WSN環境下Diff-based更新之還原機制的研究與實作

Recovery Mechanism for Diff-based Reprogramming in WSNs

連新路(Shin-Lu Lien)

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(2013年)

https://doi.org/10.6837/NCNU.2013.00210

摘要

無線感測網路的遠端動態更新機制允許運作中的感測節點經由軟體機制更新當前正在執行的應用程式。動態更新機制能幫助系統管理者調整感測節點以適應當前的監控環境或變更當前的監控任務。目前已經有許多的研究使用diff-based策略進行動態更新，此策略在更新過程中僅傳送新、舊版本之間的不同部份。當感測節點以full image策略進行更新失敗，系統通常能夠藉由重新嘗試更新或載入保存在節點上的前一版本就能夠適當地進行自我回復。通常感測節點的儲存空間十分有限，因此舊版的patch file通常會被下一個新的patch file覆蓋。雖然此方式能節省儲存空間，但覆蓋策略會在diff-based更新機制失敗時引發更多額外的問題。因為進行多次diff-based更新之後，更新回復機制會因為中間的幾個patch file版本被新的版本覆蓋，被覆蓋的patch file會造成回覆機制無法藉由保存在節點上的歷史版本漸進的回覆到當前的最新版本。要解決此問題，就需要重新傳送被覆蓋的patch file，同時也需要從主機發送控制指令，但是這些重新傳輸的動作與控制指令也會同時影響到網路中所有正常運作的感測節點。尤其重新傳送patch file對於硬體資源有限的無線感測網路更是不可接受的消耗。為了解決此問題，我們提出了兩個新的回復機制分別稱為N與N-1漸進還原機制，此兩機制針對更新期間發生的失誤進行回復。我們將機制實做在TinyOS並且選擇Two-Stage Diff做為感測網路的遠端更新機制，因為此機制完全相容於TinyOS預設的Deluge遠端更新機制。實驗結果顯示，我們提出的N與N-1漸進還原機制與預設的Deluge還原機制比較能夠大幅減少回復所需要的傳輸量與回復時間。

關鍵字

無線感測網路 ； 作業系統 ； 嵌入式系統 ； 動態更新 ； 系統容錯

並列摘要

Reprogramming remote sensor nodes on wireless sensor networks (WSNs) allows the software on them to be dynamically updated or enhanced for adapting to changed environmental conditions or application requirements. Many researches use diff-based approaches to transmit only the code difference to sensor nodes for updating. When traditional full image replacement reprogramming mechanism fails, sensor nodes can simply perform self-recovery by retrying or reloading the stored full image of the previous version, but failure of diff-based reprogramming may cause extra problems. Because sensor node’s storage is limited, the older version of patch files may be overwritten by the next one. Then after many times of update, reprogramming recovery mechanism will lack the intermediate version of patch files for incremental recovery. To recover from this situation, extra data retransmission is needed and the host machine should issue recovery commands, but both of them will also affect all other normal nodes in WSN. The data retransmission is especially unacceptable in a resource-limited WSN environment. To solve these problems, we have proposed two new diff-based recovery mechanisms named N and N-1 Incremental Recovery to recover sensor nodes from reprogramming failures. We implement the proposed mechanism in TinyOS and choose Two-Stage Diff as our diff-based dynamic update mechanism for updating the software on sensor nodes. Two-Stage Diff mechanism is compatible to Deluge mechanism which is taken as TinyOS’s default remote full image update mechanism. Experiments with the modified TinyOS show that our N and N-1 Incremental Recovery compared with the Deluge default recovery mechanism can greatly reduce the amount of retransmitted data and the processing time can be significantly reduced as well.

並列關鍵字

Wireless Sensor Networks ； Operating System ； Embedded System ； Dynamic Update ； Fault Tolerance

參考文獻

[1]. R. A. Rashid, S. H. S. Arifin, M. R. A. Rahim, M. A. Sarijari, and N. H. Mahalin, “Home healthcare via wireless biomedical sensor network,” 2008 IEEE International RF and Microwave Conference (RFM), Conference, pp. 511-514, Dec. 2-4, 2008.

[2]. G. A. Sánchez-Azofeifa, C. Rankine, M. Marcos do Espirito Santo, R. Fatland, and M. Garcia, “Wireless Sensing Networks for Environmental Monitoring: Two case studies from tropical forests,” IEEE 7th Intel. Conf. on e-Science, pp. 70-76, 2011.

[3]. M. A. Hussain, P. Khan, and K. K. Sup, “WSN research activities for military application,” 11th IEEE Intel. Conf. on Advanced Communication Technology (ICACT), Vol. 1, pp. 271-274, Feb. 15-18, 2009.

[4]. I. Saleh, H. El-Sayed, and M. Eltoweissy, “A Fault Tolerance Management Framework for Wireless Sensor Networks,” Innovations in Information Technology, pp. 1-5, Nov. 2006.

[5]. F. Koushanfar, M. Otkonjak, and A. sangiovanni-Vincentelli, “Fault Tolerance in Wireless Ad-Hoc Sensor Networks,” In IEEE Sensors, pp. 1491-1496, June 2002.