The current MAP BR logic (/plugins/map) parses certain attributes of runtime-defined rulesets and stores in hashed tables for later lookup.

Presently, the IPv6 version of this hash table uses the (typically /64) IPv6 Default Mapping Rule as the seed of this hash. Consequently, even though nothing precludes the use of the same DMR in multiple MAP rules per spec, the implementation forces uniqueness of this attribute among defined rules (Actually, it will just incorrectly route traffic runtime without error).

The proposed feature update will alter the mechanism used to generate the hashed rule lookup table to leverage the defined IPv6 rule prefix (generally a /56, /48, or even larger) as these values will indeed always be unique and non-overlapping amongst MAP rules configured per RFC.