In agreement with nit comments in the Gen-Art review. 1) Section 2. Background seems to be a duplication with the introduction section and could probably be merged with this section. 2) In general I wander wether the requirement level "SHOULD" is to week in some places. The concern (and question) here is whether this may lead to uncertainty whether a Split-DNS configuration always will provide the expected level of security (or fail), or wether such configuration may lead to successful communication without the expected level of security ( in compliance with this specification).