Current Issue

• 2012, No. 2　Published:25 February 2012 Last issue    Next issue
  Columns in this issue: 论文

• 论文

  Scheduling under the SINR Model in Ad Hoc Networks with Successive Interference Cancellation

  Lv Shaohe,WANG Xiaodong,ZHOU Xingming

  2012, 34(2): 1-8. doi:

  Abstract ( 372 )   PDF (639KB) ( 408 )   　　

  The capacity of modern wireless communication systems is limited by interference. Successive Interference Cancellation (SIC) is an effective way of multipacket reception to combat interference at the physical layer. This paper focuses on link scheduling under the SINR (Signal to Interference Noise Ratio) model in an Ad Hoc network with SIC. The facts that interference is accumulated and that the links decoded sequentially by SIC are bring about key technical challenges. To characterize the accumulative effect, for a given link,  conflict set is defined as a set of links that can interfere with the detection at the link. And then, we propose a conflict set graph (CSG) to characterize the interference and define interference degree to measure the link interference. As scheduling over CSG is NPhard, an independentsetbased greedy scheme is explored to efficiently construct a maximal feasible schedule. The performance is evaluated by simulations. As compared to the simple greedy method ［1］, the throughput gain is on average 30% and up to 60%.

  Attack Graph Optimization Based on Attack Distance

  HE Zhiqiang，LOU Fang，LI Liang

  2012, 34(2): 9-12. doi:

  Abstract ( 342 )   PDF (655KB) ( 362 )   　　

  Traditional attack graph generation methods have the problem of state explosion with the network scale expansion. Network security administrators are often overwhelmed by holding redundant attack graphs. The complex attack graph is optimized by using attack distance, which can eliminate unnecessary attack paths and choose the optimal attack paths. The administrators can refer to the optimized attack graph and get the defense information. The experimental results show that it retains the most likely attack path and simplifies the attack graph. With the expansion of network scale, the effect has become increasingly evident.

  Research on an Intrusion Detection Method Based on Rough Sets

  SHI Zhicai，XIA Yongxiang

  2012, 34(2): 13-18. doi:

  Abstract ( 314 )   PDF (568KB) ( 438 )   　　

  In order to improve the performance of intrusion detection systems, the initial data are usually preprocessed by feature extraction so as to reduce the payload of the system and increase its detection speed. At first the rough set theory is used to give a formal description to the intrusion detection systems. Information entropy is applied to the discretization of continuous numerical attributes. Attribute features for intrusion detection are extracted by knowledge reduction. Information gain is used to control the reduction procedure of attribute features. The redundant features are eliminated effectively. The processing payload of the system is reduced and its detection effect is improved. The experiments justify that the proposed method makes the training time of the system to typical attacks for DoS and PROBING is reduced by 2.8 and 3.2 times. The detection speed of the system for two attacks is increased by 3.2 and 4.5 times.

  Bloom Filters Based on the Tree Structure

  CHENG Nie，HUANG Kun，SU Xin，ZHANG Dafang

  2012, 34(2): 19-24. doi:

  Abstract ( 343 )   PDF (570KB) ( 511 )   　　

  This paper presents a multilevel structure called the Treebased Bloom Filter(TBF). Multilevel structure is the hot spots of Bloom filters and related data structure research in recent years. This structure achieves multiple levels of storage and reduces the burden of onchip memory, but it also accelerates the speed of onchip search. TBF is a more efficient algorithm which is the improvement based on the drawbacks of the BloomingTree algorithm, and TBF can reduce the conditions of the space requirements and achieve the same function of CBF under the same conditions. Our experiments show that compared with the BloomingTree algorithm, the TBF algorithm can effectively solve the index error in the logic problem of the BloomingTree algorithm, and show more time efficiency: under the conditions of the same false positiveness  and unchanged layers, the query time improve on  an average of 13.4%; under the conditions of the fixed false positiveness and the same layer changes, the time of insertion improves on an average of 17.9%, and 12% average improve the time of deletion.

