Best Way To Add Handshake DNS Resolution Support For Dapps (Discussion)

[flikites]

This can be approached two ways:

1. Adding support for applications/containers to be deployed to the flux network with a specified --dns flag:

https://docs.docker.com/config/containers/container-networking/#dns-services

This feature would allow a prospective customer to deploy their own hsd resolver container and specify the 2nd container (say a vpn) to use the DNS server deployed in the 1st container. (Bypassing the DNS server that the Flux Node uses in /etc/resolv.conf

2. Mandate node operators to deploy HSD with the recursive resolver running publicly on port 53, set up the systemd file to run HSD as a service, and then configuring their Flux node to use the newly deployed DNS server in /etc/netplan

I think number 1 seems more feasible and doesn't create friction.

While number 2 would create friction, it would create more distributed and decentralized DNS landscape for the network.

I have the guide for number 2, my thoughts initially we're to try and figure out how to incentivize operators for taking the extra steps to have decentralized DNS running on their nodes.

HSD currently requires roughly 60 GB of storage space.

Update: We can run the full node in prune mode and only ever use 400 MB:

Pruned full node

This will never require more than 400 MB total. Only the last 288 blocks are saved to disk. Wallet rescans are impossible in this mode. This node will not relay historical blocks to new bootstrapping nodes but otherwise is fully validating and just as private and secure as any full node.

Handshake DNS Documentation: https://hsd-dev.org/