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
{{ message }}
This repository has been archived by the owner on Oct 5, 2022. It is now read-only.
Ok nice idea! I thought about it when I started to build this.
Using database (postgresql is only supported currently) to manage nodes makes it harder to join our network. So I separated the logics to cloud.
If we manage it on memory, fetching and syncing huge nodes would be big bottleneck.
Cloud will acquire the high availability by #7.
Multiple cloud components can be behind a load balancer by this issue.
Maybe we can use DHT to solve this problem, pg integrated here seems too heavy to this kind of network application, if we can merge 'cloud' and 'node' into each p2p node, rather than centralized cloud point, we can defense more attacks and censorship. Each gateway ran by every end user can connect to a random node with some nodes status information in its DHT, and we can build a new load balance map behind each user gateway.
Can we accomplish it without passing through packets to any nodes except the nodes included in the route?
I mean, as my understanding, node must send a request to other nodes to know the node corresponding to the hash. Which would be vulnerability like traffic analysis attack.
But, yeah, decentralizing is one of the good approach to protect from attackers! 👍
We could change the it if we could guarantee the security.
Thanks! Let's discuss many things!
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
If the cloud part is failed, no new nodes can connect to it and participate in its network.
I think cloud part can be p2p also.
The text was updated successfully, but these errors were encountered: