Current Issue

• 2014, No. 07　Published:25 July 2014 Last issue    Next issue
  Columns in this issue: 论文

• 论文

  A novel task load balancing algorithm in the
  large-scale CFD with multi-zone structured grids             

  TANG Bo，WANG Yongxian

  2014, 36(07): 1213-1220. doi:

  Abstract ( 144 )   PDF (1118KB) ( 226 )   　　

  Aiming at the weakness of low fitness, poor scalability, and inaccurate communication overhead measurement in traditional parallel Computational Fluid Dynamics (CFD) applications, a new algorithm for the task load balancing in the large-scale CFD with multizone structured grids is proposed, which implements balancing the task load in parallel CFD applications by employing a three-phase method containing zone splitting, mapping zones to computation tasks, and adaptive adjustment of intra-task. Experimental results show that the proposed algorithm has better performance than the traditional greedy strategy on both homogeneous and heterogeneous computational platforms. By using the new algorithm, the performance of large-scale parallel CFD applications can be greatly improved.

  Research and implementation of elasticity
  supporting mechanism for distributed applications     

  JIA Xianglong，WU Gang

  2014, 36(07): 1221-1225. doi:

  Abstract ( 105 )   PDF (568KB) ( 185 )   　　

  One significant feature of cloud computing is to offer and utilize resources on demand. It is essential for distributed applications deployed on cloud computing platforms to have the capability of dynamic scaling. The paper researches on the premise of elasticity of distributed applications and approaches to dynamic scaling, and proposes a methodology whose essence is to discover the appropriate pattern for a certain class of applications. The elasticity supporting mechanism is designed and implemented for the pattern. This methodology is applied in Web 2.0 application case, and the corresponding pattern and elasticity supporting mechanism are proposed.

  A RAIDoriented optimization for SSD design               

  CHEN Bo, XIAO Nong, LIU Fang, OU Yang, HE Wanhui

  2014, 36(07): 1226-1230. doi:

  Abstract ( 122 )   PDF (583KB) ( 108 )   　　

  With the era of big data, SSDs (Solid State Disks) have been deployed in many largescale datacenters. As the most popular RAID technology, RAID5 has been used in SSD arrays in order to enhance the reliability. However, the parity information of RAID5 is frequently accessed, especially on random accesses dominated workloads. Frequent updating parity information degrades the performance as well as reduces the lifetime of SSD arrays. To solve this problem, the ParityAware SSD controller design is proposed. The ParityAware SSD controller gets the logical address of parity from RAID5 controller, and employs a cache, called Pcache, to store the updated parity. Data and parity are stored separately in the SSD. Experiment shows that, our method effectively reduces write operations of parity on SSD and the erase count, thus enhancing the performance and lifetime of SSD arrays significantly.

  论文

  Fluid simulation method based on CPU-GPU hybrid acceleration                

  HU Pengfei,YUAN Zhiyong,LIAO Xiangyun,ZHENG Qi,CHEN Erhu

  2014, 36(07): 1231-1237. doi:

  Abstract ( 174 )   PDF (854KB) ( 185 )   　　

  Fluid simulation based on the Smoothed Particle Hydrodynamics (SPH) plays an important role in the virtual reality, but it requires a lot of computing resources. The general methods are difficult to achieve the realtime requirement of fluid simulation based on SPH. The simulation of fluid consists of physical computing, collision detection and rendering and so on. The parallel computing based on GPU can speed up the computing and collision of particles and simulate the motion of fluid in real-time. In order to satisfy the realistic and real-time requirements, a novel fluid simulation method based on CPU-GPU hybrid acceleration is proposed, which consists of computing and rendering. The computing part of fluid simulation is accelerated by GPU, and the rendering part is accelerated by OpenMP running on CPU. The experiments show that the proposed hybrid acceleration method can significantly reduce the computing time in a fluid time step and complete rendering tasks more quickly.

  WAPFTL:A workload adaptive page-level
  flash translation layer with prediction        

  XIE Xuchao,SONG Zhenlong,LI Qiong,WEI Dengping,FANG Jian,XIAO Liquan

  2014, 36(07): 1238-1243. doi:

  Abstract ( 181 )   PDF (894KB) ( 181 )   　　

  NAND Flash based Solid State Drive (SSD) has been utilized in enterprise servers and high performance computing environments due to its various advantages, such as fast access speed, low power consumption, high reliability, etc. Since SSD suffers from poor performance of random writes and limited lifespan, a workload adaptive pagelevel address mapping algorithm, named WAPFTL, is proposed, which can predict the characteristics of incoming requests and dynamically adapt the cache strategy of address mapping information. Experimental results show that WAPFTL can improve the cache hit ratio and reduce the number of expensive extra read/write operations by jointly exploiting the temporal and spatial localities of workloads efficiently. Besides, WAPFTL can also reduce the number of erase operations, thus improving the overall performance of SSD.

