物聯網應用於命名資料網路之快取機制設計與分析

胡碩誠 ； 黃彥銓

《TANET2017 臺灣網際網路研討會》 2017年 (2017/10) Pp. 1374-1379

https://doi.org/10.6728/TANET.201710.0239

摘要

命名資料網路（Named Data Networking, NDN）是一個新興的網路技術，被看好成為未來網際網路的標準。不同於目前網際網路以主機為中心的設計，NDN強調直接以命名的資料取代IP位址來搜尋與傳輸資料，普遍視為適合實現物聯網應用的網路技術。物聯網的資料封包通常較小且具有時效性，與一般網際網路路由器快取中常見的多媒體資料並不相同。同時，相較於路由器，物聯網中的網路節點在能源供應，儲存與處理能力等方面是相當受限的。在本篇論文中，我們在基於NDN技術所建構的物聯網網路上，同時考量資料的即時性與受歡迎的程度，設計出一個分散型概率式快取策略。在我們的方法中，每個資料封包會攜帶有關新鮮度的參數，傳輸過程中網路節點會視情況更新該參數。過期的資料將不會在網路中傳輸以確保接收端確實收到正確且即時的資料。為了驗證提出方法的效能，我們利用ndnSIM來撰寫模擬系統，實驗結果證實了本方法能降低網路傳輸負載同時提高有效快取命中率。

關鍵字

命名資料網路 ； 物聯網 ； 快取機制

並列摘要

Name Data Networking (NDN) is emerging as one of the most promising future internet architecture. In contrast to the host-centric communication model of current Internet, NDN emphasizes content by making it directly addressable and routable, has been considered as an enabling technology for the Internet of Things (IoT). Unlike the typically cached content at routers, e.g. multimedia files, IoT data are transient and small in size. At the same time, IoT devices are typically resource-constrained compared with the Internet routers in terms of energy, storage and processing capability. In this paper, we consider NDN as a communication infrastructure for IoT and propose a distributed probabilistic caching strategy that relies on the freshness and popularity of data. We include the freshness parameter in each Data packet and will be updated when the packet is relayed by the IoT devices. No stale data will be transmitted over the network, the accuracy of data received by consumers can be guaranteed. The proposed mechanism has been evaluated through simulations with ndnSIM and results indicate that it can reduce network load, meanwhile increase valid cache hit ratio.

並列關鍵字

Named Data Networking ； Internet of Things ； cache policy