Performance Analysis of Virtualised Head Nodes Utilising Cost-Effective Network Attached St

合集下载

IE论文——精选推荐

IE论文——精选推荐

A virtual network laboratory for learning IP networkingLluís FàbregaJordi MassaguerTeodor JovéDavid MéridaInstitut d’Informàtica i Aplicacions (IIiA)Universitat de Girona (UdG)Lluís Santaló Av., 17071 Girona, SPAIN+34 972418475{ fabrega | jmassa | teo | dmerida }@eia.udg.esABSTRACTIn this paper, a network laboratory for distance learning of basic concepts in IP networking is presented. Through a web interface,students can choose various configurations of a real private network (number of routers and subnetworks, use of IPv4/IPv6,etc.), and learn about them using Linux network commands. The utilization of the virtual laboratory and how it is implemented are described.Categories and Subject DescriptorsK.3.1 [Computers and Education ]: Computer uses in education–distance learning.General TermsDesign, Experimentation.KeywordsRemote laboratory, networking.1. INTRODUCTIONThe possibilities offered by the use of Internet in teaching activities are increasingly important. However, the physical presence of students in laboratories is required when the subject has a practical component, and this makes distance learning more difficult. Virtual laboratories can be used to overcome this situation.We have built a virtual network laboratory for distance learning of IP networking concepts such as IP addressing, routing tables,address resolution between IP and Ethernet, and the combined use of IPv4 and IPv6. Students access the virtual laboratory through a web interface and can change the network configuration by choosing one of the available preset configurations. Then, using Linux network commands, they can learn about these configurations and how to test the network.The virtual network laboratory is a private IP over an Ethernet network. It consists of several PCs (with one or more Ethernet cards) connected through a configurable Ethernet switch. One ofthese PCs, which is connected to the Internet, runs the web server and performs the different network configurations upon receiving a student’s request.The paper is organized as follows. In section 2 we describe the user interface of the virtual laboratory and some examples of lab classes to show how it can be used. Section 3 deals with the implementation of the virtual laboratory, the composition of the physical network and the remote configuration method used for the switch and for IPv4 and IPv6 in the PCs. Conclusions and future work conclude the paper.2. LEARNING IN THE VIRTUAL LABThe virtual network uses IP over an Ethernet and consists of four nodes. These nodes can be grouped to build IP subnetworks in different topologies, so they act as a host or router depending on the topology. The student can choose between four available topologies (see Figure 1).Our objective was to build a tool whereby remote students could learn the basic concepts of IP networking and related Linux network commands. We want them to learn the following:-IP addresses within an IP subnetwork share a common prefix that defines the subnetwork address.-The composition of the routing tables in hosts and routers,which defines the next node to send a packet depending on the IP subnetwork where the destination host belongs.Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists,requires prior specific permission and/or a fee.ITiCSE’02, June 24-26, 2002, Aarhus, DenmarkCopyright 2002 ACM 1-58113-499-1/02/0006…$5.00.Figure 1. The configured network topologies.host routerD subnet A BCDsubnet subnet C A B 12B subnet subnet subnetC A D34Dsubnet subnet subnet B C A-Address resolution between Ethernet and IP.-The main differences between the addressing scheme in IPv4 and IPv6.-The issues derived from the combined utilization of IPv6and IPv4, and the need for tunneling that allows hosts to send IPv6 packets through IPv4 networks.-The Linux commands for network administration such as ifconfig for interface addressing, route for the routing tables,traceroute for the routing path between two nodes, ping for checking connectivity, arp for the address resolution table,and others.In a lab class the student first chooses one of the available network configurations and then studies it by using Linux network commands. Section 2.1 describes the user interface and section 2.3 shows some examples of lab classes.2.1 The user interfaceStudents access the virtual laboratory through a web interface after an authentication phase. Then they choose one of the following options, “topologies”, “protocols”, “tunneling”,“commands” or “exit”:-The “topologies” option allows the student to choose one of several network topologies that differ in the number of IP subnetworks (1, 2 or 3) and therefore in the number of routers (see Figure 1).-Using “protocols”, the student can choose to configure each network node either with only IPv4 or with both IPv4 and IPv6. Predefined addresses are assigned to each node.-Using “tunneling”, the student can choose one of several network configurations where some nodes use IPv4 and others IPv6, which have configured the required tunnels to work properly (see Figure 2).-With “commands”, the student can choose any of the nodes of the network and then study how it is configured by using Linux network commands. The student is not allowed to change any configuration parameter, but can test it.2.2 Some examples of lab classesThe basic scheme of a lab class is first to choose one of the network configurations using the appropriate options (“topology”,“protocols”, and “tunneling”), and second, to study it using Linux network commands (“commands” option).A first example is the study of IP addressing and routing tables using the 4th topology in Figure 1 (three subnetworks and one router) with IPv4 in all the nodes. This results in the configuration of IP addresses and the routing table in each node. Then, using the ifconfig command, students can see the IP addressing used, by using route , the contents of the routing table, and by using traceroute , the routing path between two nodes.A second example is the study of the address resolution between IP and Ethernet using the third topology in Figure 1 (three subnetworks and two routers) with IPv4 in all nodes. The use of arp in each node allows the student to see the contents of the ARP (Address Resolution Protocol) table listing IP and Ethernet addresses. Other commands like ping , traceroute , or route , would be also interesting for studying the configuration.The third example is the study of the combined use of IPv4 and IPv6 nodes. Here the chosen topology is the third one in Figure 1,IPv4 in the B and C nodes and IPv4/IPv6 in the A and D nodes.First, no tunnel configuration is selected. Then, students can see that there is no virtual interface created in the A and D nodes by using the ifconfig command, that the routing table is wrong by using route , and that there is no connectivity between the A and D nodes by using ping . After that, the tunnel configuration number 2(see Figure 2) is chosen and the network configuration is tested again to check the right behavior.3. IMPLEMENTATION 3.1 EquipmentThe network of the virtual laboratory consists of four PCs running the Linux operating system (kernel 2.2.12-20, RedHat 6.1distribution) and an Ethernet switch (Cisco Catalyst 2920XL with 24 ports). Each PC has one or more Ethernet cards connected to the switch as shown in Figure 3. It is a private network that is accessed from the Internet through the node A, which has two Ethernet cards, one connected to the private network and the other to the Internet. Node A does not forward packets from or to the private network and the Internet.The Ethernet switch is used to build the subnetworks of each topology. Through its configuration, each switch port is assignedFigure 2. The configured tunnels.123Figure 3. The physical network and the groups of portsfor the third topology.ABCDEthernet switchto a specific group of ports. The different groups are isolated, i.e. the switch does not provide connectivity between two nodes that belong to different groups of ports. In this way, “virtual” Ethernet subnetworks are built, as if the nodes of each subnetwork were physically connected to different Ethernet switches. The groups of ports for creating the third topology (see Figure 1) are shown in Figure 3.3.2The remote configurationThe remote configuration of the PCs and the Ethernet switch is made through a web server (Apache release 1.3.9 [1]) running in node A. Using a CGI (Common Gateway Interface), the web server executes different UNIX scripts files. In turn these scripts execute a Telnet client that connects to the Telnet server running in the target device (PC or switch), and then executes the necessary commands to configure it. The script files use the expect scripts language [2] that allows CGIs to interact with the Telnet client, so an interpreter of this language needs to be installed in node A. This is the usual method for the configuration. The next sections focus on the configuration of the switch and the nodes.3.3The switch configuration and topologies The four different network topologies (see Figure 1) are built by configuring the (virtual) subnetworks in the switch. The number of subnetworks varies from one to three, as does the number of hosts that belong to a subnetwork. Note that only one of the two interfaces of node A can belong to the different network topologies, while the other interface is always connected to the Internet.The switch configuration is made via Telnet by using some specific commands [4] for assigning ports to (virtual) subnetworks. As explained in section 3.2, this is carried out by expect scripts files, which use these commands for each configuration. An example file is shown in Figure 4. First a Telnet session is initiated and then the commands for assigning ports to subnetworks are executed.3.4IPv4 and IPv6 in the nodesThe user can choose a topology and the IP protocol version for each node, i.e. only IPv4 or both IPv4 and IPv6 (there are no only IPv6 nodes). IPv4 is the default option. According to the selected option and topology, a predefined IP address (IPv4 and/or IPv6) is assigned to each network interface of the node depending on the subnetwork it belongs to. The corresponding routing table of the node is then configured. Moreover, nodes with interfaces in different subnetworks act as IP routers, so packet forwarding must be enabled in them.Linux systems configure network aspects at the boot time by executing several UNIX scripts files and using the information written in several configuration files [3]. All nodes are dual, i.e. both IP versions are installed on them. In order to configure IPv6, we have added some specific UNIX script files and configuration files [5], which are the equivalents to the installed files for IPv4. The configuration files for IPv4 are the following: ifcfg-eth0, for information about the Ethernet interface 0, such as the assigned IPv4 address (and the corresponding files for each interface); the file static-routes, for adding static routes to the routing table; the file network, for enabling or disabling networking and forwarding; the file hosts, with the table that relates DNS names and IP addresses. Network configuration is made at the boot time by executing the script network.The configuration files for IPv6 are the following: network-ip6, for enabling or disabling networking and forwarding, and also for defining the files names where there is the information about the interfaces, the routing and the tunnels; the file network-ip6.conf (or the file name defined in network-ip6), for the interfaces (equivalent to ifcfg-ethx files in IPv4) and for the static routes of the routing table (equivalent to the static-route file in IPv4); the file tunnels.conf (or the file name defined in network-ip6) for setting up virtual interfaces that define the IPv6/IPv4 tunnels (see next section); the file hosts, with the table relating Domain Name Services (DNS) names and IPv6 addresses (the same file as in IPv4). Network configuration is performed at the boot time by executing the script network-ip6.init.For each one of the different PCs’ configurations we have created the corresponding specific files. For example, for the file ifcfg-eth0, we have created the files 1.ifcfg-eth0, 2.ifcfg-eth0 and 3.ifcfg-eth0, corresponding to network topologies 1, 2 and 3 respectively, and using IPv4. Remote configuration is performed by replacing these files (e.g., ifcfg-eth0 by 1.ifcfg-eth0), and then enforcing the system to a network reconfiguration (e.g., executing the script network). As explained in section 3.2, this is done through expect scripts via Telnet.3.5IPv4 and IPv6 interconnectionIn order to provide connectivity between IPv6 nodes through a path that involves IPv4 nodes, it is necessary to set up tunnels between them, i.e. IPv6 packets are encapsulated in IPv4 packets and IPv6 addresses are mapped to IPv4 addresses at the sending point [6]. This operation is undone at the receiving point. Tunnels are created between two nodes by configuring a network virtual interface at both ends. The interface encapsulates and sends the packets to a real interface (i.e. eth0), and it is used in the IPv6 routing tables for a specific route. In Figure 2 three examples of tunnels are shown:-In the first one, there are three IPv6 subnetworks and two IPv6 routers. The second subnetwork uses IPv6 over IPv4.A tunnel is created between theB andC routers for therouting path that connects the first and the third subnetworks.-In the second example there is a single IPv6 subnetwork with IPv6 over IPv4. A tunnel is created between the A andD hosts.-In the third example there are three IPv6 subnetworks and two IPv6 routers. The first subnetwork uses IPv6 over IPv4.A tunnel is created between A andB for the routing paththat connects the first subnetwork with the other ones.3.6Remote commands executionOur platform allows the students to execute commands in a remote way at any of the nodes, such as ifconfig, route,and traceroute. These commands are executed in a Telnet session as a non-root user using expect scripts. As mentioned in section 2.1, students can test any configuration parameter but they are not allowed to change them.4.CONCLUSIONS AND FUTURE WORK We have built a virtual network laboratory for learning basic concepts of IP networking and have shown how students can use it. Using a web interface, students can choose one of the available configurations and study them by using (and learning) the main Linux network commands.There are several areas for future work. One is to allow students not only to choose and study predefined configurations, but also to build it on demand. In this way students would create their own topologies, assign IP addresses, and configure the routing tables (graphically and/or through Linux commands).Another interesting possibility is using the Simple Network Management Protocol (SNMP) for configuration of the network laboratory. We are also planning to add an enhanced user management (e.g. session scheduling, the ability of saving the configuration for recovery in a future session), and a network monitor for capturing Ethernet packets in order to study TCP/IP protocols, such as e-mail, web, DNS and others.5.ACKNOWLEDGMENTSThis study was partially supported by the CICYT (Spanish Education Ministry) contract TEL-98-0408-C02-01 and the Acciones Integradas program ref. HI1998-0032.6.REFERENCES[1]The Apache Software Foundation. HTTP Apache Server,[2]Don Libes. Exploring Expect: A Tcl-Based Toolkit forAutomating Interactive Programs. O’Reilly and Associates(1995)[3]Richard Petersen. Linux - Manual de Referencia.Osborne/MacGraw-Hill (1997)[4]Cisco Systems, Inc. Cisco IOS Desktop Switching SoftwareConfiguration Guide – Catalyst 2900 Series XL Cisco IOSRelease 11.2(8)SA6 (1999)[5]Peter Bieringer. Linux: IPv6.http://www.bieringer.de/linux/IPv6/index.html[6]Christian Huitema. IPv6 – The new Internet Protocol.Prentice Hall (1998)。

