ISSN 1000-1239 CN 11-1777/TP

Table of Content

15 April 2005, Volume 42 Issue 4
A VLSI Architecture for PMVFAST Block-Based Motion Estimation Algorithm
Li Tiejun, Shen Chengdong, and Li Sikun
2005, 42(4):  537-543. 
Asbtract ( 441 )   HTML ( 1)   PDF (457KB) ( 541 )  
Related Articles | Metrics
A flexible, efficient and low power architecture for PMVFAST, an enhancing block-based motion estimation algorithm, is proposed in this paper. The core of the a rchitecture is a motion estimation engine, which supports independent calculatio n of SAD and several searching patterns with different sizes and types. The engi ne includes three types of delay variable units, which support arbitrary delay i n particular ranges and enable different pattern searching. The engine reuses pr ocessing elements to construct a SAD engine and enables basic support of all BMA s. Besides, the engine can reduce power consumption by gating the unused element s and reusing resources. Experimental results verify the superiority of the prop osed architecture. Its computing efficiency is about 15 times higher than the we ll-known low power FS architecture including 16 PEs and its PSNR is similiar to the FS algorithm.
Applying Aggregate I/O to Improve Performance of Network Storage
Cao Qiang and Xie Changsheng
2005, 42(4):  544-550. 
Asbtract ( 398 )   HTML ( 1)   PDF (384KB) ( 456 )  
Related Articles | Metrics
Network storage provide high scalability, availability and flexible for storage systems, and are widely applied to many fields. Small I/O seriously affects the performance of the storage system including network storage. Current several met hods want a theory model and only improve write performance. In this paper, the aggregate I/O model is presented first. Moreover, how to apply this methodology in a NAS system is also illustrated based on IP network. The aggregate I/O modul e called CFS (chunk file system) is designed at both the NAS device and the cust omer end. In the experiment, the analysis of the results shows correctness of th e model and an improvement made by aggregate I/O for network storage performance .
An Exception Handling Method Based on Grid Service Routes
Yue Qiang, Li Wei, and Xu Zhiwei
2005, 42(4):  551-556. 
Asbtract ( 342 )   HTML ( 0)   PDF (315KB) ( 362 )  
Related Articles | Metrics
A grid service is a Web service that provides a set of well-defined interfaces a nd follows specific conventions. A grid service exception may occur in a process of behaviors related to requesting, responding, or running a grid service. Grid service exceptions deal with the run-time status and contextual information of grid services. A handler of exceptions mainly holds the configuration of a group of checkpoints, trace and rollback of services. The common approach to handling an exception is using a function to check out an error and sending the error co de to the caller, which may need a multi-level and consecutive behaviors, till t he error is sent to the function that responds it. If the linkage of these funct ions is weak, the whole process would fail. In a grid computing environment, an exception handler should not only reduce the internal propagation of error codes , but also make a transferable function to send the exception to an external uni t or a grid service. In terms of the grid architecture, the exception handler of a grid service is discussed and then an A/B-Back algorithm is proposed, which i s used to handle exceptions in a grid service. Finally, the test result based on the simulated experiment in Globus Toolkit 3.0 is given.
Study of Algorithm-Oriented Mathematical Model of SIMD Computer and Its Applicat ion
Zhang Facun, Zhao Xiaohong, Wang Zhong, and Shen Xubang
2005, 42(4):  557-562. 
Asbtract ( 417 )   HTML ( 1)   PDF (349KB) ( 618 )  
Related Articles | Metrics
By the study of image processing in the field of data parallel computation, an algorithm-oriented mathematical model of SIMD computer, and a novel method which is useful for mathematical description of data parallel algorithm are proposed. By using the method proposed, detailed descriptions of data parallel algorithm o n histogram operation, region-growing image segmentation and image convolution o peration are given. This mathematical description method is simple and accurate, the communication complexity and computation complexity of the algorithm being studied can be obtained easily in the mathematical formula. The method can be us ed as mathematical description in applications of data parallel algorithm studie s.
