技術探索

適用於物聯網的記憶體式資料處理技術

中文摘要

近年來物聯網(Internet of Things, IoT)興起,IDC市場調查發現2014年物聯網的市場規模已達到6,558億美元,並預計至2020年將擴大到1兆7,000億美元[1]。在物聯網中所有智能裝置都可以上網,因此所帶來的訊息總量非常大,但每個東西所傳送的訊息卻很小,因此需要新技術快速地處理或儲存這些大量的小訊息。本論文基於批次處理循序送出(BOSC)技術,結合記憶體式處理(in-memory processing)技術、快速封包處理技術(fast packet processing)與遠端機器儲存日誌(logging)的方式,提供快速接收並處理物聯網訊息的方法,並以硬碟效能最大化的方式批次循序更新至硬碟。

Abstract

In recent years, the IoT (Internet of Things) rises. IDC market survey showed that the IoT related market in 2014 reached 655.8 billion US dollars, and is expected to rise to 1,700 billion in 2020 [1]. In IoT, all smart things can connect to the Internet and bring about large amount messages – and the size of each message is very small. Therefore new techniques are needed to process and store the large amount of small messages. This paper, based on Batch mOdifications with Sequential Commit (BOSC) technique, combined with in-memory processing, fast packet processing, and remote logging techniques, provides a solution for fast receiving and handling messages, and then flushing messages with optimal hard disk performance.

關鍵詞(Key Words)

批次更新循序送出(Batch mOdifications with Sequential Commit;BOSC)
記憶體式處理(In-memory Processing)
快速封包處理技術(Fast Packet Processing)
物聯網(Internet of Things;IoT)

相關檔案: 適用於物聯網的記憶體式資料處理技術(全文)