  An IdentityBased Short Signature Broadcast Authentication Protocol in WSNs

  YANG Lu，YOU Lin，YANG Minghui

  2012, 34(2): 25-30. doi:

  Abstract ( 378 )   PDF (495KB) ( 347 )   　　

  Broadcast authentication is a critical security service in sensor networks, since it allows a sender to broadcast messages to multiple nodes in a secure way. In wireless sensor networks, the broadcast authentication schemes based on message authentication code, like μTESLA and MμTESLA, have some disadvantages. Recent research results have shown that bilinear pairing based cryptography is applicable on resourceconstrained sensor nodes. This paper introduces an efficient identitybased certificateless short signature scheme. The size of signature generated by this scheme is approximately 160bit long, which is the shortest signature so far. And the computation cost is much less than that of other public key signatures. It also provides the feature that it can authenticate the broadcast messages from multibasestations. Based on the MICA2DOT platform, the energy consumptions on communications and computation are analyzed. And the protocol’s other performances are also analyzed. The protocol proposed in this paper can effectively reduce resource cost and satisfy some cardinal properties of broadcast authentication, so it adapts the characteristics of wireless sensor networks.

  An EnergyEfficient Routing Protocol in Wireless Sensor Networks

  CHENG Wenlong,YU Liang，SONG Ziyu

  2012, 34(2): 31-34. doi:

  Abstract ( 383 )   PDF (556KB) ( 306 )   　　

  Energy conservation is the primary consideration in wireless sensor networks. Extending the lifetime of the wireless sensor networks and balancing nodes’ load are the goal of designing the routing protocol in wireless sensor networks. Since the election of clusterhead nodes in the LEACH routing protocol does not take the energy into account, there are some deficiencies, such as the facts that the clusterhead nodes are distributed unevenly in space and that the energy is consumed too much during a longdistance data transmission. This paper proposes an improved version of the routing protocol LEACHZED, which employs the way of zonepartition, considers the nodes’ energy and the distance to Sink comprehensively, and transfers data in multihop among clusters. The simulation shows, LEACHZED extends the lifetime of the network effectively and reduces the energy consumption of the whole network, so it performs better than the LEACH routing protocol as a whole.

  The Adaptive Transmission and Control of MultipleConcurrent Flow in Live Streaming Media Systems

  SHEN Yilou,ZHU Yanqin

  2012, 34(2): 35-40. doi:

  Abstract ( 363 )   PDF (734KB) ( 467 )   　　

  To solve the bandwidth insufficiency of live streaming media's transmission in the Internet, and the diversity of users' access, the general transmission control technologies of streaming media in the IP networks are probed, and a realtime transmission control strategy of streaming media based on broadcast is proposed. The streaming media’s realtime transmission is realized through the multicast and RTP/RTCP agreement. In order to adapt to the dynamic change of network bandwidth, the monitoring of network conditions and realtime streaming switching control are realized by the multithread technology.

  Design and Implementation of a TicketBased  Single SignOn Protocol

  LI Fan1,2,WANG Liuyi3

  2012, 34(2): 41-44. doi:

  Abstract ( 391 )   PDF (436KB) ( 363 )   　　

  With the rapid development of the enterprise informatization construction, the enterprise information applications are built in increasing numbers. It is an inevitable trend to establish a unified identity management system to provide single signon among the enterprise applications. The user is able to access different enterprise applications securely and smoothly by providing his or her identity information only once in enterprise identity authentication center. In this paper, a ticketbased single signon protocol and the design of a protocol reference implementation are proposed. The new protocol improves the limitation of the classical ticketbased single signon protocol such as Kerberos. It is easier and safer to implement single signon for enterprise applications with a lot of legacy accounts.

  An Improved Algorithm for the Centroid Localization of Wireless Sensor Network

  HU Yongmei,ZHANG Huan

  2012, 34(2): 45-49. doi:

  Abstract ( 385 )   PDF (540KB) ( 469 )   　　

  In wireless sensor networks, it is critical for monitoring activities to determine the node location or the location of an incident. The exact location of a node is not only the premise of providing monitoring events or monitoring the target location information, but also the basis of providing network topology configuration, improving routing efficiency, reporting network coverage quality, providing namespace for the network and other network functionalities. Therefore, in the paper, a centroid location algorithm for wireless sensor networks’ positioning is improved. For the general location of the unknown nodes, the  algorithm has new amendments. The weighting factors in the unknown node location determining algorithm are optimized. These improvements make the unknown node positioning error and positioning precision more accurate. Compared with the previous weighted centroid location algorithm, the  simulation results in this paper show that the improved centroid location algorithm has greatly improved in positioning error and positioning precision.

