Access control mechanisms in a distributed, persistent memory system

A distributed, persistent memory system is considered, which implements a form of segmentation with paging within the framework of the single-address-space paradigm of memory reference. A peculiar problem of a system of this type is the lack of protection of the private information items of any given process against unauthorized access attempts possibly performed by the other processes. We present a set of mechanisms able to enforce access control over the private virtual space areas. These mechanisms guarantee a degree of protection comparable to that typical of a multiple-address-space system, while preserving the advantages of ease of information sharing, typical of the single-address-space model. The resulting environment is evaluated from a number of salient viewpoints, including ease of distribution and revocation of access rights, strategies for virtual space reuse, and the storage requirements of the information for memory management.