Skip to Main Content
Large parallel-processing systems are necessary for the highest performance levels but economically wasteful if the investment in hardware, building space, electricity and manpower is not compensated by the possibility of network- and/or node-sharing. When such a system is shared by multiple users, mandatory access control (MAC) becomes a critical necessity. Considering the limitations of previous access control solutions, Toinard et al. have proposed a meta-policy approach whereby a statically-checked global policy controls dynamic changes to individual access rights. This approach combines the flexibility of decentralized and dynamic changes granted to users with the need for a global overview of information flow. This paper proposes an algorithmic solution harnessing the parallel-processing systempsilas own processing power.