  A DICacheBased Hybrid Threaded Interpretation Technique

  CHEN Wei,WANG Zhiying,CHEN Xuhao,SHEN Li,LU Hongyi,XIAO Nong

  2012, 34(2): 50-55. doi:

  Abstract ( 364 )   PDF (1501KB) ( 339 )   　　

  Interpretationbased instruction set emulation provides a solution to the binary compatibility problem. The organization of the interpretation process has great impact on the performance of an interpreter. Centralized interpretation is inefficient while the traditional threaded interpretation is not suitable for interpreting the CISC Instruction Set Architecture (ISA) because of the complicated instruction decoding process. We propose a DICachebased hybrid interpretation technique. DICache can dynamically predecode the source instruction and convert them into an intermediate form. The DICache access code is appended at the end of each interpreter routine, which enables the threaded interpretation for CISC ISA. We implement an interpreter which interprets the IA32 instructions on VLIW. We conduct some benchmarks from SPEC INT 2000 to evaluate the performance of the novel threaded interpretation method. It is demonstrated that DICachebased hybrid threaded interpretation can significantly improve the performance of an interpreter.

  Design and Optimization of a FaultTolerant Deflection Router for NetworksonChip

  FENG Chaochao,ZHANG Minxuan,JIANG Jiang,LI Jinwen

  2012, 34(2): 56-61. doi:

  Abstract ( 410 )   PDF (927KB) ( 375 )   　　

  Reliability has become a key issue of NetworksonChip (NoC) as the technology scales down to the nanoscale domain. In this paper, we design and implement a faulttolerant deflection router based on reinforcement learning for NoC. The router reconfigures the routing table through a reinforcement learning method during packet transmission to achieve faulttolerance. An optimized router with 2 pipeline stages is also implemented to improve the performance of the router. The synthesized results under the TSMC 65〖WTBX〗nm〖WTBZ〗 technology show that the router with 2 pipeline stages can achieve the frequency of 750MHz, which is almost 1x more than that of the original router, while the area only increases by 22%. The simulation results under the synthetic workloads demonstrate that the average network latency of the router with 2 pipeline stages is less than that of the router with no pipeline.

  1.25GHz CMOS PLL With the Duty Optimizing Technique

  MA Zhuo,GUO Yang,XIE Lunguo

  2012, 34(2): 62-66. doi:

  Abstract ( 402 )   PDF (1046KB) ( 385 )   　　

  In highspeed SerDes with the half rate structure, the duty of the clock is seriously important, which is the decisive factor for unit intervals. In this article, a 1.25GHz ring oscillator PLL is established on the 0.13μm CMOS process, in which a duty balance circuit is integrated. The result of testing shows the stable output clock is 1.25GHz, and the duty is within the range of 49.86～51.21%, and the mean duty is 51.21%.

  Research on a MYGCCBased Algorithmfor Checking Programming Rules

  Li Feng,WEN Yanjun,QI Zhichang,LU Saiyin

  2012, 34(2): 67-72. doi:

  Abstract ( 330 )   PDF (758KB) ( 309 )   　　

  MYGCC is a tool for checking programming rules, and its current checking algorithm has defects. That is, it can not show the whole control flow paths that violate the checking rules. To eliminate the defects, this paper presents an improved algorithm to the original one of MYGCC. Experiment shows that the algorithm is effective. This improvement can help users to find the position of bugs more accurately, and benefit bug fixing.

  Research on the Expanding Intermediate Code for Control Flow Checking Based on GCC

  HE Tao1,ZHOU Huiping2,JIA Lili2,WANG Fahong1

  2012, 34(2): 73-77. doi:

  Abstract ( 344 )   PDF (557KB) ( 334 )   　　

  The paper introduces  for CFCSS, briefly analyzes the running routine of the GCC compiler, gives the concrete methods for how to extend the CFCSS algorithm in the GCC compiler and then compiles a single C language program in the extended GCC. Finally, we perform the CFCSS algorithm’s valid verification by fault injection experiments. The experiment shows the new program can check the error of the control flow when it runs, which is compiled by the GCC compiler after extending the CFCSS algorithm. The above conclusion gives great support to positioning and recovering the faults in the next step and lays a good foundation for solving the computer running faults in the space.

