(西北工业大学 计算机学院, 陕西 西安 710129)
Analytical Energy Model of Coding-Based Reprogramming Protocols in Wireless Sensor Networks
ZHANG Yan,ZHANG Yu,LI Jun-Wei,XIA Xian-Jin,LI Shi-Ning
(School of Computer Science and Technology, Northwestern Polytechnical University, Xi'an 710129, China)
Chart / table
Similar Articles
Article :Browse 1586   Download 1577
Received:May 02, 2013    Revised:August 22, 2013
> 中文摘要: 在无线传感器网络中,当需要进行软件更新、修复软件Bug 时,重编程协议能够将新程序镜像分发到多跳网络中的全部节点.近来,基于网络编码的重编程协议被用于解决在高损环境下的有效代码分发问题,但对这些协议性能的分析仍有待深入.提出一种基于时间和网络拓扑的能耗分析模型,在模型中综合考虑了页面流水和节点通信距离对协议性能的影响.该模型分析结果与基于网络编码的重编程协议Rateless Deluge 的仿真结果达到了较好的一致性(平均单节点能耗相对误差在0.23%左右),验证了能耗分析模型的有效性.分析结果揭示了基于网络编码重编程过程的网络能耗与网络密度、网络大小和分发镜像页面大小的关系:网络能耗随着网络密度的减小呈上升趋势(能耗与网络密度经拟合呈二次函数关系);网络能耗与网络大小基本呈线性正相关关系(对于n×n 的网格型网络,n从3~10,随着n 的增大,网络能耗的平均增长率为50%),并且随着网络的增大,单节点能耗有所增加(平均增长率为6.9%);页面大小增加时,网络能耗呈下降趋势(平均下降率为11.2%).
Abstract:As application requirements evolve, the software on sensor nodes needs to be updated, or patched to fix errors to all the nodes of the entire networks. Recently, coding-based reprogramming protocols are proposed to provide efficient code dissemination in environments with high packet loss rate. Methods in analyzing the performance of these protocols, however, need further study. This paper presents an analytical model for energy cost based on time of dissemination and network topology. In this model, page pipelining and communication range between nodes are taken into accounts. Results from extensive simulations of a representative coding-based reprogramming protocol called Rateless Deluge coincide with the performance predicted by this model (the average relative error of single-node energy consumption is about 0.23%), thus validating the approach. These analytical results show that network density, network size and the number of packets per page all have significant impact on energy cost of the whole network. Particularly, a decrease in density of the network harms the energy performance (the energy consumption and network density is synthesized to be quadratic function). The energy cost of reprogramming the network is linear with the size of network (for the N×N grid network, the average growth rate is 50% when n change from 3 to 10) while the energy consumption of single node increases (the average growth rate is 6.9%). The energy cost is inverse with the page size of code image (the average rate of reduce is 11.2%).
文章编号:     中图分类号:    文献标志码:
基金项目:国家科技重大专项基金(2012ZX03005007); 陕西省自然科学基金(2013JQ8041); 西北工业大学基础研究基金(JC20110268) 国家科技重大专项基金(2012ZX03005007); 陕西省自然科学基金(2013JQ8041); 西北工业大学基础研究基金(JC20110268)
Foundation items:
Reference text:


ZHANG Yan,ZHANG Yu,LI Jun-Wei,XIA Xian-Jin,LI Shi-Ning.Analytical Energy Model of Coding-Based Reprogramming Protocols in Wireless Sensor Networks.Journal of Software,2013,24(S1):125-133