Xlink

[michaelquigley]

It's looking like the underlying connectivity semantics between transwarp and the traditional transport-based protocols are different enough that we're not likely to contain all of the differences within the transport framework. In other words, the transwarp abstraction is likely to leak out of the transport implementation.

So... we're probably going to need some kind of Xlink framework, encapsulating the behavior of link management for the overlay. The current implementation will likely become xlink_transport, and we'll end up developing an xlink_transwarp.

[michaelquigley]

An added side-effect of Xlink is that it will also allow for multiple link listeners per router, which will be required to support networks with multiple link protocols (#45).

[michaelquigley]

Will evaluate and continue to polish throughout the xlink_transwarp implementation.