   

  Enterprise investment cost monitoring
  algorithm based on cloud computing      

  XU Jie1,YAO Rui1,2

  2014, 36(07): 1244-1249. doi:

  Abstract ( 90 )   PDF (561KB) ( 165 )   　　

  Nowadays it has been appearing that more and more domestic enterprises have constructed cloud computing blindly but been lacking seriously in monitoring at the course of deploying. As for the phenomenon, an investment cost monitoring algorithm based on cloud computing is proposed to solve it. The proposed algorithm analyzes and defines the types of constructing cost, counts the cumulative values of each cost, studies intensively the ROI and the fronttoback constructing cost ratio which enterprise cares, and finally finds out the optimum alert points of all types of constructing cost. Experiments show that the proposed algorithm can make enterprises get cost alerts effectively and efficiently, and provide the powerful science evidence for enterprises to take remedial measures.

  A flexible threshold proxy re-signature
  scheme with provable security           

  YANG Xiaodong,ZHANG Lei,WANG Caifen

  2014, 36(07): 1250-1254. doi:

  Abstract ( 104 )   PDF (384KB) ( 152 )   　　

  The most existing threshold proxy resignature schemes have one threshold value. In many practical applications, the number of proxies is often dependent on the significance of the message to be resigned, which requires the threshold value to be changeable. Based on Ateniese G’s proxy resignature scheme Sbi  and the Chinese remainder theorem, a flexible threshold proxy resignature scheme is proposed, and its security is proved. According to the changeable threshold value, each proxy can noninteractively produce the resignature private key and the verification public key. Compared with other existing schemes, this scheme can provide better efficiency in terms of the communication and computation cost.

  Nodes deployment approach of Internet
  of Things for monitoring applications            

  YANG Bin,HAO Yangyang,LI Junjun

  2014, 36(07): 1255-1261. doi:

  Abstract ( 102 )   PDF (1028KB) ( 164 )   　　

  To solve the conflict between higher monitoring quality and lower costs of applying Internet of Things (IoT) monitoring system, a mixed integer nonlinear programming model is built up. Various key factors, such as network scale, cost, representation level, and deployment of nodes for balance degree of IoT, is taken into consideration in this programming model comprehensively. Furthermore, the quantity and deployment solutions are solved through the genetic algorithm, also the selection of nodes is evaluated. Simulation results show that it is conducive to solve the qualitycost conflict to some extent. It is also of theoretical significance for the IoT research and design of monitoring system.

  FCM-DAFSA based multi-target tracking node task
  allocation method for wireless sensor network               

  WANG Yanchun1，SHANG Xiaoli2，LI Hui1

  2014, 36(07): 1262-1267. doi:

  Abstract ( 166 )   PDF (615KB) ( 161 )   　　

  Multitarget tracking is one of the important applications of wireless sensor networks. A DAFSA (Discrete Artificial Fish Swarm Algorithm) based multitarget tracking node task allocation method for wireless sensor network is proposed. Firstly, the class distance threshold fuzzy Cmeans clustering algorithm is used to estimate the number of potential targets and their locations in the monitoring region. Secondly, according to the objective function of task allocation, an improved DAFSA is used to optimize the objective function so as to get the task distribution and is compared with other algorithms. Simulation results show that the proposed algorithm has lower energy consumption and less task allocation time than the nearest neighbor method, MEM method, and particle swarm optimization algorithm. Therefore, it is concluded that the proposed algorithm can effectively improve the overall performance of the wireless sensor networks and meet the needs of practical application.

  A clustering algorithm for
  mobile peer-to-peer network          

  YANG Zhongyi1,2,ZUO Ke1

  2014, 36(07): 1268-1274. doi:

  Abstract ( 122 )   PDF (835KB) ( 147 )   　　

  Using the clustering algorithm to reduce the network churn effect and extend the network lifetime is one of the research emphasis points of mobile peer-to-peer networks. Based on studying the Kautz graph and its characteristics, a clustering algorithm of mobile peer-to-peer network based on the Kautz graph is proposed. In the algorithm, an address space tree is defined firstly, and then the Kautz string is used as nodes’ identifier. The breadth-first-search via post-order is used to travel a well-defined address tree for clusters creation. Besides, the mechanism for management and maintenance of cluster structure is designed to ensure the structural integrity. Theoretical proof and experimental evaluation show that the clustering algorithm can effectively reduce the churn effect and extend the network lifetime.