High Productivity Computing Systems
Meng Dan, Zhang Zhihong, and Chen Mingyu
2005, 42(4):  563-569. 
Asbtract ( 456 )   HTML ( 0)   PDF (391KB) ( 454 )  
Related Articles | Metrics
The HPC world has been obsessively focused on simple performance metrics, such a s peak floating point performance, or performance on the Linpack benchmark. Howe ver, two critical problems, the unbalanced architecture and the deficient progra mming support, have never obtained a good solution. The recent DARPA High-Produc tivity Computing Systems (HPCS) initiative has raised the issue of proper produc tivity measurements, in the context of high-performance computing. The backgroun d, objective, challenge, and main efforts toward the research of HPCS are presen ted here.
N-SHOQ(D): A Nonmonotonic Extension of Description Logic SHOQ(D)
Wang Songxin, Wang Fei, Zhou Shuigeng, and Zhou Aoying
2005, 42(4):  570-575. 
Asbtract ( 616 )   HTML ( 1)   PDF (378KB) ( 568 )  
Related Articles | Metrics
Ontologies play a central role in the development of the semantic web, as they p rovide precise definitions of shared terms in web resources. One important web o ntology language is DAML+OIL. It has a formal semantics and a reasoning support through a mapping to the expressive description logic SHOQ(D) with the addition of inverse roles. One shortcoming of the SHOQ(D) is that the incomplete knowledg e frequently emerging in the practice cannot be expressed by SHOQ(D). In this pa per, a nonmonotonic extension of the SHOQ(D), called N-SHOQ(D), is presented to allow for dealing with incomplete knowledge in the semantic web. The syntax and semantic of N-SHOQ(D) are defined and its properties are studied in detail. The N-SHOQ(D) forms a formal underpinning for the nonmonotonic extension of the DAML +OIL language.
Optimal Model Selection for Support Vector Machines
Liu Xiangdong, Luo Bin, and Chen Zhaoqian
2005, 42(4):  576-581. 
Asbtract ( 609 )   HTML ( 3)   PDF (356KB) ( 792 )  
Related Articles | Metrics
Proposed in this paper is a method of model selection based on kernel alignment for support vector machines-OMSA (optimal model selection algorithm) by means o f learning on kernel matrix. This algorithm aims at finding the optimal kernel p arameters and learning model from training data without performing the standard procedures of SVM training and testing so as to overcome the flaws of convention al methods of SVM model selection. The classification experiments on the UCI dat abase and the face recognition experiments on the FERET face database are deploy ed with this algorithm and the famous LOO (leave-one-out) algorithm. The four da tasets from UCI used in experiments are diabetis, glass, waveform and wine. By comparison with the LOO algorithm, the experimental results show that the optimal kernel parameters and kernel matrix are found by OMSA with the minimal testing error of SVM classifier. Specially, the results from face recognition experiment s are satisfactory. This algorithm provides a feasible method for SVM model sele ction as well as references for other kernel-based learning algorithms.
A New Local Search Algorithm for the Job Shop Scheduling Problem
Zeng Liping and Huang Wenqi
2005, 42(4):  582-587. 
Asbtract ( 504 )   HTML ( 1)   PDF (286KB) ( 575 )  
Related Articles | Metrics
A new local search algorithm with hybrid neighborhood for solving the minimum make span problem of job shop scheduling is presented. A new dispatching rule base d on the frontier-greed method is proposed to generate initial solution. A new c oncept of neighborhood structure involving the move of operations on the critica l path and the method of one-machine scheduling is proposed. The hybrid neighbor hood used is not only efficient in local search procedure, but also may help ove rcome entrapments effectively and carry the search to areas of the feasible set with better prospect. “Single machine scheduling” method and another stochasti c strategy used for jumping out of entrapments can help search find improved loc al optima. The proposed approach is tested on all the 10 jobs and 10 machines pr oblem instances available from the literature, including the notorious problem i nstance ft10, and some hard problem instances among those generated by Lawrence. The approach finds the optimum solutions of all these 10×10 problem instances except la 4 in a reasonable amount of computer time. Performance comparison show s that the proposed approach yields better results in several cases than the oth er approximation procedures discussed in the literature.
