Edge/cloud resource management for time-sensitive applications
The Internet of Things (IoT) is one of the most popular technology trends to have emerged in recent years. Most IoT systems require cloud computing to assist in communicating and storing data between devices. While clouds are powerful for storing and processing, it creates delays in IoT devices comm...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
Nanyang Technological University
2021
|
| 主題: | |
| 在線閱讀: | https://hdl.handle.net/10356/148074 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 總結: | The Internet of Things (IoT) is one of the most popular technology trends to have emerged in recent years. Most IoT systems require cloud computing to assist in communicating and storing data between devices. While clouds are powerful for storing and processing, it creates delays in IoT devices communicating with each other. By decentralizing cloud computing in the form of edge and mobile computing, task computation and storage are located closer to the end users, which alleviates the problem of latency, bandwidth, and data privacy. Thus, the task schedulers in this cloud/edge system play a key role in managing the activities of this system.
This project aims to simulate a cloud/edge environment for testing different task scheduling algorithm. An open-source simulation toolkit called CloudSim Plus, which runs on Java, is used to implement this system. This simulation environment simulates the core functionality of the cloud, such as job/task queue, events processing, broker policy implementation, and the communication between different entities.
Several deadline aware task scheduling algorithms have been implemented in the simulation. CloudSim Plus creates a task with characteristics similar to a real cloud system task, such as length, bandwidth, size, etc. However, the time constraint is not one of them, and it is not considered in scheduling the task queue. Therefore, new settings to the CloudSim Plus to help the scheduler aware of tasks’ deadline is implemented.
Effectiveness and performance comparison between implemented scheduling algorithms are conducted through experiments. These experiments compare the waiting time, missed deadlines count, percentage of tasks scheduled. Overall, the simulation is able to show the effectiveness and performance of tasks scheduling algorithms in a real cloud-based system. |
|---|