在資訊科技的發達下,智慧型手機及載具裝置越來越普及,再加上行動通訊技術也一直在提升,許多裝置都會透過訊息推播的方式來傳遞訊息;在智慧型手機作業系統中,android與ios佔絕大部分,若要透過訊息推播,需向android與ios申請開發者權限,讓在開發過程中變的很繁瑣,籍此本研究利用MQTT協定來取代android與ios的推播方式。MQTT協定必需要自行建置伺服器環境,本研究會採用雲端技術(IaaS)基礎設施即服務平台來建置環境,透過IaaS的虛擬伺服器技術,可以彈性調整伺服器的硬體資源,以及IaaS的管理介面和程式類別庫,可以在短時間內啟動另一台虛擬伺服器來發送訊息。 目前在一般訊息的應用,都沒有針對等級之分的訊息來處理,例如:血糖值檢測。在檢測過程中,正常血糖值會比低血糖及高血糖訊息來的多;在處理高血糖或低血糖訊息時,都必須在短時間內通知診所或是親友,若透過目前常用的佇列演算法來發送血糖訊息,可能會造成訊息的逾時發送,也無達到在用最少伺服器的情況下來發送訊息。在本研究中,會提出自動調配佇列演算法,能針對訊息等級的不同,來發送訊息,並且能在訊息的限制時間內發送;透過自動調配佇列演算法,動態調節佇列來平衡訊息流量,配合IaaS虛擬伺服器計價方式(以小時為單位),能在用最少伺服器的情況下,減少成本上的支出。 在本研究也會針對所提出的自動調配佇列演算法,進行訊息發送的壓力測試及訊息等級分佈比例測試,希望在測試結果中,達到佇列的自動調配及平衡來處理不同的訊息流量及不同的訊息等級分佈。
Smart phones and sensor devices are proliferating in today’s information technology, and many devices will use push notifications in their messaging. If a user wishes to push a message, Android and iOS, the general smart phone operating systems, require that users apply for a developer role. Since this development process is cumbersome, in this study, we established an MQTT protocol agreement in the IaaS cloud computing platform to replace the current Android & iOS send mode. The algorithms that are now processing queues were not developed to work at the message level, e.g., blood glucose monitoring. If a patient’s blood glucose level is normal, the message timing is reasonable. However, if the blood glucose level becomes emergency, a message must be sent immediately. If we use common queue algorithms to send messages, such as the RR and WRR algorithms, messages may arrive late or require more servers to be transmitted. In this paper, we propose a load-balancing algorithm for medical care applications that can handle messages at the message level as well as adjust and balance the queues and send messages with a minimum of servers, thereby reducing the cost of virtual servers on the IaaS.