臺灣博碩士論文加值系統

IOT的發展不斷躍進，也進而促進了叢集式分散系統的興起，透過機器間彼此的分工合作，以達到運算、儲存、系統維護上能接近硬體設備極好的系統甚至在設備容錯方面更勝一籌，而叢集分散系統大部分的機器都具備運算、儲存、控制以及具備網路功能，因此可被視為一個邊緣運算節點。
透過邊緣運算節點我們能將雲端服務轉移至節點內，藉此可以降低實際傳輸距離以及減少網路傳輸資訊量與無法預測之網路壅塞，也能提高硬體資源利用率，並且將其與虛擬機器(VM)結合不但有獨立的空間，也能自行建構出服務所需的執行環境，並且讓多個任務同時在同一台機器做運行並且不會互相干擾，但VM一般需要切割出獨立的硬體資源，因此常會有硬體資源不足的問題，並且利用邊緣節點協作完成一件任務會需要節點間的同步(Synchronization)，以取得任務內容切割或執行時間限制等。
本論文將上述提到之問題轉換為Sigle Pub - multiple Sub communication的問題，因此為了解決上述問題本文選用虛擬容器Linux Container (Docker)作為邊緣節點內之執行環境架構，虛擬容器彼此間不但能共享系統硬體資源，還可以將執行環境從系統中隔離出來，因此在執行上有更輕量化及更快的啟動速度，以突破VM硬體切割上的限制，並且結合SSH File System (SSHFS)以解決邊緣節點間Synchronization的問題。
鑒於這些架構本論文，第一，想以2019年屏東燈會無人機排字為例，模擬出一個能透過無線網路發送任務給集群下的機器，讓指定數量或特定的集群機器去執行任務內容並且回傳執行結果以模擬邊緣式計算。
第二，透過模擬火災自動路徑規劃逃生系統，讓邊緣節點間互相溝通以找出距離現在位置出口的最佳路徑，來製作協作式系統。

The development of IOT has continued to leap forward promoted the rise of cluster-distributed systems. Through the division of labor between machines, it is possible to achieve an excellent system for computing, storage, and system maintenance that can be close to expensive hardware devices, even better in terms of device fault tolerance. As a result, most of the machines in the cluster decentralized system have operations, storage, control, and networking capabilities, so they can be considered as an edge computing node.
Through the edge computing node, we can transfer the cloud service to the node, which not only reduce actual transmission distance, network transmission information but also avoid unpredictable network congestion. It can also improve the use of hardware resource, and combine with Virtual machine (VM) not only provides independent space, but also we can customized system environment for the specific service, and allows multiple tasks to run on the same machine at the same time without interfering with each other. But the VM generally needs to independent hardware resource to execute system, so there are often problems with insufficient hardware resources, and when we using edge nodes to collaborate on a task will require synchronization between nodes, also face task execution time limits.
This paper converts the all problem which mentioned on the above to the problem of Sigle Pub - multiple Sub communication. Therefore, in order to solve the above problems, the Linux Container (Docker) is selected as the execution environment architecture in our edge node, the virtual containers not only share the hardware resources with each other but also isolate the execution environment from the system, so there is a lighter weight and faster startup speed to break through the limitations of VM, and combined with SSH File System (SSHFS) to solve The problem of Synchronization between edge nodes.
In view of the structure of this paper, First, I want to use the screen of the 2019 performance of drones in Pingtung as an example to simulate a machine that can send tasks to the cluster through the wireless network, allowing a specified number or specific cluster machines to perform task content. And return the execution results to simulate the edge calculation.
Second, simulated fire automatic path planning escape system, let the edge nodes communicate with each other to find the best path from current position to exit door to simulated a collaborative system.

論文審定書 i
誌謝 ii
摘要 iii
Abstract iv
目錄 vi
圖次 ix
表次 xii
第一章 序論 1
1.1 研究動機 1
1.2 問題描述 3
1.3 文獻探討 4
1.3.1 Multicast 4
1.3.2 The Dynamic Discover Problem 6
1.3.3 Shared Queue (Dealer and Router Sockets) 8
1.3.4 Heartbeating 10
1.4 論文架構 13
第二章 研究背景 14
2.1. 協作式邊緣計算系統 14
2.1.1 何謂協作式邊緣計算 14
2.1.2 IOT與協作式邊緣運算 14
2.1.3 Single Pub – Multiple Sub 15
2.2. Linux Container (LXC) 17
2.3. Docker 19
2.3.1 簡介 19
2.3.2 Docker registries 21
2.3.3 Dockerfile 21
2.3.4 Docker Swarm 22
2.3.5 Docker Libnetwork (CNM) 24
2.3.6 Docker Overlay 26
2.3.7 Vxlan 27
2.3.8 安裝與啟動 31
2.4 SSH File System (SSHFS) 32
2.5 Ascii art 32
2.6 LoRa 34
第三章 問題解決 36
3.1. 系統需求 36
3.1.1 叢集硬體 36
3.1.2 Docker安裝 37
3.1.3 Docker Swarm建構 38
3.1.4 SSHFS安裝及設定 39
3.2. 模擬架構 40
3.2.2 無人機排字模擬 40
3.2.3 火災自動路徑規劃逃生系統 42
第四章 系統實作 44
4.1. 環境製作 44
4.1.1 製作Dockerfile並上傳 44
4.2. 模擬製作 46
4.2.1 邊緣計算實現(無人機展演) 46
4.2.2 協作式運算實現(智慧逃生指示燈) 54
第五章 現有系統比較 59
5.1 無人機展演 59
5.2 智慧逃生指示燈 60
第六章 結論以及未來展望 61
參考文獻 62