CAJ | 학술논문

随着物联网应用的推广,作为底层核心构件的传感器网络所承载传输业务成激增趋势,使得窄带宽无线信道成为了制约物联网应用的首要因素。合理分流流量,实现负载均衡是提升网络承载能力的有效方法。本文将ECMP ( Equal-Cost Mul-tipaths)技术与传感器网络自组织特性相融合,传感器网络多跳自组织特性为业务传输提供多条等价最短路径,ECMP将业务均衡分担到这些等价最短路径上,实现负载均衡。理论证明传统ECMP 配置方法全网节点开通ECMP 功能不仅会增加网络控制信息开销,而且在某些情况下反而会增大区域负载,形成网络拥塞。因此,本文提出动态选择开通 ECMP 算法( DC-ECMP )。算法以流入节点流量等于流出节点流量作为业务守恒约束,链路带宽上限作为链路容量约束,以最大链路利用率最小化为目标函数,建立多约束优化模型。并依据最大链路使用率和节点度判定开通ECMP 优先级,动态选择需开通节点,以获取最优网络传输性能。仿真结果表明DC-ECMP算法比已有PPV算法有效降低最大链路使用率,消除网络局部拥塞隐患,并且最大减少传输延时9.9 ms,节省网络资源消耗4.06%。
수착물련망응용적추엄,작위저층핵심구건적전감기망락소승재전수업무성격증추세,사득착대관무선신도성위료제약물련망응용적수요인소。합리분류류량,실현부재균형시제승망락승재능력적유효방법。본문장ECMP ( Equal-Cost Mul-tipaths)기술여전감기망락자조직특성상융합,전감기망락다도자조직특성위업무전수제공다조등개최단로경,ECMP장업무균형분담도저사등개최단로경상,실현부재균형。이론증명전통ECMP 배치방법전망절점개통ECMP 공능불부회증가망락공제신식개소,이차재모사정황하반이회증대구역부재,형성망락옹새。인차,본문제출동태선택개통 ECMP 산법( DC-ECMP )。산법이류입절점류량등우류출절점류량작위업무수항약속,련로대관상한작위련로용량약속,이최대련로이용솔최소화위목표함수,건립다약속우화모형。병의거최대련로사용솔화절점도판정개통ECMP 우선급,동태선택수개통절점,이획취최우망락전수성능。방진결과표명DC-ECMP산법비이유PPV산법유효강저최대련로사용솔,소제망락국부옹새은환,병차최대감소전수연시9.9 ms,절성망락자원소모4.06%。
With the development of Internet of Things(IoTs),wireless sensor networks,as the infrastructure of IoTs, should bear more and more various transmission services. The narrow wireless bandwidth becomes the first restrictive factors. Shunting flow to achieve load balance is the effective way to improve the bearing capacity of communication network. This paper integrated Equal-Cost Multi-Paths(ECMP)technique and WSN’s self-organized characteristics,in which the self-organized connection and multi-hops transmission model of WSN provide more than one shortest paths from traffic source to the destination sensor node,and then ECMP can equally apportioned traffic on these available e-qual cost multi-paths to stabilize huge traffic. It proved that the traditional ECMP model,configured all of nodes with ECMP function,will increase the overhead expenses,and even make heavier traffic load in some special cases. To solve these problems,this paper proposed the dynamic configuration ECMP algorithm-DC-ECMP. In the algorithm,it defined the traffic volume conservation constraint and the wireless transmitting bandwidth upper bound constraint, made the minimizing the maximum link utilization rate as objective function,and establish the multi-constraint optimi-zation model. Using the maximum link utilization rate and node’s degree to calculate the priority of configuration EC-MP function,each node can be dynamic configuration and achieve the optimal network’s transmitting performance. In our evaluation with simulations,the performance of DC-ECMP is measured and compared with PPV algorithms. Based on the experimental results,DC-ECMP outperforms existing algorithms in reduce the maximum link utilization rate, short the transmitting latency 9.9 ms,and save the networks’ resource consumption 4.06%.