The Harmony protocol roadmap lays out the milestones and major features that the Harmony engineering team will focus on in the foreseeable future to make Harmony blockchain a fully decentralized, secure and high-performance network that’s capable of serving millions of users.
*Written by Rongjian Lan. (To be) Approved by Leo Chen, Ganesha Upadhyaya.
Features and Tools
Harmony as a developer and user-focused blockchain platform strives to offer the most advanced and comprehensive feature set for easier development of Dapps and smoother user experience. For example, cross-shard smart contract messaging protocol will enable the communication between shards and improve the utility of multiple shards. EIP-1559 and account abstraction will improve the efficiency of the transaction gas market and bring more flexibility to how users interact with the blockchain. Data availability support will be added to shards so it can be used to better serve layer 2 applications.
Short Term (Q3 2022)
- Cross-shard smart contract messaging protocol
- Router smart contract precompile
- Outgoing and incoming message box within contract storage
- Receiver logic and cross-shard transaction gas fee model
- Cross-shard messaging SDK
Reader staking precompiled contract to support staking derivativesUpdateChainIDto support ethereum-compatible meta-transactionRevamp transaction pool logic to have better ordering and gas suggestion.
Mid Term (Q1 2023)
- Support EIP-1559 for better transaction fee market
Long Term (Q4 2023 and beyond)
- Account abstraction implemented within smart contract layer for relayer-free and gas-less smart contract wallets
- Data availability shards to boost layer 2 security and performance
- Verkle tree computation and state root addition in block header
- KZG polynomial commitment for efficient data availability checking
- Erasure encoding of block content
- Stateless client
- Migrating current Merkle-Patricia-Tree state trie to Verkle tree.
Security and Decentralization
There is no consensus without security and decentralization on Harmony network. We will continue strengthen our protocol design and add more security mechanism to achieve full decentralization in the long term. Attack vectors on the leader node will be fully eliminated and leader rotation will be implemented to prevent leader malicious behaviors. Token incentive model will be further optimized to achieve incentive-compatible gas fee and block reward distribution. The final design for resharding will be finished along with our Research DAO and its launch will enable the full decentralization of Harmony network.
Short Term (Q3 2022)
- Slashing rule update to avoid collusion of validator and reporter
- Lower validator uptime requirement to avoid malicious leaders forcing honest validators out of election
- Leader rotation within Harmony nodes
Mid Term (Q1 2023)
- Leader rotation among all validator nodes
- Double signing prevention on validator nodes
- Update block reward and transaction fee distribution logic to incentivize leader to include more signatures and more transactions
Long Term (Q4 2023 and beyond)
- Resharding
- Bounded-cuckoo-rule election and shuffling to support full security of sharded network.
- Validator node state trie downloading/verification and shard transition.
- Full decentralization with 100% external validators
- Transaction censorship resistance coupled with view change protocol
- Stale account state pruning
Efficiency and Performance
The global consensus and open economy requires an extremely fast and scalable blockchain and Harmony will be the one. Our goal is to have a stable network where general low-end machines can participate as a validator. For example, Block and state synchronization will be speed up and decentralized to avoid sync issue in our fast finality block production network. Our node binary will be further optimized in memory, disk, and CPU so low-end machine can also run as leader node. On protocol level, crosslink sending will be batched to lower the burden on beacon chain. P2p network topology will be optimized based on the shard structure to speed up message broadcasting time within shards.
Short Term (Q3 2022)
Fast synchronization of beacon chain with only epoch blocks- Decentralized block synchronization using libp2p stream channel
- Faster block synchronization
- Switch levelDB with a faster database (Erigon’s implementation)
- Cache state transition results and batch commit blocks.
- Profile and optimize block commitment logic
- Validator node history block and related db storage pruning.
Mid Term (Q1 2023)
- Migrate BLS library from herumi to blst with eth-compatible BLS12-381 base point.
- Aggregate crosslinks and send batched crosslink data every 64 blocks
- Bring all consensus-related off-chain data onchain to avoid data consistency issues
- Last continuous crosslink number
- Delegation indexes
- Validator wrapper snapshots
- List of all validator addresses
Reduce and optimize the caches across the codebase.Blockchain database cacheRPC and consensus cache
Long Term (Q4 2023 and beyond)
- ZK-based fast state verification
- Replace RLP (recursive length prefix) encoding with SSZ (Simple SerialiZe) encoding
- Update p2p network to be sharded in connectivity and network topology.
- Only broadcast block header with transaction id list and let validator reconstruct the full block.
- Break down validator wrapper data into individual state object key-value entries.