  A novel energy-saving routing algorithm in WSN          

  LI Ping，DAI Jin

  2014, 36(07): 1275-1278. doi:

  Abstract ( 104 )   PDF (504KB) ( 173 )   　　

  The problem of unbalanced energy consumption and transmission delay in Wireless Sensor Network (WSN) is studied, and a novel routing algorithm is proposed. By establishing the minimum hop count and protecting the remaining energy of the nodes, the algorithm makes packets traverse to Sink nodes through the paths with optimized energy. Simulation is carried out in MATLAB and it proves that the algorithm can balance energy consumption and increase the network lifetime.

  Prediction and analysis of the Internet access
  population based on gray Markov Verhulst model            

  ZHAO Ling1，2,XU Hongke2

  2014, 36(07): 1279-1283. doi:

  Abstract ( 202 )   PDF (440KB) ( 157 )   　　

  In order to predict the Internet access population accurately, a forecasting method based on Gray Markov Verhulst model is proposed. The method uses historical data to construct gray Verhulst model, gets the expression of time response series of the Internet access population by determining coefficients, and obtains the development sequences of the Internet access population in the near future. Based on the Markov chain, the sequence states are divided into three parts, the state probability and medium prediction value are obtained by determining the state transition matrix, and further the modification values of each sequence are obtained. Finally, the Internet access population from December 2006 to June 2012 is used as original data to establish the forecasting model so as to predict the number of Internet users from December 2012 to June 2014. The results show that the prediction accuracy of the gray Markov Verhulst forecasting model has fewer errors and better prediction precision, gives the fluctuation range and the probability of the prediction results, and provides the decisionmaking basis for network construction and management.                                               

  A novel key exchange protocol
  for frequent communication         

  YI Tong1,LI Xuebao2,CHEN Hongchao1

  2014, 36(07): 1284-1289. doi:

  Abstract ( 116 )   PDF (684KB) ( 190 )   　　

  Firstly, an efficient verifierbased threeparty passwordauthenticated key exchange protocol previously proposed is analyzed. The protocol is vulnerable to security threats such as server key disclosure attack and so on, and has a lack of forward secrecy. Secondly, on the basis of the analysis, in order to solve the problem that most of existing verifierbased 3PAKE protocols cannot resist server key disclosure attack, a novel 3PAKE protocol is proposed. Through security analysis, the new protocol can be proved to be more secure than the old one, and has the ability to resist all known attacks. In addition, compared with existing protocols, it is more efficient.

  A bluetooth node wake-up scheduling
  mechanism for opportunistic networks      

  YE Hui1,PAN Yi1,HE Wende1,PENG Shaoliang2

  2014, 36(07): 1290-1295. doi:

  Abstract ( 103 )   PDF (741KB) ( 163 )   　　

  Mobile bluetooth devices have limited battery energy in opportunistic networks whose wireless devices are carried by humans. Moreover, nodes are often in the long disconnected state in opportunistic networks. Therefore, it is an important issue that how to design an effective node wakeup scheduling pattern so as to reduce energy consumption and keep the existing network connectivity. A bluetooth node wakeup scheduling policy, named BWM, is proposed. This policy studies the power consumption problem in bluetooth devices and constructs the bluetooth energy cost of the data transmit model. Furthermore, BWM studies the relationship of key parameters in wakeup scheduling mechanism and controls the wakeup period intervals in order to ensure the effective data throughput of nodes. The simulation results show that our policy can reduce the power consumption of nodes under the premise of effective data delivery ratio.

  Design of the real-time bus position
  inquiry system on smart phone            

  ZHANG Kai,CHEN Feng,DU Jing

  2014, 36(07): 1296-1300. doi:

  Abstract ( 98 )   PDF (681KB) ( 244 )   　　

  In daily life, real-time bus position information cannot be accurately acquired, which brings a lot of inconvenience. In order to overcome this problem, the application of mobile internet in the intelligent transportation is discussed, the Baidu map API technique is used to design a realtime bus position inquiry system based on mobile client, and the offset correction method of longitude and latitude data in the electronic map is given. People can use Smart phone or PDA to enter this app to keep track of the location or number information of the bus. This web app is proved to be feasible under test.