Mining Frequent Patterns Based on IS\++-Tree Model
Ma Haibing, Zhang Chenghong, Zhang Jin, and Hu Yunfa
2005, 42(4):  588-593. 
Asbtract ( 465 )   HTML ( 0)   PDF (366KB) ( 453 )  
Related Articles | Metrics
IS-tree is a novel mathematical model presented recently, which has been success fully applied to full-text index and storage in text database. In this paper, it s application is extended to data mining and an algorithm is presented for minin g frequent patterns based on IS\++-tree. The algorithm builds frequent patterns directly, as FP-growth algorithm does. However, it has several advantages over t he FP-tree model. Firstly, it scans the transaction database only once. Secondly , the mining process is only associated with one root tree. Thirdly, IS\++-tree can be dynamically updated by increments. The performance study shows that the a lgorithm efficiency is equal to or even higher than FP-growth. Above all, IS\++- tree is a good model to index transaction database, and it can support query on transactions with high efficiency.
A Selective Approach to Neural Network Ensemble Based on Clustering Technology
Li Kai, and Huang Houkuan
2005, 42(4):  594-598. 
Asbtract ( 479 )   HTML ( 0)   PDF (310KB) ( 650 )  
Related Articles | Metrics
A neural network ensemble is a very popular learning paradigm where the outputs of a set of separately trained neural network are combined to form one unified p rediction. To improve the effectiveness of ensemble, neural networks in the ense mble are not only highly correct but make their errors on different parts of the input space as well. However, most existing approaches ensemble all the availab le neural networks for prediction. In this paper, a selective approach to neural network ensemble based on clustering technology is presented. After individual neural networks are trained, the clustering algorithm is used to select a part o f the trained individual networks in order to reduce their similarity. Then many selected neural networks are combined. Experimental results show that this appr oach outperforms the traditional ones that ensemble all of the individual networ ks.
Implicit Curve Based on Radial Basis Function Network
Li Daolun, Lu Detang, and Kong Xiangyan
2005, 42(4):  599-603. 
Asbtract ( 508 )   HTML ( 0)   PDF (280KB) ( 419 )  
Related Articles | Metrics
A new method for closed curve construction is introduced, which is based on the combination of RBF (radial basis function) neural network and the principle of i mplicit curve construction. The algorithm, firstly, constructs the input and output of the RBF neural network from the constraint points, secondly changes the implicit function that represents object boundary into explicit function, thirdly uses RBF neural network to fit the curve of the explicit function, and finally obtains the fitting curves that represent the object boundary from the simulation surface. The main difference between the new method and other methods is that it is unnecessary for the new method to minimize the sum of the squares of the E uclidean distance or to solve linear system. The method not only has better resu lts than the method based on BP neural network, and also has some merit of local ity that other methods do not have. It has good numerical stability and robustne ss in dealing with noisy or missing data. Experimental results are given to veri fy the effectiveness of recovering incomplete images and object boundary reconst ruction.
Two-Handed User Interface: Review and Research
Fu Yonggang, Zhang Fengjun, and Dai Guozhong
2005, 42(4):  604-613. 
Asbtract ( 507 )   HTML ( 2)   PDF (493KB) ( 582 )  
Related Articles | Metrics
The user interface tends to migrate from “technique centered” to the new stage of “user centered”. As a main interface that fits this trend, two-handed inte raction is becoming more and more important due to its naturalness and intuitive ness. The state of the art of the two-handed user interface is presented systemi cally. Based on the research of the behavior and psychology of bimanual operatio n, several interface design principles are concluded from the experimental evalu ation. Further more, the classification and comparison of several typical two-ha nded interaction techniques are discussed. Finally, some unresolved problems are analyzed and the future research directions are given.
Combined Feature Extraction Based on Canonical Correlation Analysis and Face Recognition
Sun Quansen, Zeng Shenggen, Yang Maolong, Heng Phengann, Xia Deshen
2005, 42(4):  614-621. 
Asbtract ( 662 )   HTML ( 0)   PDF (486KB) ( 650 )  
Related Articles | Metrics
Feature level fusion plays an important role in the process of data fusion. Acco rding to the idea of canonical correlation analysis(CCA), a novel method of comb ined feature extraction is proposed in this paper. The main idea of this method can be described as follows. First of all, two groups of feature vectors with th e same pattern sample are extracted, and the correlation criterion function betw een the two groups of feature vectors is established. Then, based on this criter ion function, their canonical correlation features are extracted to form effecti ve discriminant vectors for classification. The advantage of this method lies in the following aspects: firstly, it suits for information fusion; secondly, it e liminates the redundant information within the features, and a new way for class ification and recognition is proposed. In addition, the essence of the efficienc y is analyzed further in theory. The results of experiments on Yale and ORL face databases show that the recognition rate is far higher than that of the recogni tion adopting the single features, and that this algorithm is efficient and robu st.
Research on Key Technology in Construction of a Chinese 3D Face Database
Hu Yongli, Yin Baocai, Cheng Shiquan, Gu Chunliang, and Liu Wentao
2005, 42(4):  622-628. 
Asbtract ( 732 )   HTML ( 1)   PDF (448KB) ( 786 )  
Related Articles | Metrics
By acquiring the human faces through the CyberWare 3D scanner, a Chinese 3D colo r face database is constructed. To build 3D face database, the original 3D data should be preprocessed, and the 3D faces should be standardized in a uniform for mat. The key to the problem of standardization is the pixel-to-pixel alignments among the 3D faces, which is a difficult problem in computer graphics and comput er vision. Generally, the optical flow algorithm is used for the alignments, but the result is not so good. To get precise alignments of the 3D faces, a method based on resample algorithm is proposed. The experimental result shows the metho d is feasible.
Restoration of Human Face Illumination with Unknown Light Source
Zhang Zhibin, Lai Jianhuang, and Deng Donggao
2005, 42(4):  629-634. 
Asbtract ( 667 )   HTML ( 0)   PDF (370KB) ( 743 )  
Related Articles | Metrics
Human face recognition is an important technique that can be widely employed in many practical applications, and, specifically, recognizing human face under non-uniform illumination condition with unknown-position light source is a particul arly difficult problem. In this paper, a human-face-model based approach to rest ore the illumination of human face under the above condition is proposed, which work by recovering the intensity of the shaded areas in human face. Human face t exture is represented as a linear combination of several training images, and, b ased on the face texture in the bright area, face illumination is restored by re covering face texture through a least-square estimation. Experiments show that t he method is not only efficient but also fast in restoring the shaded area on hu man face.
A Form Frame Line Removal Algorithm Based on Gray-Level Image
Zhang Chongyang, Chen Qiang, Lou Zhen, and Yang Jingyu
2005, 42(4):  635-639. 
Asbtract ( 635 )   HTML ( 0)   PDF (395KB) ( 821 )  
Related Articles | Metrics
Preprocess procedure is an important procedure in a document image analysis (DIA) system. In practical document images, characters usually overlap with the prep rinted form frames, creating tremendous problems for the recognition engines. Mo st of the form frame line removal algorithms are based on bi-level images, which have lost much useful information during the binary stage. Proposed in this pap er is a line removal algorithm directly based on gray-level images. First, cross -points of characters and lines are detected by Soble gradient. Then the overlap ping types of characters and lines are converted into touch type or crossover ty pe by cross-points analysis. Finally, lines are removed with topological method. Experiment results on 1225 real life character string images demonstrate the ef ficiency of this algorithm. The recognition rate is improved from 75.9% to 91.4% .
An Improved Parallel Watershed Algorithm for Image Segmentation
Zhou Haifang, Yang Xuejun, Liu Hengzhu, and Tang Yu
2005, 42(4):  640-646. 