面向底层单节点失效的轻量级可靠虚拟网络映射算法

面向底层单节点失效的轻量级可靠虚拟网络映射算法

a t i f r s t , a n d t h e n t h e i s s u e i s f o r n ml a t e d a S a n I n t e g e r L i n e a r P r o g r a m( I L P ) b a s e d o n i t . F i n a l l y a n o v e l h e u r i s t i c
ma ppi n g.T he V N t o pol og y r e ma i ns c o nne c t e d e xc e pt f a i l e d vi r t ua l n ode i n t h e e v e nt of s i ngl e s ubs t r a t e node
g a i n i n g mo r e a n d mo r e a t t e n t i o n c u r r e n t l y . Th i s p a p e r f o c u s e s o n t h e i s s u e o f l e s s s t r i n g e n t r e l i a b l e v i r t u a l n e t wo r k
L i u Gu a n g y u a n S u S e n
( S t a t e Ke y L a b o r a t o r y Ne t w o r k i n g a n d S w i t c h i n g T e c h n o l o g底层单节点失效的轻量级 可靠虚拟 网络 映射算法
刘光远 苏 森
( 北京邮 电大学 网络与交换技 术国家重点 实验 室 北京 1 0 0 8 7 6 1
摘 要: 络 可靠性是虚拟 网络 设计 的一个重要指标 ,目前得到了越来越 多的关注 。该文对轻量级可靠的虚拟 网络

