dragonfly thread

dragonfly thread
7 Replies
j0rge
j0rgeOP•4w ago
GitHub
Implement local lan sharing of updates · Issue #773 · ublue-os/main
We used to have an issue like this but one meme led to another ... p5 and I were thinking if we could set up dragonfly on a node that we could have machines share their updates with each other. All...
j0rge
j0rgeOP•4w ago
Mark Russel (podman PM) was thinking the same thing us and timothee are going to talk to the dragonfly maintainers cc @Kyle Gospo this isn't as hard as it sounds, it all exists.
James
James•4w ago
This would go so hard
Robert
Robert•4w ago
If a user can opt-in to having a single machine act as the primary manager node, or somehow be able to autodiscover existing managers on a network and join the cluster (essentially making every node a manager and seed), it could be awesome. Encryption is the biggest unknown for me though.
j0rge
j0rgeOP•4w ago
Keep posting questions so I know what to ask lol
James
James•4w ago
Oof if it ran on bittorrent? idk what dragonfly is or what p2p stuff they do. But a global DHT thingy would be cool for individuals who want fast updates. I've only been thinking about the enterprise usecase propagating updates over a LAN/VPN
Robert
Robert•4w ago
No - still talking about Dragonfly and LAN 🙂 Dragonfly uses the term "seed" too
Manager: Maintain the relationship between each P2P cluster, It primarily offers functions such as dynamic configuration management and data collection. It also includes a front-end console, enabling users to visually operate and manage the cluster. Scheduler: Select the best download parent node for the download node. At the appropriate time, trigger Seed Peer to perform back-to-source downloading, or Peer to perform back-to-source downloading. Seed Peer: Provides upload and download functions and can serve as a root node in the P2P network, allowing the Scheduler to actively initiate back-to-source. Peer: Provides upload and download functions.
We'd want all nodes to act as the Seed Peer, and either a single machine acting as the Manager and Scheduler, or preferred, distributing them across all nodes on the local network

Did you find this page helpful?