This paper proposes an active networks architecture which supports both integrated and discrete operationmodel based on mobile Agent technology. Mobile Agent is the component for customized function transferring, acti...This paper proposes an active networks architecture which supports both integrated and discrete operationmodel based on mobile Agent technology. Mobile Agent is the component for customized function transferring, activenode provides software application layer, and Agent server processes mobile Agent specific customization. For devel-opment and deployment of specific application protocol, this paper also proposes an abstract protocol framework and aprotocol loading mechanism to enhance network performance.展开更多
In this paper, we describe a two-phase scheme to determine a scheduling of tasks with precedence con-straints that employ a reliability measure as one of the objectives in a Real-Time and heterogeneous distributed sys...In this paper, we describe a two-phase scheme to determine a scheduling of tasks with precedence con-straints that employ a reliability measure as one of the objectives in a Real-Time and heterogeneous distributed sys-tem. The simulation results show that, for task graphs with precedence constraints in a heterogeneous Real-Timesystem, the method performs significantly better than the two algorithms presented that do not consider reliabilitycost. Furthermore, the results suggest that higher computational heterogeneity is conducive to improving the schedu-lability of the reliability cost-driven (RCD)algorithm, while the opposite is true for the two non-RCD algorithms.展开更多
文摘This paper proposes an active networks architecture which supports both integrated and discrete operationmodel based on mobile Agent technology. Mobile Agent is the component for customized function transferring, activenode provides software application layer, and Agent server processes mobile Agent specific customization. For devel-opment and deployment of specific application protocol, this paper also proposes an abstract protocol framework and aprotocol loading mechanism to enhance network performance.
文摘In this paper, we describe a two-phase scheme to determine a scheduling of tasks with precedence con-straints that employ a reliability measure as one of the objectives in a Real-Time and heterogeneous distributed sys-tem. The simulation results show that, for task graphs with precedence constraints in a heterogeneous Real-Timesystem, the method performs significantly better than the two algorithms presented that do not consider reliabilitycost. Furthermore, the results suggest that higher computational heterogeneity is conducive to improving the schedu-lability of the reliability cost-driven (RCD)algorithm, while the opposite is true for the two non-RCD algorithms.
