改善WSN環境下動態更新的效率的研究與實作

Research on Improving the Efficiency of Dynamic Update in WSN Platform

呂宗霖(Tsung-Lin Lu)

指導教授 ： 姜美玲

國立暨南國際大學/管理學院/資訊管理學系/碩士(2011年)

https://doi.org/10.6837/NCNU.2011.00077

摘要

隨著無線傳輸與嵌入式系統相關技術的進步，無線感測網路(Wireless Sensor Network, WSN)已成為具有高度發展性的資訊技術，在許多領域已被廣泛的應用與研究。軟體動態更新(Dynamic Update)機制在WSN的相關研究中，是個非常重要且具有挑戰性的議題。感測節點藉由軟體動態的更新，可以改進與増加軟體功能，符合更多實際環境的變化與不同應用的需求。許多研究使用Diff-based的動態更新方式，在更新時不需要傳送完整的程式映像檔，只需要傳送新舊檔案差異的部分，是個較有效率的更新方式。然而，Diff-based的動態更新方式比在更新的過程中傳遞完整的映像檔的方式在處理上複雜許多，尤其感測節點的資源是很有限的，如何降低對感測節點資源的消耗，並且有效率的完成動態更新的動作，即是本論文主要研究與探討的議題。
在本論文中，針對有資源限制的WSN環境下，基於Diff-based的動態更新方式，我們提出了一個新的且更有效率的元件動態更新機制，稱為Two-Stage Diff，目的是減少傳送更新資料的大小與提昇更新時的效率，同時考量Flash Memory的抹除(Erase)與寫入的特性，並設計與優化Diff Script，使我們的機制更適合使用在資源有限的WSN環境。
我們將所提的Two-Stage Diff元件動態更新機制實作於TinyOS系統上，我們主要是修改了在TinyOS上支援的Deluge動態更新機制來實作我們的Diff-based的動態更新方式。實驗結果顯示，當對感測節點進行動態更新時，我們的Two-Stage Diff可以有效地減少所需傳送的更新資料，且只需要Deluge的0.2%-49%的封包傳輸量。同時，感測節點執行更新的時間也大幅地降低，如在小量更新的情況下，比起Deluge效能更可以提升22.64倍。

關鍵字

無線感測網路 ； 作業系統 ； 嵌入式系統 ； 動態更新

並列摘要

As the advance of wireless communication and embedded system technologies, wireless sensor networks (WSNs) which consist of lots of sensor nodes with various sensing technologies nowadays have versatile applications in lots of fields and become a highly developed nature of information technology. Supporting dynamic software update mechanism is a very important and challenging topic in the related researches of WSNs. If a WSN platform can support remote dynamic update, software of sensor nodes can be dynamically updated or enhanced to adapt to changed environmental conditions or different requirements of applications. Many researches on dynamic updates use the diff-based approaches that do not transmit the whole new program image to sensor nodes for updating. Only the code difference needs to be transmitted to sensor nodes. Compared with the approaches basing on full image replacement, diff-based approaches are more effective whereas more complicated in processing. However, sensor nodes are often of limited energy, memory, processing power, and communication bandwidth. How to efficiently provide dynamic update in WSNs while reducing resource consumption of sensor nodes is the main concern of this thesis.
In this thesis, for the resource limited WSN environments, we propose a new and more effective diff-based approach named Two-Stage Diff to dynamically update software components in sensor nodes. This mechanism aims at effectively reducing the transmission size of the update data and improving the efficiency of the update processing. Especially, flash memory characteristics in erasing and data writing are considered in our design. Besides, the design of Diff Script format is optimized to further decrease the size of transmitted update data. Such that, our mechanism can reduce the resource consumption incurred by dynamic update mechanism and be more suitable for operating in resource limited WSN environments.
We have implemented the proposed Two-Stage Diff mechanism in TinyOS for component update. We mainly modify Deluge dynamic update mechanism to support our diff-based component update mechanism. Experiments show that when performing updates, our Two-Stage Diff can effectively reduce the size of transmitted update data and require only 0.2%-49% of packet transmission as compared with Deluge. At the same time, the update processing time can be largely reduced as well. For example, Two-State Diff can perform 22.64 times better than Deluge in the case of small amount of updating.

並列關鍵字

Wireless Sensor Networks ； Operating Systems ； Embedded Systems ； Dynamic Update

參考文獻

[1] A. Arora et al, “A Line in the Sand: A Wireless Sensor Network for Target Detection, Classification and Tracking,” Military Communications Systems and Technologies, Volume 46, Issue 5, pp. 605-634, Computer Networks, 5, December 2004.

[2] K. Lorincz, D. Malan, T. Fulford-Jones, A. Nawoj, A. Clavel, V. Shnayder, M. Welsh, and S. Moulton, “Sensor Networks for Emergency Response: Challenges and Opportunities,” IEEE Pervasive Computing, v.3 n.4, pp.16-23, October 2004.

[3] J. D. Lundquist, D. R. CAyan, and M. D. Dettinger, “Meteorology and Hydrology in Yosemite National Park: A Sensor Network Application,” Information Processing in Sensor Networks, pp. 518-528. Springer-Verlag, 2003.

[4] T.Y. Huang, K.Y. Hou, H.Y. Yu, E. Chu, and C.T. King, “LA-TinyOS: A Locality-aware Operating System for Wireless Sensor Networks,” in Proc. SAC’07, Mar. 2007.

[5] J. Koshy and R. Pandey, “Remote incremental linking for energy-efficient reprogramming of Sensor networks,” In Proc. of the 2nd European Workshop on Wireless Sensor Networks, pp. 354-365, 2005.