2017美赛数学建模M奖论文

2017美赛数学建模M奖论文

For office use onlyT1________________ T2________________ T3________________ T4________________ Team Control Number70028Problem ChosenBFor office use onlyF1________________F2________________F3________________F4________________2017MCM/ICMSummary Sheet(Your team's summary should be included as the first page of your electronic submission.)Type a summary of your results on this page. Do not include the name of your school, advisor, or team members on this page.SummaryThe performance of highway toll plaza directly affects the capacity of the highway, so the design of road toll plaza is imperative.In this paper, we conduct performance analysis for a specific toll plaza in New Jersey, USA, including accident prevention, throughput and cost. First of all, we usegrey model to predict the future output of the toll plaza, and compared with the realdata, the average value of the residual value is 0.429. Then we can draw a conclusionthat the throughput performance of the toll plaza is secondary. Next, we use queuingtheory to get the service index of the toll plaza in the light and heavy traffic, and thecellular automaton model is used to consider the changing circumstances of servicelevel, uses regression model to establish a function relation between traffic accidentand four factors. Then, we find that the rate of change has the greatest influence onit and the pavement performance has the least influence . In terms of cost, weconsider the toll plaza land and road construction. And the cost of road constructionis divided into the labor cost and material cost.Next, according to the influence of road geometry on the traffic performance of Toll Plaza, we select the transition curve trajectory model to improve the toll plazatransition, which can also have an improvement on the size and shape of the toll plazaand merge mode.Finally, we do a series of performance studies for our improved toll plaza. First of all, the improvement in the square flow and car flow under the condition of servicelevel are determined respectively through simulation .Next, we draw a conclusion thatthe service performance of the toll plaza is not obvious in small car flow, but there is amarked increase in large flow. Then, due to the fact that the unmanned vehicle coulddeal with a variety of road conditions, it undoubtedly expands our improved optionalscheme. Eventually, we obtain the throughput of toll before and after the improvementunder the different proportion of mixed charge mode and find that the improvedthroughput in the toll plaza has been increased on the performance.contents1 Introduction: (1)1.1 Problem background: (1)1.2 Steps: (1)1.3 Our work: (1)2 Assumptions (2)3 Nomenclature (2)4 Throughput analysis of grey forecasting model (3)5 error analysis (4)6 Service level of toll station (5)7 Vehicle lane changing rules based on Cellular Automata (6)8 Security analysis based on multivariate statistical regression mode (8)8.1 Study on the rate of change of Toll Plaza (8)8.2 Study on the longitudinal slope of entrance section of Toll Plaza (9)8.3 Research on service level of toll station (10)8.4 Study on pavement performance of toll station (10)9 Safety performance evaluation model of toll station (11)10 Cost analysis model of toll station (11)11 Analysis of the influence of lane geometry parameters on its capacity (12)11.1 Determination of lane changing rate (12)11.2 Influence of geometric parameters on the flow of the car lane (14)11.3 Energy consumption analysis based on cellular automata model (15)Definition of energy consumption: (16)Numerical simulation and analysis of the results: (17)Influence of curvature radius on energy consumption (17)Influence of arc length on energy consumption (18)12 The effect of traffic flow on service performance based on improved queuing theory 1913 The influence of unmanned vehicles on the improved model of Toll Plaza .. 2114 The influence of charging method on improving model of Toll Plaza (21)15 Strengths and Weaknesses (22)15.1 Strengths: (22)15.2 Weaknesses: (22)15.3 Future Model Development: (22)Comprehensive improvement strategy of tollplaza1Introduction:1.1Problem background:Highway toll and toll plaza is to ensure traffic safety and unimpeded, however because of lack of unified design specification, toll station and its square construction exists many problems. Such as: low value because of the technical indicators to make square construction scale too small and cause the toll plaza opened only few years as the traffic bottleneck, and use the high value on the one hand, because of the technical indicators and make the toll station construction scale is too large, waste a lot of money and resources. Due to incorrect linear indicators, or too short, the gradual square square length is insufficient, square road centerline offset, etc., it is too difficult to use after the completion of the square.so establishing the toll gates and the toll plaza design norms, as soon as possible, has the very vital significance in standardizing the construction of the toll station, ensuring the smooth general characteristic of toll plaza and traffic safety, improving the charging efficiency and management level, reducing the land acquisition and controlling investment and so on .1.2Steps:·A performance analysis of any particular toll plaza design that may already be implemented through the following three factors: accident prevention, throughput and cost .·Determine if there are better solutions (shape, size, and merging pattern) than any in common use.·Consider the performance of your solution in light and heavy traffic.·Consider the situation where more autonomous (self-driving) vehicles are added and how the solution is affected by the proportions of conventional (human-staffed) tollbooths, exact-change (automated) tollbooths, and electronic toll collection booths (such as electronic toll collection via a transponder in the vehicle)1.3Our work:·Based on the available data ,we make a performance analysis of any particulartoll plaza design that may already be implemented .·According to the problem from the performance analysis ,we make out a better solutions (shape, size, and merging pattern) than any in common use.·Determine the performance of the solution in light and heavy traffic ,how the solution change as more autonomous (self-driving) vehicles are added to the traffic mix and how the solution is affected by the proportions of conventional (human-staffed) tollbooths, exact-change (automated) tollbooths, and electronic toll collection booths.2AssumptionsTo simplify the problem and make it convenient for us to simulate real-life conditions, we make the following basic assumptions.1. Each section of roads is one-way traffic2.Vehicles in the retention period of toll station can be neglected3.In any hour of the vehicle arrival rate is proportional to the length of time4.The probability of any vehicle arrival in one hour of time is not affected by the previous history .5. The vehicles arrive in line with the Poisson distribution, namely the headway is negative exponential distribution3Nomenclatureε(0)(t)the residual errorq(t)the relative errorc the variance ratioP the small error probabilityr the curvature of the bend radiusu the static friction coefficientl the gradual change ratiok the number of serving drivewayρ/k traffic intensityw mean time to stay at a toll stationd automotive braking distancef the tire and road surface friction coefficientY the number of traffic accidents in toll stations per year∆W Width of the gradualα1curve angle R 1the radius of convex curve points pdelay probability e(n,t) energy consumption of the first n vehicles from time t to t+14 Throughput analysis of grey forecasting modelFigure 4-0-1Schematic diagram of New Jersey toll plazaFirst of all, we chose a toll plaza on the New Jersey in the United States for a specific performance analysis of toll plaza, and it includes the accident prevention, throughput, and cost.In view of the throughput of the toll plaza, we choose the grey forecasting model GM(1,1) , to predict the throughput of the toll plaza. Due to the problem of uncertainty, so we take the grey prediction model to deal with it.Suppose x (0)(1),x (0)(2)…,x (0)(M )In order to overcome the irregular , we use accumulation processx (1)(t )=∑x (0)(i)M i<1 Such a relatively smooth new series approximation can be described by the following differential equation:dx (1)dt +ax (1)=μ Its an albino form discrete solution of differential equation is: x ̂(1)(i +1)=.x (1)−u a /e ;ai +u aThe type of the parameter a、u be determined by the least squares fitting method is as follows:(1)(2)(3)A ̂=0a u 1=(B T B);1B T Y N Among them the matrix is:B =[ −12,x (1)(1)+x (1)(2)-1−12,x (1)(2)+x (1)(3)-1⋯⋯−12,x (1)(m −1)+x (1)(m )-1] Y N =(x (0)(2),x (0)(3),⋯,x (0)(m ))TSo the original data fitting sequence is:x ̂(0)(1)=x (0)(1)x ̂(0)(i +1)=x (1)(i +1)−x (1)(i )Table 4-0-1 Traffic flow prediction table5 error analysisIn equation (11), and regulations, the original data of reducing value and its residual error and relative error between observed value is as follows{ε(0)(t )=x (0)(t )−x′(0)(t )q (t )=ε(0)(t )x (0)(t )×100%The following inspection of the accuracy: x(0)=1M ∑x (0)(t )M t<0 ε(0)=1M;1∑(ε(0)(t )−ε0M t<2)2Second, calculate the variance ratio c =s 2s 1and small error probability P =2|ε(0)(t )−ε(0)|<0.6745s 13(4)(5)(6) (7) (8) (9) (10) (11)Figure 5-0-2comparison chart of grey prediction modelWe use m、p、v max to represent quality of the vehicle, random delayprobability and maximum speed respectively, g represents the local acceleration of gravity, r represents curvature of the bend radius and u represents the static friction coefficient . With the road statistical analysis carried out on the real value and the error of predicted value, we obtain the following res ults:It shows that the GM(1,1)model prediction results have a better response to .reflect the actual situation.6 Service level of toll stationThe direct feeling of the driver to the traffic environment of the toll station is from the queue length of the toll lane, and the length of the queue depends on the service level of the toll station V/C. In this regard, we use the queuing theory model of multichannel Queuing service, in which the vehicle arrival time is in a Poisson distribution, which is the negative exponential distribution; Suppose m is random arrival rate ,c i is output rate,k is the number of serving driveway, ρ=m c .There is the probability of having no vehicle in the queuing theoryρ(0)=1,∑1n!k−1n=0p n :1k!