基於分散式服務的可擴展容錯物聯網協同閘道器系統

Abstract

物聯網(IoT)是現今最熱門的議題之一，藉由種類繁多的傳感器(sensor)收集現實中的資訊，包括溫濕度、氣壓、亮度等等資訊。這些資訊可以藉由執行器(actuator)設定在不同的情況執行不同的動作，但更讓物聯網受到關注的原因是這些經由傳感器獲得的資料，可以經由網路將這些收集到的資訊傳送到計算中心，將這些資訊做為大數據(Big Data)的基礎。到目前為止已經有眾多的分析表明傳感器的數量正在飛速成長到了2020年將會達到約目前的兩倍左右。物聯網的架構由傳感器、閘道器、雲端計算中心構成，傳感器負責收集各種資訊，閘道器負責傳感器與雲端計算中心之間的連結，雲端計算中心負責資料的分析。當傳感器數量增加時，所需的閘道器數量也隨之增加，由於數量的上升，會造成出錯的次數也可能跟著上升，此時閘道器的穩定度成為新的需求。在本論文中，提出了基於分散式系統概念，具備容錯能力的協同閘道器系統。在接收傳感器資料的同時能夠就閘道器系統中發生的錯誤在一定的程度內予以修復。此外，考慮到要為雲端計算中心分擔計算上以及網路頻寬的負擔，我們設計了一個預處理資料的模組，這個模組可以和我們的閘道器系統進行整合，讓資料在送到雲端前先經過預處理。最後我們將會就可擴展性、容錯能力、計算能力三方面來對我們的閘道器系統進行評估。

With the developing of embedded devices and the widespread of communication network, the Internet of Things (IoT) is one of the most popular topics nowadays. The data produced by sensors can be analyzed then used by governments or businesses. The number of IoT sensors grows fast; in fact, there are many researches shows that the number of sensors will be twice of the number of sensors now. The architecture of IoT consists of sensor network, gateways and cloud. Sensors are responsible for sensing information around; gateways are in charge of bridging sensors and cloud, and cloud is the place where the sensor data be analyzed. When the number of sensors increasing, the need of gateways increases; also the frequency of failure might increase, too. Therefore, the stability of gateway is the new requirement. In this thesis, we propose a scalable IoT coordinated gateway system based on distributed system with fault tolerance. This can fix the faults happen to gateway system when sensor data is receiving. Besides, we implement a data preprocessing module in our gateway system, in order to ease the burden of network bandwidth and save computing power of cloud. Last, we evaluate our gateway system from three aspect: scalability, fault tolerance and computing power.