  VMSF—A Virtual Machine Scheduling Framework for Hosted VMMs

  LIU XiaoJian,DAI Huadong,YAN Yuejin

  2012, 34(2): 78-81. doi:

  Abstract ( 402 )   PDF (576KB) ( 310 )   　　

  Virtial machine is a new kind of application domain, which has many different scheduling requirements from the traditional software applications. It is the Virtual Machine Monitors’ duty to satisfy the running virtual machines’ QOS requirements. However, in hosted VMMs, the function of virtual machine scheduling is delegated to host operating systems, which can not handle the new situation well. In order to facilitate the adoption of more suitable scheduling algorithms, a new scheduling framework, VMSF, is introduced in this article. VMSF makes no modifications to the host operating systems’ scheduling mechanism, and at the same time, permits the adoption of the third party’s virtual machine schedulers. Finally, experiments are conducted on KVMs to verify the effectiveness of the framework.

   Simulating System for Scheduling CoAllocation Tasks in Distributed Computing Environments

  LI Bo1，ZHOU Enwei1,SHEN Bin2

  2012, 34(2): 82-86. doi:

  Abstract ( 417 )   PDF (438KB) ( 391 )   　　

  Resource coallocation is one of the crucial technologies affecting the utility and quality of services of largescale Gridlike distributed environments by simultaneously allocating the requested multiple resources to a single submitted application to meet the specific performance requirements. This paper presents a discreteevent driven simulator for studing the coallocationrelated resource management and scheduling algorithms. This tool models the main Grid components and their functions for coallocating grid jobs, including Grid user，metascheduler, coallocator, and coreservator. Some main scheduling algorithms and policies, including First Come First Served, First Processor First Served and Backfilling, are  implemented.

  HighPerformance Computing Environments for Ensemble Prediction

  LIU Cancan,ZHANG Weimin,LUO Zhigang,REN Kaijun

  2012, 34(2): 87-92. doi:

  Abstract ( 385 )   PDF (816KB) ( 365 )   　　

  There is an urgent demand for High Performance Computing (HPC) resources to analyze and process huge data in ensemble prediction. It will improve the effect of the prediction result to provide effective resources and data management. With the huge data and high demand for HPCs, we provide a data management strategy based on meta data distillation and a resource management method based on virtual organization. Besides, the grid technology is used for data and resource sharing and management. As a result, a coordinated development environment is built for the meteorologists distributed all over the world and different organizations, which accelerates the development of meteorology as a result.

  Reconstruction and Improvement of the Radiation Hydrodynamics Code RH2D

  REN Jian1,2，SHEN Weidong1

  2012, 34(2): 93-98. doi:

  Abstract ( 343 )   PDF (645KB) ( 538 )   　　

  In scientific computing and industrial applications, some programs integrated with physical models and numerical methods and verified by theories and the experiment in the past long time, have much more confidence in numerical simulation. But these programs are not suitable for the research requirements in numerical accuracy and computing scales at present. In this paper, the code RH2D have been reconstructed based on JASMIN, and the radiation heat conduction numerical method has been improved from the splitting method to the nonsplitting method. The numerical results for practical problems are presented to show that the nonsplitting method has smaller energy conservation errors and much higher parallel efficiency than the splitting method.

  A Dynamic Updating Mechanism of Devicesin Pervasive Computing Environments

  XIAO Changbo，WU Gang

  2012, 34(2): 99-103. doi:

  Abstract ( 322 )   PDF (737KB) ( 339 )   　　

  Devices are the key of the pervasive computing applications to acquire the context changing and also take the responding actions. Since device updating happens occasionally in a running pervasive application, it is necessary to study how to minimize the cost of renewal to popularize the application.Based on our former work, which  is an adaptive and scalable contextaware architecture for pervasive computing and its prototype is implemented based on OSGi/ROSGi, this paper proposes a dynamic updating mechanism of devices to solve the problem related with the device heterogeneity. With this mechanism the developers of the application need not to be involved in device updating.

  SubCounter:A Node Subset Size Estimation Approach Based on Semantic Clustering

  ZHENG Zhong,WANG Yijie,MA Xingkong

  2012, 34(2): 104-110. doi:

  Abstract ( 296 )   PDF (482KB) ( 362 )   　　

  Many P2P applications need the size values of node subsets in the system to enhance performance. The existing subset size estimation approaches are based on applying the size estimation approach directly. This paper proposes SubCounter, a node subset size estimation approach based on semantic clustering. SubCounter maintains a semantic clustering neighbor list for each node by view exchange, so each node can keep contacts with others in the same subset. Based on the semantic clustering, SubCounter realizes the estimation of subset sizes, through antientropy aggregation. The experimental results show that compared with the existing approaches, SubCounter converges more quickly when each node belongs to many subsets simultaneously, and ensures the same precision and similar robustness with less communication and storage cost.