無線感測器網路中動態資料中心儲存方法之研究

Abstract

我們探討一個距離基地很遠的無線感測器網路，其中之感測器節點可以偵測不同型態的資料。使用者在網路內部發出即時且集合性的資料詢問。例如：過去四小時內的平均溫度為多少？為了避免不必要的資料傳輸以造成能量消耗，以及加強資料安全性，節點收集到的資料將會被儲存在網路內部，而不是傳送回在遠方的基地。針對此問題，我們提出一個雙層、群集式、資料中心的儲存方法。我們的目標是達到有效率的資料擷取，以及延長網路生命。我們把網路虛擬地切割成群集，不同型態的資料將會被儲存到不同的群集。而在群集內部，資料將會以環狀複製法來儲存。我們提出兩種資料型態和群集的對應方法：雜湊法與查表法。

We study the problem of a wireless sensor network system which is deployed far from the base station and is for collecting heterogeneous types of data. Users make queries of real-time, collective data inside the network. For example, what is the average temperature in the last 2 hours? For avoiding unnecessary data transmissions and enhancing data security, the collected data are stored inside the sensor network itself instead of transmitting all the data back to the base station. We propose a two-tier, cluster-based, data-centric storage system for the studied scenario. Our goals are to achieve efficient storage and retrieval of collective data, and to extend system lifetime through load balancing. In our system, the network is virtually divided into clusters. Different types of sensed data will be stored in their corresponding clusters. Inside a cluster, data is replicated in a ring-based fashion. As for deciding which cluster the data will be forwarded to, we propose two methods: hashing and table mapping. Both methods provide load balancing to prolong network lifetime.