ρk k k−ρ- Average number of vehicles in queueing theory:n =ρ+p n ρ(0)k!k n−k (1;ρk )2 (12)(13)queue length: q =n −ρ=p n ρ(0)k!k n−k (1;ρk )2 Average number of waiting vehicles per lanea =q kAverage waiting time in queue systems:d =n m =q m +1c Average waiting time in queue:W =q mMean tardinessDeceleration time of vehicle entering toll stationt 1=v 03.6a 1Mean time to stay at a toll stationw =E ,S -+W qVehicle acceleration time of leaving toll stationt 2=v 03.6a 2 In this equation, v 0 is the normal traffic flow (km/h); a 1 、a 2 are deceleration of the vehicle (m/s 2); W q is average queue time (s); E ,S - is expected service time (s);7 Vehicle lane changing rules based on CellularAutomataWe apply the previous cellular automata model, which is now extended to multi Lane case. The main difference between multi lane and single lane is to consider the model of lane changing. In this paper, we take 4 lanes as an example.In reality, it may be possible to change lanes when the driver is found to be close to the exit and the front of the adjacent lane is empty. If you want to change lanes ,you should consider the vehicle behind the adjacent lane. When the distance (14)(15) (16) (17) (18) (19) (20)to the rear of the adjacent lane reaches to a certain length, you can change the road. Lane change scenarios can be shown in figure (), when the c car on the 1 Lane is blocked by the c 1 car, while the c 2 and c 3 cars on the 2 lanes are relatively large. in order to maintain the speed, c car will change to the road lane 2.Figure 7-0-3Schematic diagram of lane changingWhether or not the driver chooses the lane change is mainly decided by the d 0,d n,otℎer 、d n ,back three indicators, through the previous research, this paper thinks that the lane changing rule is:When d n,back >v maxC n ={1−C n d n <min{v n +1,v max } d n,otℎer >d n ,d n,back >v max c n Otℎer circumstancesWhen d n,back ≤v max ,C n ={1−C n d n <min{v n +1,v max } d n,otℎer >d nv max −θ(−∆x )α>1+min{d n,otℎer +1,v max }−min *V n +1,v max +c n Otℎer circumstancesAmong them, C n is the n car in the lane , C n =0 or 1,d n 、d n,otℎer andd n,back are the distance between the first n vehicle and the front vehicle, the distance from the adjacent lane and the distance from the vehicle in the adjacent lane, respectively. d safe is safety lane change model.d n,back −v max , ∆x <0, v max −θ(−∆x )α is the distance between the vehicle and the vehicle in the adjacent lane after correction by the value function, 1+min{d n,otℎer +1,v max }−min *V n +1,v max + is Limit Lane distance. The parameters α and θchange according to the psychological status of driver. If α>1, the greater α is, the more careful the driver is. If θ>1, the greater θis, the more careful the driver is. When α=1,θ=1,that ’s Lane changing model.(21) (22)In order to discuss the αandθ, we use Cellular automata simulation. In a two lane road with a length of7.5km, adopting the open boundary condition, each lane is composed of 1000cells with a length of7.5km, the maximum speed of vehicle v max=5. The random slowing down rate was 0.2.8Security analysis based on multivariate statistical regression modeAimed at the prevention of the accident, we use multiple linear regression to establish a function between the number of traffic accidents and the following four factors: toll square gradient, service level, Toll plaza entrance section of the longitudinal slope, the Pavement performance of Toll station .Figure 8-1Cause analysis of accident8.1Study on the rate of change of Toll PlazaFan in and fan out area of toll plaza are designed to make the gradual vehicles more natural smoothly in and out of the toll plaza. In order to drive vehicle easily , there has a requirement on its gentle gradient change. Otherwise the driver could produce driving deviation, which may cause improper operation and endangers safety.The relationship is as follow:(23)l=b,LAccording to the experience, the vehicles with straight into another lane deviation than at around 0.9m s⁄, drivers usually have no move feeling and uncomfortable feeling.Figure 8-1 The relationship between Accident number and Toll plaza ramp rateFigure 8-1 shows the relation curve between highway toll plaza ramp rate and traffic accident, the figure demonstrates that as the toll plaza ramp rate increases, the traffic accidents will increase, whereas the security of the toll plaza will decrease.Through the data regression analysis, we get the related models between toll plaza ramp rate and the number of traffic accidentsY =1.423e .0064xIn this equation, Y is the forecasted numbers of traffic accident corresponding to the toll plaza ramp rate , x is the toll plaza ramp rate of toll plaza.The correlation coefficient in the model R 2=0.8621, it shows that description model of correlation is higher, From the model ,we can learn that the occurrence of traffic accident frequency is proportional to the toll plaza ramp rate. Gradient length is insufficient, so it can't meet to slow down and change lanes entering the toll plaza vehicle safety requirements, resulting in the occurrence of traffic accidents .8.2 Study on the longitudinal slope of entrance section ofToll PlazaHighway toll entrance section of the longitudinal slope design without fully considering the characteristics of vehicles entering the toll plaza, a long downhill or turn downhill and so on bad road alignment, those will affect the normal operation of the pilot and make the vehicles entering the toll plaza slowdown not sufficient, longitudinal safe driving distance not enough and driving direction can't adjust to the charge lane ,which will causetraffic accidents. This will lead to serious losses. (24)Figure 8-0-4 entrance section of the longitudinal slope and accident numberThrough regression analysis, we get the relevant model between the toll plaza entrance section of longitudinal wave and traffic accidentsY =2.6254e 0.638xIn this equation, Y is the forecasted numbers of traffic accident corresponding to the toll plaza ramp rate , x is the longitudinal wave of t oll plaza’s entry section .The correlation coefficient in the model R 2=0.9219,it shows the correlation of this model is relatively high. But we can learn that toll station ‘s traffic accident and its entrance section of longitudinal wave have a positive correlation from figure model representation ,.The greater the slope, the lower charge war security.8.3 Research on service level of toll stationBased on the previous research of service performance of toll station, we take V C as the measure of service level and Cite previous results. 8.4 Study on pavement performance of toll stationAccording to the vehicle dynamics, the vehicle's braking distance can be expressed as follows:d =u 257.9(f:I) In this equation, d is automotive braking distance , u is the speed at the beginning of the automobile brake, f is the tire and road surface friction coefficient, Iis road longitudinal slope(25)9 Safety performance evaluation model of toll stationBased on the above analysis, the evaluation model of descriptive can be written as the equation form, using multiple linear regression model .Y is the number of traffic accidents in toll stations every year , x 1=1l ,x 2=V C ,x 3=i,则Y =β0+β1x 1+β2x 2+β3x 3N is sample size , Y i (i =1,2,…,N ) represent the Y value of sample i , x i 1,x i 2,…x i n (i =1,2…,N) represent the value of each variable insample I, respectively.