jvm-libp2p

a libp2p implementation for the JVM, written in Kotlin 🔥

⚠️ This is heavy work in progress! ⚠

Roadmap

The endeavour to build jvm-libp2p is split in two phases:

• minimal phase (v0.x): aims to provide the bare minimum stack that will allow JVM-based Ethereum 2.0 clients to interoperate with other clients that rely on fully-fledged libp2p stacks written in other languages.

  • To achieve this, we have to be wire-compliant, but don't need to fulfill the complete catalogue of libp2p abstractions.
  • This effort will act as a starting point to evolve this project into a fully-fledged libp2p stack for JVM environments, including Android runtimes.
  • We are shooting for Aug/early Sept 2019.
  • Only Java-friendly façade.

• maturity phase (v1.x): upgrades the minimal version to a flexible and versatile stack adhering to the key design principles of modularity and pluggability that define the libp2p project. It adds features present in mature implementations like go-libp2p, rust-libp2p, js-libp2p.

  • will offer: pluggable peerstore, connection manager, QUIC transport, circuit relay, AutoNAT, AutoRelay, NAT traversal, etc.
  • Android-friendly.
  • Kotlin coroutine-based façade, possibly a Reactive Streams façade too.
  • work will begin after the minimal phase concludes.

minimal phase (v0.x): Definition of Done

We have identified the following components on the path to attaining a minimal implementation:

• multistream-select 1.0
• multiformats: multiaddr
• crypto (RSA, ed25519, secp256k1?)ç
• connection bootstrapping
• secio
• mplex as a multiplexer
• stream multiplexing
• TCP transport (dialing and listening)
• identify protocol
• peer ID

We are explicitly leaving out the peerstore, DHT, pubsub, connection manager, etc. and other subsystems or concepts that are internal to implementations and do not impact the ability to hold communications with other libp2p processes.

License

Dual-licensed under MIT and ASLv2, by way of the Permissive License Stack.