Scheduling on-demand broadcast with timing constraints
Journal of Parallel and Distributed Computing
| Hi-index | 0.00 |
There are many situations in which we need to incorporate timing constraints in broadcasting systems, such as for mobile real-time database systems. In a broadcasting environment with real-time constraints, the goal is to determine how well the scheduling algorithms ensure that the database server does not miss deadlines, instead of minimizing wait time. Future mobile real-time databases could be large scaled and may require accommodating a large population of users, and thus, require on-line scheduling algorithms that can efficiently handle user requests associated with timing constraints. Various work can be found on real-time transaction scheduling, as well as off-line real-time broadcast scheduling. However, very few studies address on-line broadcast strategies that consider timing constraints. This dissertation focuses on the study and design of efficient algorithms for on-line scheduling of mobile real-time broadcast systems. Efforts in this dissertation include the mathematical formulation of a real-time broadcast system, analysis of the system as a Markov Decision Process (MDP), the design of good broadcast strategies based on the analysis, and a performance evaluation of the proposed strategies. We propose two scheduling strategies, called Aggregated Critical Requests-ω (ACR-ω) and Aggregated Critical Requests-β (ACR-β), which are designed for timely delivery of data to mobile clients. Simulation results show that the ACR strategies outperform the existing algorithms when both the timing constraints and the benefits of broadcasting are considered.