2010_IEEE_NB-TRACE_MC-TRACE { Title{ //Multicasting vs Broadcasting: What are the Trade-offs? 多播与广播的权衡利弊 } Abstract{ //Network-wide broadcasting and multicasting //are two importmant routing schemes used in group communications. 组播通信中网络广播和多播是两个重要的路由模式 //In network-wide broadcasting, //generated packets at the source node are distributed to all nodes in the network, //while multicasting delivers the packets only to a subset of the nodes //by creating and maintaining a data dissemination structure. 广播中,源节点生成数据并发布给网络中所有节点, 多播只创建一个包并分发给网络中的部分节点 //Due to the overhead in multicasting,in certain situations, //it is more efficient to use network-wide broadcasting instead of multicasting, //even when the data is destined to a subset of the nodes. 在某些情况下,由于多播的开销,使用广播反而比多播更有效 即使只有一部分节点接收数据 //In this paper we analyze a specific broadcasting protocol //and a specific multicasting protocol for two performance metrics, //namely spectrum efficiency and energy efficiency, //and determine the conditions that make one of the preferable over the other. 本文具体分析了两种协议的两个指标,分别是有效率和能耗 依次确定在各种协议的适用条件 } 1.Introduction{ 1.{ //Group communications is essential for many applications in mobile ad hoc networks, //including supporting electronic classrooms,tactical military communication, //and communication in disaster recovery missions. 组播通信对于许多移动自组织多跳网络都非常重要 包括支持电子教室,军事战术通信,灾区救援任务通信 //One to many group communications are generally classified into two groups: //network-wide broadcasting and multicasting. 一对多通信通常分为两类:广播和多播 //In network-wide broadcasting the objective is to //distribute the generated data to all the nodes in the network. 广播的目标是将生成的数据分发给网络中的所有节点 //On the other hand, //the objective of multicasting //is to deliver the data to a subset of the nodes in the network. 而多播的目标是将数据传输给网络中的部分节点 //By using a data dissemination structure, //multicasting protocols limit //the diffusion of the data to a certain subset of the entire network, //namely to the multicast members. 多播将数据的传播限制在整个网络的某个子集中,即多播组的成员节点 //[1] and [2] present resent surveys on various group communication protocols. 文献1和文献2展示了各种组播通信协议的最新研究 } 2.{ //There is additional overhead incurred in multicasting protocols //compared to broadcasting protocols. 相比于广播,多播协议会有额外的开销 //In certain scenarios, //the cost of collecting and processing the additional information overwhelms the gains //in limiting the data dissemination structure to the multicast members. 在某些情景下,对于多播组成员,收集并处理额外信息的代价大于限制数据传播所获得的收益 //As one might expect, //in scenarios where the majority of the nodes are part of the multicast group, //one can increase the efficiency by using a broadcasting protocol //instead of using a multicasting protocol. 当网络中大多数节点都是组播成员时,使用广播可以提高效率 //In this paper, our objective is //to investigate the trade-offs between multicasting and broadcasting //in order to determine the conditions //that make one of them preferable over the other. 本文的目标是权衡多播与广播,确定其各自表现优异的条件 } 3.{ //There are similar that point out the trade-offs between multicasting and broadcasting. 针对多播与组播的权衡也有许多研究。 //Researchers in [3] compare a selected multicast protocol, //namely on demand multicast routing protocol(ODMRP), //with a selected network-wide broadcast protocol, //namely scalable broadcast algorithm (SBA). 文献3对比了多播协议ODMRP和广播协议SBA //The authors report that while multicasting is preferable for small group sizes, //as the group size increases, broadcasting becomes more efficient. 作者表示对于较小的分组,多播是有效的。 随着分组大小的增加,广播的效果好于多播 //However, the protocols considered in this work are quite different in nature. 但是在自然情况下工作状态会不同。 //While the aim of SBA is //to minimize the number of redundant routes are created intentionally. SBA算法的目的是最小化冗余路径的数量 //Thus the comparison between those protocols //does not provide a full understanding of the trade-offs //between multicasting and broadcasting. 这些协议之间的比较并不能为广播和多播的权衡提供全面的认识。 } 4.{ //Towards the goal of investigating the trade-off between multicasting and broadcasting, //we perform extensive simulation studies on a chosen protocol from each class: 为了研究多播与广播的平衡,从每类中选择协议仿真分析: //Network-wide broadcasting //through time reservation using adaptive control //for energy efficiency(NB-TRACE)[4] for broadcasting; 文献4中NB-TRACE协议是采用自适应控制的能量有效的广播协议 //and Multicasting //through time reservation using adaptive control //for energy efficiency (MC-TRACE)[5]for multicasting. 文献5中MC-TRACE协议是采用自适应控制的能量有效的多播协议。 //The first reason for choosing these protocols //is that they have been shown to outperform many other protocols in their class. 选择这些协议的首要原因是因为它们的表现优异。 //Moreover, those protocols are built on top of the same MAC structure, //and their sensitivity to MAC layer issues such as mobility and link errors are similar. 另外,这些协议建立在相同的MAC层结构上, 对于MAC层移动性和链路错误问题的敏感度是相同的。 //Finally, the data maintained by the protocols are very similar to each other, //and any additional burden of multicasting can directly be observed. 最后,协议间所维持的数据相似, 多播产生的任何额外负担都可以被直接观察到。 //Consequently, the protocols //can be combined into a unique framework and coexist simultaneously. 可以将协议整合到框架当中并同时共存。 //Ultimately, this approach potentially yields a unified protocol //where the better approach (broadcasting or multicasting) //can be used depending on the situation. 使用哪种方法更好取决于具体情况。 } 5.{ //The rest of the paper is organized as follows. 文章剩余部分结构如下。 //In Section II, the chosen candidate protocols are introduced. 第2节介绍选择的候选协议。 //Then, we analyze the efficiency of the protocols for a sample scenario in Section III. 第3节通过一个简单的场景对协议的效率进行分析。 //The effect of node density on the relative efficiency of the protocols //is investigated in Section IV. 第4节研究了节点密度对协议效率的影响。 //Finally, we conclude the paper in Section V. 第5节总结全文 } } 2.Description { //Description of the Chosen Multicasting and Broadcasting Protocols 所选择的多播和广播协议的描述 1.{ //The main purpose of this section //is to give the reader insight into the differences between the protocols. 本节的主要目的是使读者理解协议之间的不同点。 //The details of the protocols, NB-TRACE and MC-TRACE, //can be found in references [4] and [5], respectively. NB-TRACE协议与MC-TRACE协议的详细描述可以分别在文献4和文献5中找到。 } 2.{ //Both MC-TRACE and NB-TRACE are cross layer approaches //where the MAC layer and the routing layer functionalities //are implemented together in a unique framework. MC-TRACE协议和NB-TRACE协议都是交叉层方法 MAC层和路由层的方法在一个框架下统一实现。 //The MAC scheme of the protocols follows from MH-TRACE[6], //where the network is organized into overlapping clusters, //each managed by a clusterhead(CH). MAC层的模式遵从文献6, 网络被划分成相互重叠的簇 每个簇由一个簇首管理。 //Each CH holds the channel access rights in its neighborhood //and grants channel access to the neighbor nodes on demand. 每个簇首掌握着其邻域的访问控制权 并按照需求为邻节点分配信道 //Clusters and CHs are chosen dynamically //to ensure the existence of at least one CH in the neighborhood of each node. 簇和簇首是动态选择的 确保在每个节点的邻域内至少有一个簇首 //In other words, CHs form a dominating set. 也就是说,簇首组成一个支配集。 //Fig. 1 depicts the clustering structure of MH-TRACE. 图1描述了MH-TRACE协议中的成簇结构。 //Each CH choose one of the frames and operates in that frame. 每个簇首选择一个帧并在该帧内进行操作。 //Beacon and clusterhead announcement(CA) slots //are used to creat and maintain the clustering structure. 信标和簇首广播时隙用于创建和维护分簇结构。 //Nodes pass their channel access request to a nearby CH using the contention slots. 节点使用竞争时隙向邻近簇首申请信道。 //After receiving the channel access requests, //the CH sends a header message //that includes the transmission schedule //that will be followed for the rest of the frame. 接收到信道访问请求后,簇首发送一个包含传输安排的簇首消息。 //Nodes that acquire channel access from the CH //transmit an information summarization(IS) packet //that contains the unique packet ID //that will be transmitted together //with the routing information prior to the data payload //that is transmitted at the corresponding data slot. 