Asbtract ( 657 )   HTML ( 2)   PDF (414KB) ( 487 )  
Related Articles | Metrics
Watershed transform is a classical and effective method of image segmentation in mathematical morphology. Nevertheless classical watershed algorithms have stron g recursive nature, so straightforward parallel ones have a very low efficiency. Firstly, the advantages and disadvantages of some existing parallel algorithms are analyzed. Then, an improved parallel watershed algorithm is presented based on the definition of watershed transform with condition of distance. For this al gorithm, a new definition of components graph is described and good locality of algorithm is developed in use of lower distance. Through introducing pseudo-seed pixels, the process of region growing and region merging is separated, and for this reason the cost of re-flooding is eliminated. At the same time, the algorit hm presented can create watershed lines directly. As the experiments show, the a lgorithm optimizes both running time and relative speedup, and has more strong f lexibility.
A Distributed PEP-Based TCP Performance Enhancing Protocol for Satellite Networks
Huo Longshe, Zheng Yanfeng, and Gao Wen
2005, 42(4):  647-654. 
Asbtract ( 887 )   HTML ( 4)   PDF (444KB) ( 634 )  
Related Articles | Metrics
Based on distributed performance enhancing proxies, a new transport protocol for the satellite environment, the XP protocol is presented. It is designed and opt imized to overcome the performance degradation that TCP experiences in satellite networks due to their long latency, high bit error rate, and bandwidth asymmetr y characteristics. Main contributions include a reliable two-and-half-way handsh ake connection establishment mechanism, rate control and a measurement-based ada ptive bandwidth allocation algorithm, and a delayed acknowledgements technique b ased on active sender-request. Interactions with the TCP connections in terrestr ial segments are also considered. Emulation and real operating environment exper iments show that the throughputs achieved by XP improve substantially in compari son with the regular TCP, and the fairness is also maintained perfectively when the bandwidth is shared among multi-connections.
An Adaptive Robust TCP/IP Header Compression Algorithm
Wu Yichuan, Huang Kui, Zheng Jianping, Sun Limin, and Cheng Weiming
2005, 42(4):  655-661. 
Asbtract ( 730 )   HTML ( 1)   PDF (373KB) ( 476 )  
Related Articles | Metrics
In order to improve the spectrum efficiency of wireless channel, header compression schemes are used to reduce the TCP/IP protocol header size in wireless IP ne tworks. However, all existing header compression schemes do not take wireless ch annel state into consideration. For better usage over wireless links, a new adap tive robust TCP/IP header compression algorithm is proposed and analyzed. Throug h adjusting the dimension of variable sliding window (VSW) of W-LSB encoding in header compressor with the accurate estimation of wireless channel state, this a lgorithm can achieve a good balance of compression ratio and error-resistant rob ustness. Simulation results are given to demonstrate the effectiveness of this a daptive algorithm over wireless link.
Research on Ensuring QoS and Its Admission Control in Web Servers
Li Zhendong and Xie Li
2005, 42(4):  662-668. 
Asbtract ( 486 )   HTML ( 1)   PDF (407KB) ( 456 )  
Related Articles | Metrics
The explosive growth in WWW is the main cause of Internet success. But because t he current Web servers architecture is short of QoS ensuring mechanism, it leads to significant revenue losses while server is overloaded. A QoS ensuring archit ecture for Web servers which provide multi-class service is presented in this pa per. It uses (SRD) service response delay as the core measurement criterion of t he architecture. With SRD, one class service is abstracted to a virtual server, and every component of a system are glued together. And a dynamic adaptive servi ce admission control algorithm (DASAC) is also presented. Experiments show that, although suffering severe overloading and burst, the user-perceived QoS can be ensured.
Research and Implementation of Web Service Differentiated QoS
Xu Mingwei, Hu Chunming, Liu Xudong, and Ma Dianfu
2005, 42(4):  669-675. 
