search

nasa / bplib

Apache License 2.0
27 stars 13 forks source link

Incomplete data flow identification when handling ACS bundles #10

Open jpswinski opened 5 years ago

jpswinski commented 5 years ago

Currently bplib only checks that the destination node and service of an ACS bundle matches the current channel's node and service numbers; and if it does, it goes ahead and processes the ACS.

This is problematic if the current channel's node and service is connected to multiple endpoints. If that is the case, then the channel needs to check that source node and service of the ACS also matches the destination node and service of the channel.

This also has implications for how routing is handled.

jpswinski commented 4 years ago

Current thinking is that a source node and service could be used to uniquely identify a flow and the channel will maintain unique CIDs per node.service.