  Software quality evaluation method based on
  fuzzy neural network with fuzzy triangle numbers         

  LI Kewen,ZHANG Yu,MA Jingfeng,LIU Hongtai

  2014, 36(07): 1301-1306. doi:

  Abstract ( 117 )   PDF (586KB) ( 144 )   　　

  The results of software quality evaluation are closely related to users’ experience, but current software quality evaluation methods lack attention to user requirements because of the abstraction and complexity of software products and the fuzziness of customer requirements, thus ignoring the importance of user requirements on software quality evaluation. In consideration of the user requirements’ influence on the quality of software, a fuzzy neural network based on fuzzy numbers is proposed in order to imitate the nonlinear function between customer needs and software characteristics considering the influence of the user requirements on software quality, which can meet the complexity characteristics of software products, and make the results of software quality evaluation more objective and comprehensive. Results indicate that the method can more accurately imitate the nonlinear function between customer needs and software characteristics and is an effective measurement for studying software quality evaluation.

  Study of the test instances sets selection
  in project scheduling problems          

  TIAN Wendi1,XU Jing1,BIE Li2,CUI Nanfang3

  2014, 36(07): 1307-1315. doi:

  Abstract ( 108 )   PDF (1472KB) ( 150 )   　　

  In order to test and compare the performance of algorithms in the project scheduling problems, test instances sets are required. Some literatures are reviewed and surveyed on the instances sets. As internationally and commonly used, two basic instances sets (Patterson sets and PSPLIB) and two instances sets generators (the single project generator RanGen and the multiproject generator RCMPSP) are introduced. Finally, the selection flow of test instances sets in project scheduling problems and the general method of constructing test instances sets are proposed. Two cases are used to illustrate the effectiveness and application prospects of these methods.     

  A multi-user detector based on the binary artificial
  bee colony algorithm with estimation of distribution              

  LIU Ting1,ZHANG Liyi1,2,ZHANG Jinbin3

  2014, 36(07): 1316-1323. doi:

  Abstract ( 117 )   PDF (795KB) ( 135 )   　　

  In order to enhance the global exploration ability of the binary artificial bee colony algorithm, a binary artificial bee colony algorithm based on estimation of distribution is proposed. The proposed algorithm is applied on the optimal multiuser detection technique, and the multiuser detection scheme based on the algorithm is designed. The algorithm directly adopts the multidimensional neighborhood search strategy in the discrete domain, thus quickening the convergence speed and avoiding the conversion from continuous domain to discrete domain. Meanwhile, it makes use of the global statistics information obtained by estimation of distribution to generate the candidate solutions. The simulation results show that, compared with the conventional detectors, the convergence speed of the designed detector is faster, the bit error rate and the nearfar resistance ability is significantly improved.

  A benefit based collision avoidance
  strategy for multi-robot systems           

  YAO Zhifeng1,2,YE Xiufen2,DAI Xuefeng1,ZHU Ling1,SUN Ming1

  2014, 36(07): 1324-1329. doi:

  Abstract ( 117 )   PDF (854KB) ( 203 )   　　

  The collision among robots often occurs when a multirobot system performs the SLAM operation. However, the collision avoidance is different from the general obstacle avoidance, because the obstacles are generally not moving in the obstacle avoidance problem. In order to solve the collision avoidance problem among robots, a benefit based collision avoidance strategy for multirobot systems is proposed. The strategy mainly focuses on improving the efficiency of multirobot systems, and determines the order when the robots pass the intersection. Besides, the dynamic collision avoidance is considered and the algorithm of determining the order of robots’ passing the intersection is given. Finally, a simulation example of robots’ avoiding collision is given in order to illustrate the process of the strategy and the simulation results are analyzed. And the reasonable assumption is given for the spatial relationship between the robots and targets.

  A multi-factor grey model based on SVM residual
  error compensation for telephone traffic prediction         

  GUO Qin1,JIA Zhenhong1,QIN Xizhong1,SHENG Lei2,CHEN Li2

  2014, 36(07): 1330-1335. doi:

  Abstract ( 106 )   PDF (681KB) ( 144 )   　　

  In order to improve the prediction accuracy of telephone traffic and the speed of modeling, aiming at that the current mobile telephone traffic forecasting is influenced by many factors, a multifactor grey traffic prediction model based on Support Vector Machine (SVM) residual error compensation is proposed. In this model, the main factors affecting the traffic variables are determined by the grey correlation analysis method, and the multivariable grey model is used to do prediction. Then, the residual error sequence is predicted by the leastsquares SVM optimized by the particle swarm optimization so as to realize the residual error compensation. The experimental results show that the required samples of this prediction model are small, the model has the advantage of high accuracy, and the model provides a new forecasting tool for the telephone traffic network management.