This paper proposes a method called implicit encapsulation for creating layered stimulus models for use in constrained random verification environments. The proposed method relies heavily on the VMM class library – the atomic generator and scenario generator in particular – along with techniques typical of transaction level modeling. A set of generic components and a scalable structure is described that enables the creation of layered protocol stacks in a variety of arrangements. The intended audience includes those verifying medium to large RTL blocks that process layered protocols.
