摘要:

AbstractThe construction of large‐scale ATM switching networks is done mainly to configure switching networks only with buffered unit switches and to establish a simple path on a call‐by‐call basis. With this approach it is desired that ATM switching networks be nonblocking at call level qr there always exists a path consisting only of inner links with sufficient idle capacity.An efficient configuration of a new class of ATM switching networks is proposed which are nonblocking at the call level. First, the proposed network is shown in the most generalized form and the nonblocking condition is derived. Then, by reforming this generalized configuration into a more practical one, two types of configurations are derived. Through comparison of the two types, it is clarified that Parallel Benes Configuration is optimum. It has the following features: (1) It can be composed by only one kind ofn × nsquare unit switches; (2) it is very fault‐tolerant; (3) besides non‐blocking at the call level, good cell‐level traffic performance is expected. It is shown also that a high‐speed unit switch is mandatory for very large‐scale ATM switching networks to reduce the hardware volume and the numbe

ISSN:8756-6621    

DOI:10.1002/ecja.4410770401

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractA video codec for ATM (Asynchronous data Transfer Mode), which is the transfer scheme in the era of B‐ISDN (broadband aspects of ISDN) is constructed and the performance is evaluated. The experimental equipment can realize the variable‐rate coded video transmission corresponding to CCIR 601 recommendation, through the experimental switching network. The periodic cyclic refresh, packeting of the coded information, and the interframe prediction DCT layered coding are employed, realizing a robustness against the cell loss which is the cause of a drastic image quality deterioration in ATM.As a result of the subjective evaluation experiment, the image quality of MOS 3.5 or above is obtained for the cell loss ratio of 10−6using only the cyclic refresh and the packeting of the coded information, and 10−2to 10−4of low‐priority order information using the interframe prediction DCT lay

ISSN:8756-6621    

DOI:10.1002/ecja.4410770402

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractThis paper discusses the multipath allocation problem in the communication network and proposes a multipath allocation algorithm suited to the distributed environment. The multipath allocation problem is a problem where multiple paths are to be allocated without a conflict. This problem has been difficult to solve in the distributed environment since relations among various routes must be considered. The algorithm proposed in this paper solves this point by realizing the asynchronous path search.In this paper, the basic characteristics of the proposed algorithm are described through the comparison between the proposed and the traditional algorithms for the case of serial computation under the centralized control. Then the proposed algorithm is applied to several kinds of distributed cooperative multiagent models, and it is demonstrated that the proposed model can easily be applied to the distributed environment. A simulation is executed actually using several workstations, and it is shown that the proposed algorithm is desirable as a solution method for the multipath allocation problem in the distributed environment.

ISSN:8756-6621    

DOI:10.1002/ecja.4410770403

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractISDN network resources such as links and nodes can be shared by various classes of calls with different bit‐rate service quality, holding time characteristics, etc., raising expectations for economical design and efficient operation of the networks. Heterogeneous‐traffic network dimensioning is defined as a problem of determining each link size so as to minimize the total network cost while satisfying the individual loss probability specifications. This is formulated as a nonlinear programming problem and an iterative algorithm for solving it is proposed.The iterative algorithm is similar to the one used in single‐class traffic network dimensioning, but the optimality conditions are constructed by temporarily using the individual loss probability constraints for the traffic classes for which the equality is held in the preceding iteration. The algorithm then is applied to dimensioning networks serving multiple bit‐rate traffic, and its dimensioning examples are shown. The result shows design characteristics peculiar to a heterogeneous traffic environment, such as the possibility of two local minima, and network topologies with zero direct link size and nonzero direct link size ar

ISSN:8756-6621    

DOI:10.1002/ecja.4410770404

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractVarious communications services are being provided over ATM networks, and among them, file transfer is considered to be an important service. In file transfer, cases in which immediacy with respect to the connection request of a call is not as strong as in voice and picture may also be considered. In addition, in band allocation necessary during the call setup time, the transmission rate may be established freely to a certain extent by a terminal.In view of the foregoing, in this article, performance of file transfer multiplexed with other traffic in ATM networks is evaluated considering the forementioned characteristics.First, the effects of the file transmission rate and the performance of the system maintaining the file transfer over a single channel are studied and then the performance over multichannels is evaluated. As a measure of these performances, the throughput, which is important in file transfer, also is evaluated using the call loss probability considered to be important from the viewpoint of the user.

