Currently all routes obtained via BGP are written to default VPP table.

It might be useful for user to configure non default VRF.

To support it, some kind of BGP peer/AS to VRF mapping is needed

(depending whether iBGP or eBGP is being used).