从簇首获得信道访问权的节点发送一个包含唯一ID的信息摘要包 在相应的数据时隙,路由信息和数据有效载荷信息一起传送 //By listening to the relatively shorter IS packets, //nodes become aware of the data //that are going to be sent and may choose to sleep during the corresponding data slots. 通过监听相当短的信息摘要包,节点获知数据正在传送并可能在相应的数据时隙选择睡眠。 //Acknowledgements are embedded into the IS messages. 确认信息嵌入在信息摘要包当中。 } A.NB-TRACE{ //Routing in NB-TRACE makes use of the clustering structure. NB-TRACE协议中的路由利用成簇结构。 //The protocol sends a copy of each data packet to all of the CHs, //and the CHs retransmit these packets to their cluster members. 将数据的拷贝发送给所有的簇首,簇首再将数据包重新发送给他们的簇成员 //Each data session starts with an initial flooding stage //where each rebroadcasting node implicitly acknowledges its upstream node //through IS packets as a part of its transmission. 每个数据会话起始于一个初始泛洪阶段, 每个广播节点通过IS信息摘要包隐式地向其上游节点发送确认信息。 //In the case of the existence of more than one uptream node, //only one of them is selected and announced in the IS packet. 当存在不只一个上游节点时,只能选择一个并在IS信息摘要包中宣布。 //A node drops its relaying status and stops retransmitting the packets //when it does not receive an acknowledgement for a certain amount of time. 当节点在某段时间内未收到确认信息,则取消中继状态并停止发送数据包。 //Only the CHs keep retransmitting the packets even //when they do not receive any downstream acknowledgement. 只有簇首持续发送数据即使他们没有接收到任何下游确认信息。 //This behavior prunes the redundant retransmissions //and creats a tree that starts from the source node and ends at the CHs. 该方法减少冗余重传并创建一棵从源节点开始到簇首节点的树。 //The dynamic behavior of the network is handled by a local branch repair mechanism. 网络的动态性能通过本地分枝恢复机制处理。 } B.MC-TRACE{ 1.{ //MC-TRACE implements multicast routing on top of MH-TRACE using a mixed layer approach. MC-TRACE协议在MH-TRACE协议的基础上采用混合层的方法实现多播路由。 //Like NB-TRACE, MC-TRACE also starts with an initial flooding stage. 与NB-TRACE协议相似,MC-TRACE协议也从初始泛洪阶段开始。 //Nodes that do not receive a downstream acknowledgement stop retransmitting. 没有接收到下游确认信息的节点停止重发数据包。 //However, in MC-TRACE, CHs do not take a special role in routing. 然而,在MC-TRACE协议中,簇首在路由中没有特殊的作用。 //Instead, the member nodes keep sending an acknowledgement to their upstream node //even when they do not receive any downstream acknowledgements. 相反,成员节点需要向它的上游节点持续发送确认信息 即使他们没有接收到任何下游确认信息。 //Therefore, the tree is kept alive directly by the group members. 因此,组成员使树保持活跃。 } 2.{ //Furthermore, in MC-TRACE, //retransmitting nodes also choose and announce a downstream node //in addition to their upstream node. 另外,在MC-TRACE协议中,重发节点需要选择并告知一个除他们上游节点以外的下游节点。 //The first node that sends an upstream acknowledgement //is selected as the downstream node //and announced in the following transmissions. 第一个发送上游确认信息的节点被选为下游节点并在接下来的传输中被告知。 //The node that is announced as the downstream node //is responsible for sending upstream acknowledgements and keeping the branch alive. 被告知作为下游的节点有责任发送上游确认信息并保持分枝活跃。 //With the help of this mechanism, //in the case of more than one leaf member node receiving the data from the same branch, //only one of them sends the acknowledgement messages. 在该机制的帮助下,不只一个叶子成员节点从相同分枝接收到数据 只有其中的一个发送确认信息。 //Although this mechanism eliminates redundant acknowledgements, //the need for acknowledgements //makes MC-TRACE consume considerably more resources compared to NB-TRACE //when the cluster members are spread throughout the region. 尽管该机制减少了冗余的确认信息,相比于NB-TRACE协议, 对确认信息的需求使MC-TRACE协议消耗了更多的资源 当簇成员在区域内传播时 } } } 3.Comparing Multicast and Broadcast{ 1.{ //In general, multicasting protocols eliminate redundant retransmissions //by confining the data dissemination to a limited area. 通常,多播协议减少冗余重传,通过将数据传播限制在一个区域内。 //However, this comes with the additional cost of overhead //to keep the data distribution structure alive. 然而,保持数据分发结构需要花费额外的开销。 //Intuitively, while multicasting //is expected to be a more efficient method of data distribution for small group sizes, //broadcasting would be more efficient for large group sizes. 