令Y =[Y 1Y 2⋮Y n], X =[11⋮1x 11x 21⋮x n 1⋯⋯⋮⋯x 1n x 2n ⋮x n n ] β=[β0β1⋮βn ] Y =Xβ,making maximum likelihood estimate of each variable coefficient β1,β2,…βn , it can get a normal equations:X T Xβ=X T YSo you can get the following regression equationY =−4.4012−9.947511l +10.098V C +11.25i 10 Cost analysis model of toll stationWe selected the American New Jersey a toll plaza to make cost analysisFirstly, according to relevant data, we learn that New Jersey’s average price is (26) (27)(28)(29)(30)$3500 per mu,And the toll plaza which we analyzed occupies about 5 mu, therefore, the land price of the toll plaza is about $17500;Second, the road construction costs include labor and material cost, and the local construction industry ’s average monthly salary is $3000, we use it to calculate labor, this occupies the largest in the road construction costs; As for material cost, we calculate by the current prices in the United States, is about $40 per cubic meter, then according to the size of the toll plaza, it will cost about $45000.In conclusion, the cost of toll plaza spend mainly on the labor cost of highway construction, the material cost also accordingly account for part of it.11Analysis of the influence of lane geometry parameters on its capacity11.1Determination of lane changing rateAccording to the analysis of vehicle trajectory and running state of vehicle , vehicle trajectory in the middle of the gradual path is similar to vehicle lane changing trajectory, and considering the factors when the driver turns, we select the easement curve trajectory model to design the gradual change section of toll plaza. And in the middle of the two convex type curve , we join a long for L straight section , it is shown in the figure belowFigure 11-0-5Toll plaza improvementsAccording to characteristics of convex curve geometric elements, we can use the following formula to calculate the first period of convex curve of easement curve tangent length T1:T1=(R1+p1)tanα1+q1(31)2In this equation, R 1 is the radius of the first section of convex curve points , ρ1 is Within shift, q 1 is tangent increment, α1 is curve angle, and α1=2β1, β1 is easement curve angleSuppose the first and second convex curve gradient width are ∆W 1 and ∆W 2 respectively, the width of one side with the gradient is ∆W .Depending on the figure with the easement curve in orbit, there are: ∆W 1=T 1∙sin α1∆W 2=T 2∙sin α2∆W =∆W 1+∆W 2+Lsinα1∆W =0(R 1+p 1)(1−cos L S1R 1 )+q 11∙sin L S1R 1 +0(R 2+p 2)(1−cos L S2R 2)+q 21∙sin Ls2R 2 +Lsinα1 L S1 and L S2 are the length of easement curve of two convex curve respectivelyL is radial tangent of two convex curve, so α1=α2,then it Can be introduced as follows:L S1R 1 =L S2R 2 Associate (38) and (39),we can get the length of easement curve of two - Section convex curve L S1 and L S2, then the transition section longitudinal distance L y can use the following formula to calculate:L y =[(R 1+p 1)tan L S12R 1 +q 1+(R 2+p 2)tan L S22R 2 +q 2](1+cos L S1R 1 )+Lcosα1 Suppose the ramp rate of transition period is K ,then we can adopt the following equation:K =∆WL y From this equation , we can learn that the driving radius and the straight line segment L have a great influence on the length and the gradient of the gradient. The greater the radius, the longer the straight line, the longer the length of the gradient, the smaller the rate of change(32) (33) (34) (35) (36)(37)(38)11.2 I nfluence of geometric parameters on the flow of thecar laneAssuming C 0 and C 1=dC dl represent respectively bend and itsgradient , l represents the length of the curve itself , we can get C (l )=C 0+C 1lso ,the bend of the direction Angle isφ(l )=φ0+∫C(τ)l 0dτ=φ0+C 0l +12C 1l 2 The bend of the longitudinal distance x(l) and transverse distance y(l) are{x (l )=x 0+∫cosφ(τ)dτl 0y (l )=y 0+∫sinφ(τ)dτl 0 Assuming sinφ≈φ,cosφ≈1,and when x 0(l )=0,x (l )=l , then the bend of transverse distance y(x) and direction angle φ(x) can be expressed{φ(x )=φ0+C 0x +12C 1x 2y (x )=y 0+φl +12C 0x 2+16C 1l 3 Using the ideas of analytical mechanics, assuming that the longitudinal velocity along the x axis for x ′, along the y axis transverse speed for y ′ , along the z axis of horizontal pendulum angular velocity as the bits of ψ′, then from The Lagrange's equations we can get{ d dt .ðE T ðẋ/−ψðE T ðẏ=F Q 1d dt .ðE T ðẏ/+ψðE T ðẋ=F Q 2d dt .ðE T ðψ/+ẋ ðE T ðẏ−y ðE T ðẋ=F Q 3 Defining the system kinetic energy E T =12m (ẋ+ẏ)+12I z ψ2In the formula, m,I z respectively represent Vehicle quality and Rotary inertia take the derivative of (46),we can get{ d dt .ðE T ðẋ/−ψðE T ðẏ=d dt(mẋ)−ψ (mẏ)d dt .ðE T ðẏ/+ψðE T ðẋ=d dt (mẏ)−ψ (mẋ)d dt .ðE T ðψ/+ẋ ðE T ðẏ−y ðE T ðẋ=d dt (I z ψ)−x (mẏ)−y (mẋ) (39)(40)(41)(42)(43)Delimiting generalized force: {F Q 1=∑F xF Q 2=∑F y F Q 3=∑M zIn summary we can get the Vehicle longitudinal coupling model.We mainly consider the lateral situation∑F y =F yr +F xf +F xf cosδ If the vehicle driving in the bend is only disturbed by small disturbance near the equilibrium state, the front wheel angle is small enough , so cosδ≈1,sinδ≈δ ∑F y =−(C f +C r )y ẋ−(aC f −bC f )ψẋ+(F xf +C f )δ We put the formula () and formula () into ()y =−d 2ẏẋ−.ẋ+kd 3ẋ/ψ−(F xf :C f m )δ In the formula d 2=C f :C r m ,d 3=aC f ;bC rI z ,k =I z mThen, the resultant force ∑M z along the vertical direction is∑M z =aF xf sinδ+aF xf cosδ−bF yrWhen sinφ≈φ,cosφ≈1,then∑F y =−(a 2C f +b 2C r )ψẋ−(aC f −bC f )ẏẋ+a(F xf +C f )δψ=−d 4ψẋ−−d 3y ẋ+a I z (F xf :C f m )δ In the formula, d 4=(a 2C f :b 2C r )I z 11.3 E nergy consumption analysis based on cellularautomata modelConsidering the influence of different shapes on traffic performance is mainly reflected in the curve, we mainly study the influence of the curve on the whole problem. On the road segment, Lane set of sections containing only one plane curve, the curve is provided with the deceleration section of L , the road will be regarded as the length of the L 1D discrete lattice chain, each lattice point at each moment or is empty or occupied for a car.m 、p and v max represent the quality of the vehicle, the (44) (45)(46)(47) (48) (49) (50)(51)stochastic delay probability and maximum speed ,respectively, g is the local acceleration of gravity, r and u represent the static friction coefficient of curvature radius and static coefficient of friction between wheel and road, respectively. The vertical direction of the vehicle is subjected to a pair of balance forces, and the influence of tangential friction on the vehicle is mainly reflected in the change of the speed, Therefore , the centripetal force required for the safety of the vehicle is provided by the normal static friction force,v safe is maximum speed of safetyturning, then mv safe2r =μmg,⁄v safe =√μgr .In each step of t →t +1 , all vehicles are in accordance with the following rules of the evolution of the speed and location of the synchronization update :Determine the vehicle delay probability p :When the vehicle is in the buffer section , if v >v safe,take the probability of delay p =p 1 (larger), in other cases, take p =p 2 (smaller),Acceleration process: v n (t)→min (v n (t )+1,v max );deterministic deceleration process: v n (t)→min (v n (t ),gap n (t))Stochastic deceleration process with probability p :v n (t)→max (v n (t )−1,0) deceleration process :When the vehicle is in the corner of the road, and the speed v (t )>v safe , in order to turn the corner ,it must be slowed down :v n (t)→min (v n (t ),v safe )location update process: x n (t )→x n (t )+v n (t)Among them, v n (t) and x n (t ) are the speed and position of the first n vehicle at time t respectively , x n:1(t ) is the position of the first n +1 vehicle at time t . gap n (t )=x n:1(t )−x n (t )−1is the spacing between the first n car and the foregoing vehicle which is close to it.Definition of energy consumptionSuppose the mass of vehicle is m , when it slows down, its kinetic energy is reduced, we define the kinetic energy reduction for energy consumption, e(n,t) represents that energy consumption of the first n vehicles from time t to t+1 .e (n,t )={m,v 2(n,,t );v 2(n,,t:1)-2v (n,t )>v (n,t +1);0,v (n,t )≤v (n,t +1)The average energy consumption per vehicle per unit time:E d =1T 1N ∑∑e(n,t) N n<1t0:T;1t<t0 N is the total number of vehicles on the driveway, t 0 is relaxation time. For(52) (53)the energy consumption of the vehicle, if it is because the speed of t moment is greater than the Vehicle-to-vehicle distance v(n,t)>gap n(t), the vehicle decelerates, thatis defined as the interaction energy, denoted by E di; If it is because of the random deceleration caused, defined as the random deceleration energy consumption, denotedby E dr;if it is because the car speed In the corner v(n,t)>v safe, there is deceleration for the sake of driving safely, defined as safe energy consumption, denoted by E ds.Then total energy consumption is:E d=E di+E dr+E ds(54)Numerical simulation and analysis of the resultsTo simplify the problem, assuming that the length of actual road is 7.5km, Divided into 1000lattices, equivalent to the actual length of each grid correspondsto 7.5m, Delay probability p1=0.8,p2=0.25,Quality unit is defined 1. Entering probability changes from 0~1.0.The state of each vehicle is represented by its own speed v, v∈,0,v max-We let v max=5cell he actual speed is135km/h.We take8×104time steps every run .Influence of curvature radius on energy consumptionThe arc length s, the friction coefficient μand the radius of curvature of r are carried out numerical simulation. parameters are as follows: s=30m,μ=0.5,r=10、50、100、200、300m.According to v max=5cell/s,the maximum speed of the vehicle v max=37.5m/s. Results show that when r=300m, the safetyspeed v safe=√μgr=38.73m/s,v safe>v max, the bottleneck of the curve disappears and the speed limit is lost. The change of the probability in_p of therandom energy consumption(E di、E dr、E ds、E d)is shown in the figure.。

