DRNG Trees and RBGC Constructions (N22a)
Software implementations of random number generators (RNGs) may lack direct access to an entropy source and rely on other software or system components to seed and reseed their deterministic random number generators (DRNGs). Unlike other functionality classes and constructions, these DRNGs may be added and removed during operation in an ad-hoc fashion, forming a dynamic tree of dependencies. This talk will discuss rules and requirements for seeding, reseeding, composition, and operation of AIS 20/31 DRNG trees and SP 800-90C RBGC constructions, as well as examples highlighting the concepts presented.