直观上,多播对于小组数据分发更有效,而广播对于大组更有效。 //In this section, we show that //this is indeed the case by analyzing broadcasting and multicasting //through extensive simulations of a select broadcasting protocol,NB-TRACE, //and a select multicasting protocol, MC-TRACE. 通过对选择的广播和多播协议进行大规模的仿真进行详细的分析。 } 2.{ //In particular, the number of multicast group members beyond which NB-TRACE //becomes more efficient for data dissemination, //called the cross-over point, //is determined for various scenarios. 广播更有效时的组播成员节点个数叫做交叉点。 //By comparing the simulation results, //we can analyze the effect of the total number of nodes in the network //and the size of the region in which the nodes are distributed //on the value of this cross-over point. 交叉点处网络中节点的个数和分簇大小。 } 3.{ //Two performance metrics, energy efficiency and spectrum efficiency, are considered. 考虑两个度量标准,能效和范围效率。 //Specifically, the average energy spent per node per generated packet //and the total number of transmissions per generated packet are measured for each scenario //to compare the energy efficiency and the spectrum efficiency of the protocols,respectively. 用节点生产数据包的平均能耗来比较能效, 用数据包的发送次数比较范围效率。 } 4.{ //We begin with describing the simulation environment //and the parameters selected for the scenarios under concern. 首先描述仿真环境,和参数选择。 //Then, for a network of 100 nodes distributed in a 1000×1000m area, //the bandwidth efficiency of the protocols and their energy consumptions are compared. 在1000平方米的区域内分布100个节点,比较协议的带宽利用率和能效。 //Finally, the analysis is extended by varying the number of nodes in the network //and the size of the area in which nodes are distributed. 进一步分析网络节点个数和区域大小对网络的影响。 } A.Simulation Environment{ 1.{ //We conduct ns-2 simulations of NB-TRACE and MC-TRACE under different network scenarios. 使用ns-2仿真环境。 //We used default energy and propagation(two-ray ground) models in ns-2. 采用默认的能量传播模型。 //Both path loss and interference //are taken into account in determining a successful reception. 路径损坏和干扰都纳入考虑当中。 //The receiver can receive only those packets //whose received power is above a certain threshold. 接收端只能接收功率大于某一阈值的数据包。 //For a successful reception, //the receiver has to be within 250m of the transmitter //with the given propagation model, //reception threshold and transmission power. 接收端必须在给定模型250米的范围内,接收阈值和发送功率。 //However, simultaneous transmissions interfere with each other //and prevent successful reception. 节点之间的同时发送会对成功接收造成干扰。 //In the case of simultaneous transmissions, //a successful reception is only possible if, at the receiver side, //the power of one of the packets is 10 times larger than any other packet. 同时发送数据时,只有在接收端一侧,某数据包的功率比任何其他数据包大10倍时才能成功接收。 //The transceivers are fixed at 2 Mbits/sec data rate. 发送端的发送速度固定为每秒2M比特。 } 2.{ //The packet generation model assumes a 16-bit voice coder //that generates 100 byte packets every 25ms. 数据包生成模型假设为一个声音编码器,每25毫秒生成100字节数据包。 //The length of an IS slot is 12 bytes long in NB-TRACE, //while it is 15 bytes long in MC-TRACE //due to the extra routing information requirements. 在NB-TRACE协议中IS信息摘要的长度为12字节, 而在MC-TRACE协议中由于需要额外的路由信息,IS信息摘要包的长度为15字节。 //The superframe period is fixed to the packet generation period, //and the number of frames per superframe is fixed at 6 //for both NB-TRACE and MC-TRACE. 超帧周期固定为包的生产周期,每个超帧的帧数量固定为6 //However, due to the extra bits in the IS slots, //MC-TRACE has 6 data slots per frame whereas NB-TRACE has 7. 然而,由于IS信息摘要包中额外的数据位,多播协议每帧有6个数据时隙,而广播有7个 //Each node transmits or relays the data packets of the stream //using one of the available data slots. 节点使用其中的一个数据时隙发送或转发数据包。 } 3.{ //In order to have a fair comparison in terms of energy consumption, //the concept of group members is introduced to NB-TRACE. 为了公平的比较能耗,对广播协议引入分组成员的概念。 //Nodes that do not belong to the group do not listen to the data slots of the stream. 