It was decided in Quarkus that for a consistent user experience we actually wanted a HTTP stack based on Vert.x, so the existing Netty work was modified to have a connector SPI, and then a Vert.x implementation of this (leaving open the possibility of a pure Netty implementation as well). 

My gut feeling is that even though the Vert.x backend makes sense for Quarkus this does not make sense for Undertow as a standalone project, so we are going to continue this work as 'Undertow 3.0' we would need a pure Netty backend as well as the existing vert.x one. Even though it is based on the same code base I am not sure if it actually makes sense to try and move this code back to the Undertow project, it may make more sense to just consider it a Quarkus specific fork. It would be good to get some feedback from the community about how to handle this, as I am not really sure what the best solution is here.