Today, Flow has 440+ staked validators in the ecosystem, which secure the protocol and ensure its resistance to censorship from corporations, jurisdictions, or individuals. Validators also help protect Flow’s open access, decentralized nature, and resilience to cyber attacks and natural disasters. Nevertheless, there are still more defenses to be added in the future so that Flow automatically detects attacks and the protocol can autonomously defend itself against malicious nodes within the network. Within a broader context, this is a continuation of the recent evolutionary steps taken with Flow’s light client and consensus upgrade.
Initiatives
Flow nodes need the ability to generate cryptographic proofs as evidence for the consensus committee to slash byzantine peers. Through the combination of network layer hardening, Dynamic Protocol State, and message forensics, honest nodes are enabled to maintain a healthy network and autonomously control and suppress byzantine behavior.
Dynamic Protocol State is the foundation for countermeasures such as slashing misbehaving nodes or entirely revoking their authorization to participate. In a nutshell, the protocol state tracks the blocks produced by the network – and for each block, its finality and sealing status – as well as the nodes permitted to be part of the network as of that block. The protocol state is collectively maintained by all nodes in the network and protected by Flow's consensus mechanism. Each node replicates the protocol state locally and only accepts updates if they have been approved by consensus.
Currently, the set of nodes allowed to participate is fixed for an entire epoch within the implementation, which is too limiting. After implementing the dynamic protocol state, the information about participating nodes can be updated dynamically throughout every epoch. This will enable nodes to revoke their keys if they’re suspected of being compromised or for the network to autonomously slash nodes for protocol violations.
Flow’s multi-node, pipelined architecture is the foundation for its unmatched scalability and efficiency. At the same time, having specialized node roles also means that there is a highly diverse set of messages exchanged at runtime. Established networking libraries for blockchains – like libP2P, which Flow uses – are not equipped to protect the higher-level application logic from spamming attacks or to rank a peer's reliability across multiple communication topics heuristically based on the traffic received from them.
As part of network layer hardening, we encompass the defense and mitigation strategies described in this research paper and a unified interface to rank peers holistically. At completion, a node will independently score its peers based on their reliability and protocol-compliant behavior. Nodes will autonomously limit and reduce resources available to peers overusing libP2P control messages, sending repetitive requests or broadcasts, or sharing messages that the protocol layer classifies as invalid or malicious. As a result, the honest supermajority becomes resilient against a byzantine minority trying to consume network resources exploitatively.
Message forensics is another initiative that will start soon. If individual nodes observe protocol violations, the message forensics framework will enable them to generate cryptographic proofs as necessary evidence for the consensus committee to slash byzantine peers.
Current progress
Over the last months, the core protocol team has made significant progress on the Dynamic Protocol State: R&D phase with a detailed software design proposal has concluded. The implementation is currently ongoing and expected to complete within the next few months.
Earlier this year, the network reached a significant milestone for attack-resilient message propagation by becoming fully permissionless for developers - which allows autonomously operation of Access nodes. Also, this milestone laid the foundation for the anonymous participation of other node types.
Lastly, Cryptographic proofs of protocol violations are still at an early stage. While research is at an early stage, the development is likely to be intricate with many requirements to consider.
Roadmap goals
Keep reading
Quickstarts
Stay up to date with the latest news on Flow.
Stay up to date with the latest news on Flow.