ISSN:8756-6621    

DOI:10.1002/ecja.4410770405

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractLiNbO3Mach‐Zehnder optical modulators, which are expected to be used for multi‐Gbit/s optical transmission scheme, have one problem which is a dc drift phenomenon. In order to operate this modulator for a long time, a control unit to adjust a bias voltage for such dc drift is therefore necessary. So far, a method to superimpose a control signal with a main signal has been commonly used. However, the control signal may influence the output signal in this method. A novel automatic bias voltage control scheme utilizing the feature of Mach‐Zehnder modulator with traveling‐wave electrodes is proposed in this paper. A dc drift can be detected with a probe light propagating along a direction opposite to that of a modulated voltage waveform, and therefore very little influence of the probe signal on the output signal would be expected. A theoretical discussion on a dc drift detection and control method is described. This method has been successfully demonstrated to work for a mark density of 1/4, 1/2,

ISSN:8756-6621    

DOI:10.1002/ecja.4410770406

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractThe Virtual Path concept has several beneficial features to construct an economical and efficient Asynchronous Transfer Mode (ATM) network. The Virtual Path capacity control technique can increase network flexibility and improve transmission efficiency through statistical sharing of facilities.This paper presents an adaptive virtual path capacity control method which maximizes transmission efficiency under the condition of limited processing load. Evaluation shows that the proposed method successfully improves network transmission efficiency with a slight increase in processing load compared to the fixed virtual path capacity scheme.

ISSN:8756-6621    

DOI:10.1002/ecja.4410770407

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractPipelined protocol processors are used in high‐speed networks such as FDDI for achieving high‐throughput data communications between end users. In this paper, we present a model for computing the maximum throughput that can be delivered by such processors.The parameters of the model are: input packet length, length of PDU (Protocol Data Unit) and SDU (Service Data Unit) headers, packet processing time of pipeline stages, capacity of the FIFO channels that interconnect stages, and average number of packets assembled/segmented at a stage.As an example, the maximum throughput of an idealized pipelined processor that executes the FDDI, LLC Class I, IP, and TCP protocols is computed. Also, we show that the formulas given in [1] and [2]are particular cases of the formulas presented in this paper.In a typical application of the proposed model, some desired, or measured, throughput characteristics of a protocol processor will be given, and the parameters of the processor will be compu

ISSN:8756-6621    

DOI:10.1002/ecja.4410770408

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractA highly efficient random access scheme ICMA‐PE (Idle‐Signal Casting Multiple Access with Partial Echo) is proposed for digital mobile communications. In this scheme, partial echo (PE) bits are broadcast with idle busy bite and receive nonreceive indicating bits as collision control bits to control mobile transmission, thus improving the random access efficiency.First, frame formats, behavior of mobile and base stations, and the transmission procedure are explained.Next, the packet failure rate and throughput are evaluated with both theoretical analysis and computer simulations. As a result it is clarified that about 20 percent improvement in maximum throughput is achieved compared to conventional schemes. The improvements achieved with transmissions also are mentioned and the transmission characteristic when various message length signals coexist is clarified. Finally, the effect of the coincidence of a PE with one from another station is evaluated, and it is shown that the PE bit length of about 7 bit is necessary to maintain the characteris

ISSN:8756-6621    

DOI:10.1002/ecja.4410770409

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY

摘要:

AbstractThe Viterbi equalizer is effective for frequency selective fading in digital mobile communication, but as the multipath component propagation delay time difference increases, the number of states increases exponentially and the computation amount increases drastically. Combining the Viterbi equalizer and interference reduction operation based on the equalization result, this paper proposes a new type of Viterbi equalizer with iterative equalization. It can expand the equalizable range without increasing the number of states. Computer simulation result shows that its performance is better than that of a conventional one.

ISSN:8756-6621    

DOI:10.1002/ecja.4410770410

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1994

数据来源: WILEY