Asbtract ( 605 )   HTML ( 1)   PDF (433KB) ( 533 )  
Related Articles | Metrics
With the development of web service applications, QoS becomes a severe problem. By using the average response time as the performance metric, a way about web se rvices differentiated QoS is verified and then a method to model target response time is proposed. Meanwhile, an efficient method is proposed to solve dynamic p rogramming questions. A prototype system is designed and implemented in the web service runtime environment. Experimental results show that this method is effec tive to control workload balance and provide web service differentiation.
Formal Specification and Verification for Group Communication Algorithm Suiting Extended Virtual Synchrony
Deng Yanjun and Xu Xuezhou
2005, 42(4):  676-683. 
Asbtract ( 599 )   HTML ( 0)   PDF (431KB) ( 375 )  
Related Articles | Metrics
Group communication systems(GCSs) are communication middleware systems which fac ilitate the development of fault-tolerant distributed systems. One of the import ant concepts in GCSs is virtual synchrony(VS), the essence of which is that it g uarantees the order of group membership messages and application multicasting me ssages which are delivered to group members. In order to handle network partitio ns, the extended virtual synchrony(EVS) model is introduced. Moreover, this pape r incrementally constructs the specifications of GCSs which satisfy the EVS mode l using the inheritance-based methodology of the I/O automata formalism. Conside ring the characteristics of the EVS model, this paper presents an architecture o f GCS which is based on a client/server model, and formally formulates the servi ce and algorithm of the system using an I/O automata model. The algorithm of sui ting the EVS model is emphasized. It is formally specified and verified using th e inheritance-based methodology of the I/O automata formalism.
A Study of the Relationship Between Performance of Topology-Based MANET Routing Protocol and Network Coverage Density
Wen Yingyou, Zhao Jianli, Zhao Linliang, and Wang Guangxing
2005, 42(4):  684-689. 
Asbtract ( 447 )   HTML ( 2)   PDF (342KB) ( 648 )  
Related Articles | Metrics
In most existing performance evaluations of ad hoc routing protocols, connectivi ty is not considered properly. To solve this problem, a new concept of network coverage density is adopted, which reflects communication ability of nodes and connectivity of ad hoc network. Based on this, a research is conducted on the rela tionship between performance of topology-based MANET routing protocols and netwo rk coverage density. Simulation results show that, packet delivery of routing pr otocol heightens with the increase of network coverage density, and they form an S curve. Through the definition of minimal network coverage density, a judgment criterion for applicability of routing protocol is provided in the process of a d hoc networking.
Two Results of Decomposing Weakly Invertible Finite Automata
Wang Hongji
2005, 42(4):  690-696. 
Asbtract ( 437 )   HTML ( 0)   PDF (386KB) ( 419 )  
Related Articles | Metrics
It is very important to investigate the decompositon of weakly invertible finite automata, since it could provide an approach to cryptanalyzing finite automaton public-key cryptosystem (FAPKC), which was proposed by Tao Renji and Chen Shihu a in 1985. This paper deals with this topic. For an n-ary weakly invertible fini te automaton (WIFA, for short) M with delay τ, a necessary and sufficient cond ition is obtained, that M can be decomposed into a WIFA with delay 0 and a so-ca lled τ-order delay unit, i.e., the τ-output weight of s is 1 for any state s i n M. Meanwhile, based on a class of weakly invertible finite automata that canno t be decomposed into a delay unit, an example is hereby constructed. Thus, a neg ative answer to an open question presented by Bao Feng in 1993 is given.
DOS Attack and Defense in Mobile Ad Hoc Networks
Yi Ping, Zhong Yiping, and Zhang Shiyong
2005, 42(4):  697-704. 