Bellhop模型在水声网络仿真中的实现和应用

Bellhop模型在水声网络仿真中的实现和应用

Bellhop 模型在水声网络仿真中的实现和应用刘奇佩 1, 刘 琨 2, 罗逸豪 1, 吴鑫莹 3*, 周河宇1(1. 中国船舶集团有限公司 第710研究所, 湖北 宜昌, 443003; 2. 国家计算机网络应急技术处理协调中心 黑龙江分中心, 黑龙江 哈尔滨, 150001; 3. 华东理工大学 艺术设计与传媒学院, 上海, 200030)摘 要: 随着水声技术的发展, 水声网络(UANs)因其在海洋监视、灾害预警和海洋安全等领域的表现而备受关注。

水声信道是影响UANs 性能的重要因素之一, 其复杂特性直接影响着UANs 相关协议的前期设计和评估, 对于协议走向实际应用至关重要。

有别于传统理论模型, Bellhop 水声信道模型通过跟踪射线计算海洋声场, 提供了一种更准确的获得不同海洋环境下信道特性的方法, 但该方法不能直接用于网络仿真。

针对此,文中在目前主流的网络仿真平台NS3上构建了基于Bellhop 的水声信道模型, 将高斯射线模型用于水声网络仿真。

对比结果表明, 该模型能够有效仿真声信号在水下的传播特性, 可为实际UANs 协议开发提供参考。

关键词: 水声网络; 信道模型; Bellhop; NS3中图分类号: TJ630.34; U674.76 文献标识码: A 文章编号: 2096-3920(2024)01-0124-06DOI: 10.11993/j.issn.2096-3920.2023-0015Implementation and Application of Bellhop Model in Underwater AcousticNetwork SimulationLIU Qipei 1, LIU Kun 2, LUO Yihao 1, WU Xinying 3*, ZHOU Heyu1(1. The 710 Research Institute, China State Shipbuilding Corporation Limited, Yichang 443003, China; 2. Heilongjiang Branch of National Computer Network Emergency Technology Processing and Coordination Center, Harbin 1500011, China; 3. School of Art Design and Media, East China University of Science and Technology, Shanghai 200030, China)Abstract: With the development of underwater acoustic technology, underwater acoustic networks(UANs) have attracted much attention due to their performance in marine surveillance, disaster warning, and ocean security. The underwater acoustic channel is a crucial factor affecting the performance of UANs, and its complexity directly affects the pre-design and evaluation of UAN-related protocols, which is crucial to the practical application of protocols. Unlike traditional theoretical models, the Bellhop underwater acoustic channel model provides a more accurate method to obtain channel characteristics under different oceanic environments by calculating their acoustic fields via ray tracing. However, it cannot be directly applied to network simulation. This paper implemented a Bellhop underwater acoustic channel model based on NS3, the current most popular network simulation platform, and applied the Gaussian ray model to UAN simulation. The comparison results show that the model can effectively simulate the underwater propagation characteristics of acoustic signals and provide a reference for practical UAN-related protocol development.Keywords: underwater acoustic networks; channel model; Bellhop; NS3收稿日期: 2023-02-20; 修回日期: 2023-05-10.作者简介: 刘奇佩(1988-), 男, 博士, 工程师, 主要研究方向为水声网络技术及水声信号处理.* 通信作者简介: 吴鑫莹(1989-), 女, 博士, 讲师, 主要研究方向为计算美学和智能算法.第 32 卷第 1 期水下无人系统学报Vol.32 N o.12024 年 2 月JOURNAL OF UNMANNED UNDERSEA SYSTEMS Feb. 2024[引用格式] 刘奇佩, 刘琨, 罗逸豪, 等. Bellhop 模型在水声网络仿真中的实现和应用[J]. 水下无人系统学报, 2024, 32(1): 124-129.0 引言水声网络(underwater acoustic networks, UANs)可以用于海洋资源探索、辅助导航、自然灾害预警以及海域监控等多个领域[1-5], 在军事和民用方面表现出巨大潜力, 近年来受到各国研究人员的广泛关注。

Flattened-Butterflay-A-cost-Efficient-Topology-for

Flattened-Butterflay-A-cost-Efficient-Topology-for
The performance of the interconnection network plays a central role in determining the overall performance of the system
Low-radix networks, such as k-ary n-cubes, are unable to take full advantage of the increased router bandwidth
Flattened Butterfly : A Cost-Efficient Topology for High-Radix Networks
John Kim, William J. Dally & Dennis Abts Presented by
Ajithkumar Thamarakuzhi
1
Outline
2-ary 4-fly butterfly
7
Butterfly to Flattened Butterfly
Flattened butterfly can be constructed by combining or flattening the routers in each row of the conventional butterfly network a into a single router.
number of routers = N/k Radix k’ =n(k-1)+1 The routers are connected by channels in n’ = n − 1
dimensions
8
Butterfly to Flattened Butterfly

基于CPN的安全协议形式化建模及安全分析方法

基于CPN的安全协议形式化建模及安全分析方法

2021年9月Journal on Communications September 2021 第42卷第9期通信学报V ol.42No.9基于CPN的安全协议形式化建模及安全分析方法龚翔,冯涛,杜谨泽(兰州理工大学计算机与通信学院,甘肃兰州 730050)摘 要:为了解决有色Petri网(CPN)对安全协议进行形式化建模分析时,仅能判断协议是否存在漏洞而无法找出漏洞具体位置和攻击路径的问题,以及CPN建模时随着攻击者模型引入,安全协议的形式化模型可能的消息路径数量激增,状态空间容易发生爆炸导致难以提取准确攻击路径的问题,改进了基于CPN的安全协议形式化建模方法,验证并提取攻击路径的同时,采用更细粒度的协议建模及控制。

在状态空间收敛方面提出了CPN模型不同进程在各分层模型中等待−同步的方法控制状态空间规模。

通过针对TMN协议的安全评估分析,成功提取出该协议25条攻击路径,评估了该协议安全性的同时证明了所述方法的有效性。

关键词:有色Petri网;安全协议;形式化分析;状态空间;攻击路径中图分类号:TP393.06文献标识码:ADOI: 10.11959/j.issn.1000−436x.2021175Formal modeling and security analysis method ofsecurity protocol based on CPNGONG Xiang, FENG Tao, DU JinzeSchool of Computer and Communication, Lanzhou University of Technology, Lanzhou 730050, China Abstract: To solve the problem of modeling and analyzing with colored Petri net (CPN), which was determining vulne-rabilities in hole location but couldn’t identify any attack path, and the problem of when the introduction of the attacker model, the number of possible message paths in the CPN formal model of security protocol surges the state space prone to explosion, which made it difficult to extract accurate attack paths, the formal modeling method of security protocol was improved base on CPN, the attack paths were verified and extracted, further the fine-grained protocol modeling and control were adopted. As well as in the aspect of state-space convergence, and a waiting-sync method for different processes of CPN model in each hierarchy model was proposed, which effectively controlled the state-space scale of the model. Through the security evaluation and analysis of TMN protocol, 25 attack paths of the protocol are extracted suc-cessfully, the security of the protocol is evaluated, and the effectiveness of the proposed method is proved.Keywords: colored Petri net, security protocol, formal analysis, state space, attack path1 引言安全协议已成为现代计算机网络正常运转的基础,但由于其设计阶段的规范缺失和不可避免的逻辑缺陷,常会带来潜在的安全隐患,使各种协议的开发和安全性验证成为一项艰巨的任务[1]。

Synergistic policy and virtual machine consolidation in cloud data centers-slides

Synergistic policy and virtual machine consolidation in cloud data centers-slides

how are servers managed?
cloud data centre server management
A large shared compute environment where tenants (users) come and go unpredictably - hence, unpredictable dynamic workload. Virtualisation (e.g. virtual machines) enables sharing and provides elasticity (i.e. dynamic scaling). Virtual machines can be dynamically consolidated – via live migration – to improve resource utilisation, hence the revenue ($).
4 Turner, D. et al., “NetPilot: Automating Datacenter Network Failure Mitigation.” ACM SIGCOMM 12 Lin Cui Richard Cziva Posco Tso Dimitrios P.Synergistic Pezaros Policy and Virtual Machine Consolidation in Cloud INFOCOM, Data Centres 2016 9 / 29
network policy
All networks are governed by network policies, which are high-level networking objectives derived from network-wide requirements3 : Internet client traffic must be checked by Firewall first and then forwarded to a LB. Traffic to access database server must be checked by an IPS.

