Abstract | ||
---|---|---|
In designing a distributed hard real-time system, it is important to reduce the end-to-end delay of each real-time message in order to ensure quick responses to external inputs and a high degree of synchronization among cooperating actuators. In order to provide a real-time guarantee for each message, the related literature has focused on the analysis of end-to-end delays based on worst-case task phasing. However, such analyses are too pessimistic because they do not assume a global clock. With the assumption that task phases can be managed by using a global clock provided by emerging real-time fieldbuses, such as EtherCAT, we can try to calculate the optimal task phasing that yields the minimal worst-case end-to-end delay. In this study, we propose a heuristic to manage the phase offsets in the distributed tasks to reduce the theoretical end-to-end delay bound. The proposed heuristic reduces the search time for a solution by identifying time intervals where actual communication occurs among inter-dependent tasks. Furthermore, to analyze the distribution of end-to-end delays in different phases, we implemented a simulation tool. The simulation results showed that the proposed heuristic can reduce worst-case end-to-end delay as well as jitter in end-to-end delays. |
Year | DOI | Venue |
---|---|---|
2017 | 10.15837/ijccc.2017.5.2894 | INTERNATIONAL JOURNAL OF COMPUTERS COMMUNICATIONS & CONTROL |
Keywords | Field | DocType |
task phasing, end-to-end delay, real-time, fieldbus, EtherCAT | End-to-end delay,Heuristic,Synchronization,Computer science,EtherCAT,Phaser,Real-time computing,Jitter,Periodic graph (geometry),Distributed computing,Fieldbus | Journal |
Volume | Issue | ISSN |
12 | 5 | 1841-9836 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Sang-Hun Lee | 1 | 169 | 25.36 |
Hyun-Wook Jin | 2 | 383 | 41.06 |
Kanghee Kim | 3 | 223 | 20.24 |
Sangil Lee | 4 | 1 | 1.03 |