Asbtract ( 671 )   HTML ( 0)   PDF (415KB) ( 543 )  
Related Articles | Metrics
Mobile ad hoc networks will often be deployed in environments where the nodes of the networks are unattended and have little or no physical protection against tampering. The nodes of mobile ad hoc networks are thus susceptible to compromise . The networks are particularly vulnerable to denial of service (DOS) attacks la unched through compromised nodes or intruders. Presented in this paper is the ad hoc flooding attack, a new attack, which results in denial of service when used against on-demand routing protocols for mobile ad hoc networks, such as AODV, D SR. The intruder broadcasts mass useless route request packets or sends a lot of useless DATA packets to exhaust the communication bandwidth and node resource s o that the valid communication can not be kept. After analyzing the attack, floo ding attack prevention (FAP) is developed, which is a generic defense against th e ad hoc flooding attacks in mobile ad hoc networks. The FAP is composed of neig hbor suppression and path cutoff. When the intruder broadcasts exceeding packets of route request, the immediate neighbors of the intruder observe a high rate o f route request and then they lower the corresponding priority according to the rate of incoming queries. Moreover, not serviced low priority queries are eventu ally discarded. When the intruder sends many useless DATA packets to the victim node, the node may cut off the path and does not set up a path with the intruder any more. The results of the implementation show that FAP can prevent the ad ho c flooding attack efficiently.
A Security Enhancement Method of a Verifiable Threshold Signature Scheme Based on the Elliptic Curve
Wu Zhongdong, Xie Weixin, and Yu Jianping
2005, 42(4):  705-710. 
Asbtract ( 510 )   HTML ( 1)   PDF (317KB) ( 507 )  
Related Articles | Metrics
Proposed in this paper is an efficient (k, n) threshold digital signature scheme based on the elliptic curve. In the scheme, k out of n signers cooperate to iss ue a signature, while Nyberg-Ruepple digital signature scheme and Pedersen's ver ifiable secret sharing (VSS) are employed. This new scheme enhances the security of the existing threshold digital signature scheme by redistributing the privat e key d of the signature periodically into the different access structures. This design is not only more difficult for adversary to attack the threshold digital signature scheme, but also flexible for adding or deleting one or more signatur e participants. Assuming that secret communication between the signers is availa ble, the security performances of the scheme are discussed. The advantage of the new scheme has capability of intrusion tolerance besides the property of verifi able partial signature which verifiable threshold signature scheme should have.
Migrating Entities and Their Time Synchronization Issue in Large-Scale Federated Simulation (Part One): Protocol and Implementation for Migrating Entities
Liu Xiaojian, Zhong Hairong, and Jin Shiyao
2005, 42(4):  711-715. 
Asbtract ( 388 )   HTML ( 1)   PDF (304KB) ( 574 )  
Related Articles | Metrics
Entity migration can help to sustain large scale, long running distributed simul ations. The shortcomings of the high level architecture (HLA), when dealing with this issue, are presented, and the ad hoc approaches for an HLA-compliant feder ate to move entities are analyzed. A protocol used to move entities under HLA/RT I is introduced, then its implementation issues are thoroughly discussed, includ ing the starting of the migration, the entity state transfer mechanism, the mess age relaying and processing logic. The executing framework of the protocol is al so presented. Experiments prove the feasibility of this protocol. Compared with relative studies, this approach can avoid missing messages during migration, and guarantee the correct message process sequence.
Migrating Entities and Their Time Synchronization Issue in Large-Scale Federated Simulation (Part Two): Time Synchronization for Migrating Entities
Zhong Hairong, Liu Xiaojian, and Jin Shiyao
2005, 42(4):  716-720. 
Asbtract ( 445 )   HTML ( 0)   PDF (285KB) ( 495 )  
Related Articles | Metrics
Migrating entities are beneficial to improve performance of large-scale distribu ted simulation systems, while time synchronization during migration is the key a nd difficult issue. The time synchronization mechanism for migrating entities in HLA (high level architecture) federated simulation is studied. Transition of fe derates' states is analyzed, and synchronization strategies for the old federate to relay messages and for the new federate to receive and process messages are proposed. Under these strategies, messages are processed in time-stamp order, th e logical time of both sides is synchronized, and messages are never lost or dup licated. Message retraction is also supported to enable optimistic synchronizati on.