You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Our project currently uses Ethereum's discovery protocol, discv5, for node discovery between peers. While discv5 functions as intended, we are encountering limitations and issues because discv5 is specifically designed for Ethereum nodes, and our network is not part of the Ethereum ecosystem.
Issues with discv5:
Network Compatibility: discv5 is optimized for Ethereum nodes and networks, which can lead to inefficiencies and suboptimal performance in non-Ethereum environments.
Limited Customization: The protocol’s tight coupling with Ethereum’s specific requirements limits our ability to customize it for our network's unique needs.
Potential Scaling Concerns: As our network grows, the scalability of discv5 outside of Ethereum’s context might pose challenges.
Proposal:
To address these issues, I propose we consider switching to libp2p’s discovery mechanisms, which offer a more flexible and modular approach to peer discovery. Libp2p is designed to be network-agnostic, making it a potentially better fit for our project’s needs.
Benefits of libp2p:
Modular Design: Allows us to tailor the discovery protocol to our specific requirements, independent of any particular blockchain network.
Cross-Network Compatibility: libp2p is widely used across various projects, providing robust support and a larger community for troubleshooting and enhancements.
Scalability: Proven to scale efficiently across different types of decentralized networks.
Open questions
Since this is a different discovery system (though based on similar primitive like kademlia dht for example), would this require forking? how hard will it be to use the existing identity scheme (enrs)?
Next Steps:
Evaluate the feasibility of migrating from discv5 to libp2p’s discovery.
Identify potential challenges in integrating libp2p with our existing infrastructure.
Conduct a proof-of-concept implementation to compare performance and reliability.
Request for Feedback:
Please share your thoughts on this proposal, including any potential concerns or suggestions for the migration process. If there are alternative discovery protocols we should consider, feel free to mention those as well.
The text was updated successfully, but these errors were encountered:
Description
Our project currently uses Ethereum's discovery protocol, discv5, for node discovery between peers. While discv5 functions as intended, we are encountering limitations and issues because discv5 is specifically designed for Ethereum nodes, and our network is not part of the Ethereum ecosystem.
Issues with discv5:
Proposal:
To address these issues, I propose we consider switching to libp2p’s discovery mechanisms, which offer a more flexible and modular approach to peer discovery. Libp2p is designed to be network-agnostic, making it a potentially better fit for our project’s needs.
Benefits of libp2p:
Open questions
Since this is a different discovery system (though based on similar primitive like kademlia dht for example), would this require forking? how hard will it be to use the existing identity scheme (enrs)?
Next Steps:
Request for Feedback:
Please share your thoughts on this proposal, including any potential concerns or suggestions for the migration process. If there are alternative discovery protocols we should consider, feel free to mention those as well.
The text was updated successfully, but these errors were encountered: