[j-nsp] Next-table performance

[Nathan Ward]

Hello,

Does anyone know how next-table behaves, in terms of packet processing path and such?

I’m looking at using it for a reasonable volume of traffic, but before I commit to that I want to understand if there’s a performance hit, and, what that is.
The intent is to use it to point a default to a VRF with a big route table, rather than importing those routes in to several VRFs.
Does it mean packets are (partially?) processed twice, or, does it mean that routes are installed twice in to the FIB - or something else?

I couldn’t find anything about this in Juniper docs or on Google - but - may have been hitting the wrong search terms.

--
Nathan Ward