 30
- 31. Simulation - High-Level Overview https://observablehq.com/@cdetrio/shasper-viz-0-4 Credits: Casey Detrio, Ethereum Foundation 31
- 32. Simulation - Consensus https://github.com/leobago/shardSim Credits: Leo Bautista Gomez - Barcelona Supercomputing Center 32
- 33. Available research simulators ● Jannik Luhn - https://github.com/jannikluhn/sharding-netsim (sharding) ● Leo Bautista Gomez - https://github.com/leobago/shardSim (consensus) ● EF + libp2p + Whiteblock - https://github.com/ethresearch/sharding-p2p-poc (sharding) ● Consensys - https://github.com/ConsenSys/wittgenstein (consensus) ● Protolambda - https://github.com/protolambda/lmd-ghost (consensus) ● Vitalik ○ https://github.com/ethereum/research/tree/master/clock_disparity ○ https://github.com/ethereum/research/tree/master/ghost ● Whiteblock - https://github.com/zscole/nonce (whole blockchain, needs client) And all client teams coming soon™ 33
- 34. Looking for a place to start? 34
- 35. Thank you! 35