北京邮电大学研究生英语学术写作作业-分析一篇英文文献

北京邮电大学研究生英语学术写作作业-分析一篇英文文献

《研究生英语学术写作》期中作业学号:姓名:1.Baic InformationTitle: Multi-UAV Routing for Persistent Intelligence Surveillance & Reconnaissance Missions Author:Satyanarayana G. Manyam, Steven Rasmussen, David W. Casbeer, Krishnamoorthy Kalyanam and Suresh ManickamSource:2017 International Conference on Unmanned Aircraft Systems (ICUAS)2.Key sentence patterns in each sectionAbstract:We consider a Persistent Intelligence, Surveillance and Reconnaissance (PISR) routing problem, which includes collecting data from a set of specified task locations and delivering that data to a control station.①We consider a … problem,which includes …The algorithms are tested on several instances and their performance is compared with respect to the optimal cost and computation time.①The algorithms are tested on … and their performance is compared with…I.IntroductionThe various NPSBN stakeholders, (e.g., network planners, public safety agencies, and equipment vendors) have an interest in being able to quantitatively measure the resiliency of proposed and fielded networks.We consider a persistent monitoring scenario, where a set of task locations needs to be visited persistently by multiple UA Vs.We assume all the available UA Vs are homogeneous.①We consider a… scenario, where….②We assume….We are interested in two metrics viz. data latency or delivery time (to the control station) and revisit rate or revisit period. We define the data delivery time (or latency time) as the time elapsed from collection of data from a task to the time the data is delivered to the control station.①We are interested in …②We define the…as…In this article, we consider a persistent routing of tasks that are spatially distributed. Also the data collected at the task locations needs to be delivered at a control station (also referred to as depot). In the existing literature concerning persistent routing, the concept of a control station is not considered and delivery time is not addressed.①In this article, we consider… defined …②In the existing literature concerning …the concept of… is not considered and…is notaddressed.We model this persistent routing problem as a multiple traveling salesman problem with revisit period constraints, and formulated as a mixed integer linear programming (MILP) problem.①We model this … problem as a … problem, and formulated as a … problem.The contributions of this article are the following: (i)We present a novel formulation addressing two important metrics, delivery time and revisit period for the tasks in ISR missions and model it as a multiple vehicle path planning problem with cycle length constraints. (ii) We present two different MILP models to find optimal solutions to the corresponding path planning problem. The two MILP models constitute novel constraints to address the cycle length limits, which could be applied to any general routing problem involving constraints on cycle length. (iii) A heuristic via assignment-tree search is presented that produces good sub-optimal solutions, and it could beeasily generalized to address different cost functions and/or constraints. (iv) We test the algorithms on several random instances and computational results are presented.① The contributions of this article are the following: (i) … (ii) … (iii) …② We present a novel formulation addressing two important metrics …③ We present … model to find optimal solutions to … problem.④ We test the algorithms on …II. PRELIMINARIES AND ASSUMPTIONSSome of the advantages of this class of solutions are the following: we do not need to have communication between UA Vs to update the scheduled tasks at each planning time interval. ① Some of the advantages of this class of solutions are the following: …There are two important metrics that needs to be addressed in PISR missions. The first one is the data delivery time or data latency (i D ) for each task i t . The other metric that we consider is the revisit period of each task.① There are two important metrics that needs to be addressed in … missions. The first one is … The other metric that we consider isIII. PROBLEM FORMULATIONWe model the MILP using node based and arc based formulations; these models are akin to the models in [14] and [12] used to solve the traveling salesman problem with time windows and the distance constrained vehicle routing problem. Similar formulations were also used to solve fuel constrained multiple vehicle routing problem in [13]. In these articles, the constraints on the length of a tour starting from a depot are constrained. In the formulation presented here, the length of a tour starting from a depot to each task and the length starting from the task returning to the depot together are constrained.① We model the …using …② Similar formulations were also used to solve … problem in …③ In these articles, … are constrained.The problem can be stated as the following: find at most v n cycles that minimizes the maximum delivery time such that, ()i each task T is covered by one cycle, and ()ii if a task i t assigned to one of the UA Vs,v , with cycle length v L , then ,v i L R i T ≤∀∈.① The problem can be stated as the following: …()i … ()ii …In the above formulation, the big M -in the constraints (4), (6) is known to cause computational problems [14], [17], and hence make the MILP model computationally less efficient. We propose a second formulation without big M - constraints and compare the computational performance of these two formulations.① In the above formulation, … is known to cause computational problems② We propose a second formulation … and compare the computational performance of these two formulations.IV. ASSIGNMENT TREE SEARCH HEURISTICIn this section, we present a heuristic to solve the PISR routing problem. The heuristic is a greedy assignment tree search, based on the prior work in [18], [19], for planning missions involving multiple UA Vs. Here, we present a synopsis of the tree search algorithm, however one can refer to [18], [19] for further details. This tree search follows a best first search pattern until it finds a feasible assignment.①In this section, we present a ... to solve ... problem②Here, we present a synopsis of the tree search algorithm, however one can refer to [18], [19] for further details. This tree search follows a best first search pattern until it finds a feasible assignment.We use this tree search heuristic to find feasible paths for the PISR routing problem. The algorithm is adapted to find feasible paths, such that the cycle lengths of each UA V adheres to the revisit period constraints of the tasks the UA V is assigned, and minimizes the maximum delivery time of all the tasks.①We use this tree search heuristic to find feasible paths for the PISR routing problem.②The algorithm is adapted to find feasible paths, such that the cycle lengths of each UA V adheres to the revisit period constraints of the tasks the UA V is assigned, and minimizes the maximum delivery time of all the tasks.Next, we discuss how the analysis helps identify high-impact sites that are candidates for additional hardening or redundancy, and how it also identifies low-impact sites whose loss does not dramatically reduce performance. The results in Fig. 5 show that we have a single site whose loss reduces the user coverage from 95.7 % to 92.7 %. In Fig. 6, we show the location of this site. We also show, in green and red respectively, the regions of above- and below-threshold coverage with the site lost. Comparing this figure to Fig. 4b, we observe that the negative impact is confined to the area around Congress Heights and Joint Base Anacostia-Bolling in the southeastern portion of the city. Also, the decrease in coverage is due to the relative isolation of this site; the presence of an additional site would be likely to reduce the impact of the site’s loss.①Next, we discuss how … helps identify …②The results in Fig. 5 show that …③In Fig. 6, we show the …Culling low-impact sites also reduces the ability of the network to handle large-scale incidents that require the deployment of large numbers of public safety resources. To demonstrate this, we simulated the gas leak incident use case defined by NPSTC in [7], which models the “report of a toxic gas leak in a large public assembly building near the National Mall in Washington, DC.” We considered 118 possible locations for the incident, which we show in Fig. 9. At each location, the incident occupies a 1.6 km × 1.6 km square. The incident command personnel are concentrated in a small area; the remainder of the 327 responders and 127 vehicles that compose the response force are deployed uniformly in the incident area. We simulated background traffic associated with routine public safety operations, and superimposed it on the incident traffic. We used the mix of applications and associated offered loads in Exhibit 9 on p. 26 of [8], but reduced by half, to generate the background traffic.①To demonstrate this, we simulated the …②We simulated …③We used the mix of …V.ConclusionsThe resiliency analysis technique in this paper gives network operators a useful tool to assess site deployments. The operator can identify high-impact sites whose loss has a disproportionate effect on performance. The network operator can then expend its limited resources to harden these sites or perform additional monitoring and maintenance, using some of th e best practices identified by NPSTC [1, Section 5.3]. Operators can estimate the probability of below-threshold performance with respect to multiple performance metrics, which also can inform decisions regarding which sites to harden. When examining multiple candidate deployments, an operator can identify thosedeployments that “bake in” resiliency (e.g., by having few high-impact sites) and that in turn free resources that can be expended on building additional capacity.We also showed that while low-impact sites need less hardening or redundancy than high-impact sites, removing lowimpact sites increases a deployment’s vulnerability to failures. An excessive reduction in the number of sites can leave a network vulnerable to even a single failure, if a high-impact site is lost. In addition, the overall network capacity will be decreased by removing sites, which reduces the network’s ability to handle the additional traffic associated with largescale incidents.①… in this paper gives …①We also showed that …3.Specialized termsresiliency 弹性base station deployment 基站部署threshold 阈值network operator 网络运营商antenna gain pattern 天线增益模式transmit power 发射功率signal strength 信号强度Monte Carlo simulator 蒙特卡洛仿真器uplink/downlink coverage 上行/下行覆盖load 负载throughput 吞吐量small deployments 小型部署heuristic algorithm 启发式算法iteration 迭代reliability 可靠性pruning algorithm 剪枝算法4.My reviewThis assignment requires us to find a high-quality scientific paper written by a native English speaker for the sake of their professional academic writing skills and mindset. Through this assignment, I’ve learned a lot of useful sentence patterns in an academic writing. Meanwhile, my skill of finding required academic paper is enhanced. I’m glad to gain such a big improvement, which will certainly enhance my writing ability and Make my research easier, from this English class.。

  1. 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
  2. 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
  3. 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。

PerformanceAnalysisofVirtualisedHeadNodesUtilisingCost-EffectiveNetworkAttachedStorage

A.P.Gerdelan,M.J.Johnson,andC.H.MessomInstituteofInformationandMathematicalSciencesMasseyUniversity,Albany,NewZealandEmail:{A.Gerdelan|M.J.Johnson|C.H.Messom}@massey.ac.nz

InmodernsystemslocaldiskI/OhassignificantlylowerbandwidththannetworkI/O.ThishasleadtothedevelopmentofStorageAreaNetworks(SANs)oftenimplementedwithexpensive,highbandwidthswitchingfabrics.WiththedevelopmentofInfinibandand10GigabitEthernetswitchingfabrics,highbandwidthI/Oisbecomingcommoditised,butasyetisnotcosteffectiveascomparedtoGigabitEthernetsolutions.ThispaperanalysestheperformanceofiSCSIandAoEbasednetworkattachedstorageovermultipleGigabitEthernetchannels.AfterthisanalysisweconsiderI/OperformanceofavirtualmachineaccessingstorageontheSAN.WeshowthatI/OperformanceissignificantlydegradedinavirtualmachinetothepointthatvirtualisationshouldbereservedforCPU-intensivejobsratherthanI/Ointensiveones.

Keywords:StorageAreaNetwork,AoE,iSCSI,Clustering,HighPerformanceComputing,

RAID,VirtualMachines.

1IntroductionCommodityhardwarehascontributedsignificantlyinthehistoryofgridcomputingsystems,fromtheoriginalBeowulfclustersbasedoncommodityprocessingandnetworkingtomorerecentlythehighthroughputcomputingbasedonworkstationanddesktopresourcesthathaveenabledgrid-basedscience.Theadventoffaster,higherbandwidthaccesstoremotestorageaswellashardwaresupportformachinevirtualisationhasallowednovelapproachestoGRIDcomputingtobeadopted.Theseincludetheuseofvirtualmachinestohouseworkloadsthataccessdataonaremotefileserver.Thevirtualmachinehelpsisolatethehostmachinefrompotentiallymaliciousapplicationswhiletheremotefileaccessisolatesthelocalfilesystem.Filesystemvirtualisationalsoallowsworkloadanddatatobemigratedindependentlyallowingoptimalconfigurationstobeconstructed[1].Highthroughputcomputingapplications(basedonCondoretc.)makeuseofcommoditycomponentsformingadhocclustersofworkstationanddesktopmachines.ModerncorporateandeducationalnetworksmakeuseofGigabitEthernetasthedefactotransport.Commodityserver-basedarraysofdiskscanprovidecosteffectivenetworkedattachedstorage(NAS)intheseenvironments.Thekeybottleneckinamulti-userenvironmentisthesinglegigabitEthernetlinkfromtheswitchtotheNAS.Toovercomethisbottleneck,thispaperinvestigatesthealternativestrategiesandprotocolsthatcanbeadoptedtosignificantlyincreasethethroughputbandwidthtothediskarray.Oneareathatpresentsaperformancebottlenecktogrid-enabledclustersisstorageread/writeaccessoveraStorageAreaNetwork(SAN).OurHPCusersdemandhighbandwidthstorageandweneedtoprovidethehighestpossibleread/writeaccesstolow-costsharedstorage.OtherresearchershaveidentifiedtheuseofCPUfornetworkpacketprocessingasthecauseofanetworkingbottleneck,andpresentnovelmeansfordedicatingtheuseofoneprocessoronmulti-coremachinesforthistask[2].Anareathathasnotbeenthoroughlyexamined,however,isthecomparativeefficiencyandeffectivenessofthevariousstorageexportprotocols.Ananalysisoftheoverallthroughputefficiencyandeffectiveness,andalsoefficiencyintermsoftheCPUresourceconsumptionfootprintofthesestorageaccessprotocolswouldpresentthearchitectsofhigh-performanceclustersandgridcomputingnetworkswithavaluabletemplateforfurthersystemoptimisation.Wehavethusinvestigatedand,inthispapercomparevarioussoftwaretechnologiesfortheexportofstoragefromaSAN:

•iSCSI•AoEiSCSIimplementstheSmallComputerSystemsInterface(SCSI)protocoloveraTCP/IPnet-workandhasbecomethedefactostandardforsoftwareaccesstodiskresources.iSCSIisavailableasacost-free,open-sourceimplementationthatwillrunonmostsystems,inmanycaseseliminatingtheneedforexpensivehardwarecontrollers[3],[4].AoEisasimilarapproachwhichimplementstheAdvancedTechnologyAttachment(ATA)protocoloveranEthernetnetwork.AoEwasdevelopedbyCoraidLtd,aproducerofstoragehardwaresolutionsandbladedevices,andisavailableforimplementationonsystemsotherthanthoseproducedbyCoraid.NotethatwewillbeanalysingtheperformanceofthesoftwareAoEimplementationwithvirtualblades,andnottheproprietaryhardwareblades.BothAoEandiSCSIcanexportstorageblockdevicesoverEthernet,whichareencapsulatedbytheprotocoltomimicphysicalharddisksonthemachinemountingthestorage;SCSIdisksinthecaseofiSCSIandATAdisksinthecaseofAoE.Themachineexportingthestorageiscalledthetargetandmachineremotelyaccessingthestorageistheinitiator.TheinitiatorcantreatremotestoragedevicesasiftheywerelocalattacheddisksandcreatefilesystemsorcombinethemusingsoftwareRAID[5].IthasbeenpreviouslyascertainedthattheperformanceofiSCSIdependsonthebalanceofvariouscomponentialsystemresources[6]inordertoaccommodatedifferenttypesofnetworktraffic.Wehavethereforecollectedperformanceresultsforarangeofdifferentnetworkoperationsincludingcharacterandblockoperations,randomseeks,andCPUutilisationperunitthroughput,inordertodetermineifoursystemconfigurationhasadifferenteffectonthesedifferentpropertiesofperformancewithiSCSIthatitdoeswiththeAoEprotocol.WedidnotconsiderNFSinourcomparisonsinceotherstudieshaveshownthatiSCSI-basedfileaccessprovideshigherperformancethanNFS-basedfileaccess[7],[8],andcurrentimplementationsofNFSscalepoorlyanditthusprovidespoorperformancewhenusedconcurrentlybymanyclients.ForconcurrentaccesstosharedstorageaclusterfilesystemsuchasGFS,OCFSorLustremustbeused.ForsimplicityhoweverwetestedusingXFSandasingleinitiator.Forourtestsweusedbonnie++tomeasurefilesystemperformanceontheinitiator,withproblemsizesgreaterthantwicethetotalsizeofavailableRAMtoavoidcachingeffects,andalsousednetperftotesttheavailablebandwidth.Wehavealsousedcustomwrittenscriptstofindrawblocktransferrates.Eachofourtestscomprises5runs,whicharerepeatableasasetof4resultswithin10%,excluding1spurriousoutlyingresult.Virtualmachinesarebeingincreasinglyadopted,particularlyingridenvironments,asclus-tergateways,networkservers,andstorageheadnodes,whereseparatephysicalmachineswouldnormallybedelegatedfortheserˆoles.Thisapproachcandrasticallyreducethecostsofbuildingacluster,allowingseveralvirtualmachinestobehostedsimultaneouslyononephysicalservermachine;arealisticscenariowiththegrowingavailabilityofmulti-coreservermachines.Xenvirtualmachinesarereportedtooperateat90-100%ofthehostmachine’sCPUcapacity,undervarioustestconditions[9].Whilstthevirtualmachineapproachisbecomingwidelyadopted,

相关文档
最新文档