组外的节点不监听组内的数据流时隙。 } 4.{ //The power spent by each node varies according to the operation performed by the node. 节点能量消耗根据节点的具体操作来确定。 //During successful reception, collision and carrier sensing periods, //the node consumes power at the rate of the reception power level. 在接收数据,碰撞和传输阶段的节点能量速率为接收能耗水平。 //There is also an idle state //where only the power needed to run the circuitry //is dissipated without any actual packet receptions. 当没有任何接收任务时则处于空闲状态。 //The nodes are assumed to turn off any circuitry //when they go into the sleep state, //where the power consumption is minimal. 节点将关闭电路并进入睡眠状态使能耗为最小。 } 5.{ //All the nodes except the source node are initially distributed //according to a uniform random distribution, //and during the course of the simulation //the nodes move following the random way-point mobility model[7][8] //with nodes speed chosen from a uniform random distribution //between 0.0m/s and 5.0m/s with zero pause time. 节点随机分布,采用随机路点移动模型,如文献7和文献8,移动速度0至5米每秒,无停顿 //The source node starts from the center of the region //and follows the same random way-point mobility model. 源节点从区域中心开始按照路点移动模型随机移动。 } 6.{ //100 repetitions are performed for each scenario, //and the presented results show the averages //and the standard deviations of the results. 每种场景重复实验100次,取结果的平均标准差。 } } B.Bandwidth Efficiency{ 1.{ //In the TRACE frame structure, //there are fixed number of data slots per superframe. 在TRACE的帧结果中,超帧中的数据时隙个数是固定的。 //The data slots are used to transmit the payload of //both the generated data and the relayed data. 数据时隙用于发送和转发生成的数据。 //Any redundant use of the data slots //wastes the available network resources //and may prevent another stream's information //from being disseminated over a region. 数据时隙的冗余利用将浪费网络可用资源,阻止其他数据信息流的区域传播。 //Thus, efficient use of the data slots //is one of the goals of both NB-TRACE and MC-TRACE. 数据时隙的有效利用是广播和多播协议的目标之一。 } 2.{ //The number of data packet transmissions per generated packet //is depicted in Fig.2 for both NB-TRACE and MC-TRACE. 图2显示了广播协议和多播协议的数据包的传输次数。 //The change in the number of multicast members //does not effect the number of data transmissions in NB-TRACE significantly //since data is broadcasted to the entire network //regardless of the locations and number of multicast members. 多播组成员个数的变化并不影响广播协议中数据的传输次数 因为数据被广播至整个网络而不是本地或组播成员。 //The slight variation in NB-TRACE //is due to the additional restriction we impose on the algorithm //to make a fair comparison between NB-TRACE and MC-TRACE. 广播协议中的轻微变化是由于算法中额外的限制,为了公平比较广播和多播协议。 //We prevent non-group members from listening to the ongoing stream //unless they are relaying the data to save energy. 除非节点转发数据,否则限制非组成员监听数据流以节省能量。 //However, this reduces the efficiency of the recovery of link failures //through local branch repair mechanism, //since the neighbor nodes that are not in the multicast tree //do not have any data in their transmission queue. 由于邻节点不在多播树中,它们的传输队列内没有任何数据, 通过本地分枝恢复机制修复链路失效的效率被降低。 } 3.{ //Unlike NB-TRACE, in MC-TRACE, //the number of transmissions increases as the number of group members increase. 与广播不同,多播中,传输数量随多播组成员的个数增加而增加。 //Nodes are distributed with an independent identical random uniform distribution. 节点随机分布。 //Hence, as the number of group members increases, //the size of the multicasting tree increases. 随着多播组成员个数的增加,多播树的大小也在增长。 //Hence, the number of transmissions per generated packet increases. 数据包的传输次数也增加了。 } 4.{ //It can be observed from Fig.2 //that for small group sizes the data dissemination //is more efficient in MC-TRACE //compared to NB-TRACE, and vice versa. 图2显示,对于小的分组,多播的数据分发比广播更有效。反之亦然。 //NB-TRACE requires 4 times more transmissions compared to MC-TRACE //when there is only one group member. 当组成员只有一个时,广播的传输量是多播的4倍。 //At the other extreme, //when all the nodes are in the multicast group, //the number of data transmissions required using NB-TRACE //is half of the transmissions required when using MC-TRACE. 另外,当所有节点都是多播组成员时,广播的通信量是多播的一半。 //The cross-over point of MC-TRACE and NB-TRACE //occurs at a multicast group size of 24 nodes, //above which NB-TRACE provides more efficient data dissemination and vice versa. 两个协议的交叉点出现在多播组成员个数为24个的时候, 多播组成员个数多于24个节点时,广播的效果更好,反之亦然。 } } C.Energy Efficiency{ 1.{ //In Section III-B, //we compared the number of data transmissions for NB-TRACE and MC-TRACE. 上节比较了广播与多播的数据量。 //The number of transmissions and receptions of other packet types, //namely beacon, CA, contention, and header are expected to be comparable. 其他类型包的接收和发送数量,如信标,簇首选举,竞争,都需要比较。 //Thus, a similar trade-off that exists in Section III-B //is also expected for the energy consumption metric. 在能耗上也存在一个相似的权衡。 } 2.{ //On the other hand, since each IS slot //is matched with its corresponding data slot, //the difference in the number of data transmissions between protocols //is also expected to be observed in the number of IS transmissions. 既然IS信息摘要包与其相应的数据时隙相对应, 两协议的数据通信量上的差异也会体现在IS的通信量上。 //Furthermore, since the length of IS slots in MC-TRACE //are longer compared to the ones in NB-TRACE, //the energy consumption of MC-TRACE is excepted to be even higher. 多播协议中的IS包比广播协议长,多播协议的能耗也更高。 //This behavior is expected to shift the energy consumption curve of MC-TRACE //in such a way that the cross-over point will //occur with fewer multicast group members. 多播组成员个数较少时能耗会增加。 } 3.{ //The energy consumption per node per generated packet is depicted in Fig.3 //as the number of group members is varied from 1 to 99. 如图3所示,随着多播组成员节点个数的增加,数据包消耗的能量也逐渐增加。 //The energy consumption increases with an increase //in the number of multicast members for both protocols. 随着多播组成员个数的增加,两个协议的能耗都有所增加。 //Although with an increasing number of multicast group members //there was a slight decrease in the number of transmitted data messages for NB-TRACE, //shown in Fig.2, the energy consumption increases. 如图2,多播组成员增加,广播的数据通信量下降,能耗增加。 //This is due to the fact that the energy consumption in the reception //and transmission states of a node are of the same order. 实际上节点接收数据的能耗与发送状态下的能耗相当。 //As the number of multicast group members increases, //the number of receptions increase, //which in turn increase the energy consumption. 多播组成员增加,接收数据增加,能耗增加。 } 4.{ //It can be observed from Fig.3 that for small multicast group sizes //MC-TRACE is more energy efficient while for large group sizes //NB-TRACE performs a more energy efficient operation. 图3表明当多播组成员很少时,多播更有效,成员多时,广播更有效。 //NB-TRACE consumes 13% more energy compared to MC-TRACE //when there is only one group member. 当组成员只有一个时,广播比多播多消耗13%的能量。 //On the other hand, when all 99 nodes are in the multicast group, //the energy consumption of NB-TRACE //is 21% lower than the energy consumption of MC-TRACE. 当组成员有99个时,广播比多播少消耗21%的能量。 } 5.{ //The energy consumption of MC-TRACE increases faster than the increase in NB-TRACE //and goes above the energy consumption of NB-TRACE //at the cross-over point of 11 multicast group members. 多播协议的能耗比广播增长的快,在多播组成员个数为11个时,多播能耗超过了广播。 //The cross-over point in Fig.3 //is lower than the cross-over point observed in Fig.2 图3中的交点值比图2小。 //This is excepted since the length of the IS slots in MC-TRACE //is higher than the length of the IS slots in NB-TRACE, //which increases the energy consumption of MC-TRACE compared to NB-TRACE. 由于多播协议中IS包的长度比广播协议长,所有能耗多。 } } } IV.Effect of Node Density{ 1.{ //In Section III we observed that, depending on the number of multicast group members, //NB-TRACE and MC-TRACE can be advantageous over the other one in terms of bandwidth //and energy efficiency. 由于多播组成员个数的不同,广播和多播在带宽和能效上各有所长。 //We have identified the cross-over points for a network //consisting of 100 nodes distributed on a 1000m×1000m area. 在1000平方米的区域内分布100个节点,验证网络协议的交点。 //In this section, we extend our analysis to other scenarios //to see the effects of node density on the efficiency of the protocols. 分析节点密度对协议效率的影响。 } 2.{ //In considering the density, //both the number of nodes and size of the area in which the nodes //are distributed are of concern. 对于密度,分别考虑节点的个数和区域大小。 //The nodes close to the edge of the large area //have fewer neighbors and dissemination close to the edges //follows a pattern different than the data dissemination pattern of the nodes //residing on the middle of the region. 区域边缘的节点邻节点个数比区域中心的节点少,数据传播模式也不同。 //In order to investigate the edge effects accurately, //variation in the number of nodes and the size of the area are considered separately. 为了准确衡量边缘影响,分别考虑节点数量变化和区域大小变化的影响。 } 3.{ //Specifically, we investigate the cross-over points //on the number of multicast group members above //which NB-TRACE is more efficient compared to MC-TRACE //and vice versa in terms of the number of data packet transmissions //and energy consumption for a range of total number of nodes and area size. 多播组成员个数影响多播与广播的数据传输效率,在总节点个数和区域变化范围内。 //The observed cross-over points are presented in Table I. 表1列出了不同衡量标准下多播与广播的交叉点。 } 4.{ //The first observation from Table I is that for a given number of nodes, //the cross-over point for the number of data transmissions //increases with an increase in the size of the area. 节点数量一定时,区域面积越大,传输量越大。 //For larger areas, the separation between //the source and the multicast members is largeer. 区域较大时,源节点与多播组成员间隔较大。 //This translates into a larger number of hops between the source and a destination, //and hence a higher number of transmissions in both NB-TRACE and MC-TRACE. 区域增大会使源节点到目的节点的跳数增加,无论是多播还是广播传输量都会增大。 //However, the increase in the number of transmissions in NB-TRACE //is larger than it is in MC-TRACE //since the number of data transmissions in NB-TRACE //is effected not only by an increase in the expected distance //between the source and the multicast members //but also by an increase in the expected separation //between the source and non-multicast members. 区域增大后,广播消耗的能量比多播增长的快。 不仅是由于源到多播成员的距离增加,也包括源到非组播成员的传播 //As a result, the value of the cross-over point //for the number of data transmissions metric increases //with an increasing size of the network. 总之,随着网络规模的增大,数据传输量的交叉点的值也在增长。 } 5.{ //Table I also shows that, for a fixed network size, //the network with more nodes has a higher cross-over point. 网络规模固定时,节点越多,交叉点的值越大。 //Since the node locations are independent, //the separation between the source and the multicast members //is independent of the number of nodes in the network. 节点位置不固定,源节点与多播组成员之间的间隔与网络中节点的个数无关。 //Hence, the number of transmissions in MC-TRACE //does not deviate significantly as the total number of nodes increases. 多播中的传输量并没有随着节点总数的增加有明显的偏离。 //On the other hand, as the total number of nodes in the network increase, //they are excepted to cover a larger region on the simulation area, //and the number of CHs are excepted to increase. 随着网络中节点总数的增加,网络覆盖区域也增加,簇首个数也增加了。 //As a result of this, the number of transmissions in NB-TRACE increases up to a limit //where the entire area is covered as the total number of nodes increases. 随着网络中节点总数的增加,整个区域都被覆盖,广播的传输量的增长达到极限。 //Thus, the cross-over point also increases for an increasing number of nodes. 节点数量增加,交点值也增加。 } 6.{ //The energy consumption of a node consists of a variable part //that is incurred for data packet transmissions and receptions //and a relatively constant part for the control messages. 节点能耗由传输和发送数据包的变量部分和控制信息的常量部分组成。 //For a fixed network size and a fixed number of multicast members, //the number of nodes that part in the data dissemination tree created in MC-TRACE //is independent of the total number of nodes in the network. 网络规模和多播组成员个数固定时,多播中所创建的数据传播树中的节点个数 不依赖于网络中节点的总数。 //Hence the variable energy consumption for data dissemination //is also independent of the number of nodes. 数据分发产生的能耗变化也不依赖于节点个数。 //When the number of nodes in the network increases, //the energy consumption per node in MC-TRACE decrease. 当网络中节点个数增加时,多播协议时每个节点的能耗下降。 //On the other hand, for a fixed network size and a fixed multicast group size, //the energy consumption in NB-TRACE is independent of the number of nodes in the network //since the data dissemination tree covers the entire network. 广播协议中数据分发树覆盖整个网络,网络规模和分组大小固定时,协议能耗不依赖于所有节点个数。 //As a result, the cross-over point //increases with increasing number of nodes in the network, //as can be observed from Table I. 网络节点个数增加,交叉点的值也增大。 } 7.{ //It is interesting to note that, when considering energy efficiency, //the cross-over point is approximately constant //for a fixed number of nodes as the network area increases. 随着网络区域的扩大,节点数量固定时,在能耗方面,交叉点接近常数。 //This is because the energy consumption of both NB-TRACE and MC-TRACE //increases as the size of the area increases, //mainly due to the increase in the number of clusters //and in turn the additional control messages. 区域扩大,能耗增加,簇首增加,额外的控制信息增加。 //Hence the increase in the average energy consumption //is of the same order in NB-TRACE and MC-TRACE //and does not alter the value of the cross-over point significantly //for a fixed number of nodes. 节点数量固定时,能耗增加,交叉点的值并无明显改变。 } 8.{ //To sum up, for the goal of minimizing the number of transmissions, //both an increase in the number of nodes and an increase //in the size of the network favor multicasting over network-wide broadcasting, //making multicasting the optimal choice for a larger set of multicast members. 组成员个数增加,多播的性能不如广播。 //Similarly, considering the goal of minimizing the energy consumption per node, //an increase in the number of nodes in the network makes multicasting a better choice //up to a larger number of multicast members. 最小化每个节点的网络能耗。 //On the other hand, the relative efficiency of multicasting and network-wide broadcasting //in energy consumption is independent of the size of the network. 协议能耗与网络规模无关。 } } V.Conclusion{ 1.{ //In this paper, we examined the effect of the number of multicast members //on the relative efficiency of multicasting and broadcasting. 多播组成员个数与协议效率的关系。 //We consider two performance metrics: //energy efficiency and bandwidth spectrum efficiency. 衡量标准:能效和带宽利用率。 //We showed that for large multicast groups, //using broadcasting instead of multicasting //leads to 75% savings in the number of transmitted data packets //and up to 21% savings in the average energy consumption. 分组较大时,广播比多播节省75%的数据包传输,节省21%的平均能耗。 //Similarly, for small multicast groups, //multicasting reduces the number of data transmissions //and the average energy consumption by 50% and 13%, respectively. 分组较小时,多播可以减少50%的数据传输量和13%的能耗。 } 2.{ //We also showed that an increase in the total number of nodes in the region //decreases the relative efficiency of broadcasting compared to multicasting //for both performance metrics, //and hence the cross-over occurs at a larger number of multicast members. 区域节点个数的增加增加能耗。 //On the other hand, //the increase in the size of the area does not effect the cross-over point significantly //for the energy efficiency metric, //while it increases the cross-over point for the bandwidth efficiency metric. 区域大小对能耗交叉点没有明显的影响,但区域增大,带宽利用率的交叉点的值会增大 } 3.{ //Due to the similarity between the protocols,NB-TRACE and MC-TRACE, //they can be combined into a new unfied protocol, U-TRACE. 由于两个协议的相似性,可以被组合到一个新协议中 //Based on the a-priori information about the number of nodes and the size of the network, //the source node can choose the appropriate type of operation //considering the number of nodes in the multicasting group. 调整多播组成员个数根据网络规模和节点个数选择合适的操作。 } } Reference }
2010_IEEE_NB-TRACE_MC-TRACE
猜你喜欢
转载自blog.csdn.net/sun2014moon/article/details/80336447
今日推荐
周排行