Skip to content

Each connection should run in its own process #24

@robinhilliard

Description

@robinhilliard

I almost did it this way but ran into the problem of sharing the routing table and over corrected, creating a single process bottleneck. When we spawn the connections they should remain the controlling process of their respective ports, we can eliminate the double handling/delegation of incoming network messages.

Router scope should reduce to managing the routing table for these processes. Does routing info need to be updated every message or just with heartbeats? Connection processes delegate to router functions to forward directly to outgoing connections. Should local subscriptions get registered as extra connections?

Move the main API and entry point to a new MAVLink module.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions