ZeroConf: Service Discovery with mDNS

ZeroConf is a pure Golang library that employs Multicast DNS-SD for

• browsing and resolving services in your network
• registering own services

in the local network.

It basically implements aspects of the standards RFC 6762 (mDNS) and RFC 6763 (DNS-SD). Though it does not support all requirements yet, the aim is to provide a compliant solution in the long-term with the community.

By now, it should be compatible to Avahi (tested) and Apple's Bonjour (untested). Target environments: private LAN/Wifi, small or isolated networks.

Install

Nothing is as easy as that:

$ go get -u github.com/viamrobotics/zeroconf

This package requires Go 1.7 (context in std lib) or later.

Browse for services in your local network

// Discover all services on the network (e.g. _workstation._tcp)
resolver, err := zeroconf.NewResolver(nil)
if err != nil {
    log.Fatalln("Failed to initialize resolver:", err.Error())
}

entries := make(chan *zeroconf.ServiceEntry)
go func(results <-chan *zeroconf.ServiceEntry) {
    for entry := range results {
        log.Println(entry)
    }
    log.Println("No more entries.")
}(entries)

ctx, cancel := context.WithTimeout(context.Background(), time.Second*15)
defer cancel()
err = resolver.Browse(ctx, "_workstation._tcp", "local.", entries)
if err != nil {
    log.Fatalln("Failed to browse:", err.Error())
}

<-ctx.Done()

Lookup a specific service instance

// Example filled soon.

Register a service

server, err := zeroconf.Register("GoZeroconf", "_workstation._tcp", "local.", 42424, []string{"txtv=0", "lo=1", "la=2"}, nil)
if err != nil {
    panic(err)
}
defer server.Shutdown()

// Clean exit.
sig := make(chan os.Signal, 1)
signal.Notify(sig, os.Interrupt, syscall.SIGTERM)
select {
case <-sig:
    // Exit by user
case <-time.After(time.Second * 120):
    // Exit by timeout
}

log.Println("Shutting down.")

Credits

Great thanks to hashicorp and to oleksandr and all contributing authors for the code this projects bases upon. Large parts of the code are still the same.

However, there are several reasons why I decided to create a fork of the original project: The previous project seems to be unmaintained. There are several useful pull requests waiting. I merged most of them in this project. Still, the implementation has some bugs and lacks some other features that make it quite unreliable in real LAN environments when running continously. Last but not least, the aim for this project is to build a solution that targets standard conformance in the long term with the support of the community. Though, resiliency should remain a top goal.