GaAs-based micromachining is a very attractive technique for integrating mechanical structures and active optical devices, such as laser diodes and photodiodes. For monolithically integrating mechanical parts onto laser diode wafers, the micromachining technique must be compatible with the laser diode fabrication process. Our micromachining technique features three major processes: epitaxitial growth (MOVPE) for both the structural and sacrificial layers, reactive dry-etching by chlorine for high-aspect, three-dimensional structures, and selective wet-etching by peroxide/ammonium hydroxide solution to release the moving parts. These processes are compatible with laser fabrication, so a cantilever beam structure can be fabricated at the same time as a laser diode structure. Furthermore, a single-crystal epitaxial layer has little residual stress, so precise microstructures can be obtained without significant deformation. We fabricated a microbeam resonator sensor composed of two laser diodes, a photodiode, and a micro-cantilever beam with an area of 400700 µm. The cantilever beam is 3 µm wide, 5 µm high, and either 110µm long for a 200-kHz resonant frequency or 50 µm long for a 1-MHz resonant frequency. The cantilever beam is excited by an intensity-modulated laser beam from an integrated excitation laser diode; the vibration signal is detected by a coupled cavity laser diode and a photodiode.

A system for measuring the low frequency amplitude and phase noises was set-up, with employing a phase sensitive detector and phase-shifter. It is noted that both noises were partly correlated. The phase noise was explained by the transit time fluctuation due to the fluctuating diffusion coefficient. The amplitude noise reduction was demonstrated by applying the inverted output of the phase noise to the amplitude noise.

For symmetric cryptosystems, their transformations should have nonlinear elements to be secure against various attacks. Several nonlinearity criteria have been defined and their properties have been made clear. This paper focuses on, among these criteria, the propagation criterion (PC) and the strict avalanche criterion (SAC), and makes a further investigation of them. It discusses the sets of Boolean functions satisflying the PC of higher degrees, the sets of those satisfying the SAC of higher orders and their relationships. We give a necessary and sufficient condition for an n-input Boolean function to satisfy the PC with respect to a set of all but one or two elements in {0,1}n{(0,...,0)}. From this condition, it follows that, for every even n 2, an n-input Boolean function satisfies the PC of degree n 1 if and only if it satisfies the PC of degree n. We also show a method that constructs, for any odd n 3, n-input Boolean functions that satisfy the PC with respect to a set of all but one elements in {0,1}n{(0,...,0)}. This method is a generalized version of a previous one. Concerned with the SAC of higher orders, it is shown that the previously proved upper bound of the nonlinear order of Boolean functions satisfying the criterion is tight. The relationships are discussed between the set of n-input Boolean functions satisfying the PC and the sets of those satisfying the SAC.

This paper presents improvement of data error rate against burst noise by using both chip interleaving and hard limiter in direct sequence spread spectrum (DS/SS) communication systems. Chip interleaving, which is a unique method of DS/SS systems, is effective when burst noise power is small. However, when the burst noise power is large, date error rate is degraded. While, though hard limiter suppresses burst noise power, it gives little effectiveness when the burst noise length is long. Using chip interleaving and hard limiter together, as they work complementary, stable and considerable improvement of data error rate is achieved.

Exploiting the full potential of a multiprocessor system requires a good job scheduling algorithm. In this paper we analyze three dynamic job scheduling algorithms in multiprocessor systems. These algorithms are based on static job scheduling algorithms, LPT (longest processing time first), SJF (shortest job first), and LPR (largest processor requirement first), each of which exhibits good performance in terms of asymptotic upper bound on the makespan of the schedule generated by it. We analyze their performance in the dynamic case experimentally, where we have a stochastic stream of jobs with arbitrary processing time and processor requirement. We compare their performance with the FCFS algorithm and its simple extension. Except for LPT, the algorithms are found to perform significantly better than FCFS, while among themselves SJF performs the best, followed by K-LPR, a variation of LPR. We also consider the fairness aspect of these algorithms and propose a general technique to impose fairness on these algorithms. Finally, we analyze the impact of imposing fairness on the performance of these algorithms.

In this paper, we propose a new class of ID-based non-interactive key sharing scheme with a trusted center which generate a common key on the basis of a linear combination of the center secrets. We also discuss the security of the proposed schemes, and show that the proposed schemes prevent the conventional collusion attack, by adding another random integers unique to each user, on the secret vector that is assigned to the user. Furthermore, we present a new type of a statistical collusion attack which is more suitable for the proposed schemes. We also present the lower bound of the threshold of the statistical collusion attack on the proposed schemes. The proposed schemes can be easily implemented compared with other schemes as they require only computing of the inner product of two vectors over finite ring (including finite field) and an Euclidean quotient, for generating the common key. Our proposed schemes can be regarded as modified versions of the Blom's original scheme. However our proposed schemes are secure against our new type of the attack, as well as the collusion attack based on the solving of the linear equations, although Blom's scheme is insecure against both of these collusion attacks.

Graph unification is doubtlessly the most expensive process in unification-based grammar parsing since it takes the vast majority of the total parsing time of natural language sentences. A parsing time overload in unification consists in that, in general, no less than 60% of the graph unifications performed actually fail. Thus one way to achieve unification time speed-up is focusing on an efficient, fast way to deal with such unification failures. In this paper, a process, prior to unification itself, capable of filtering or stopping a considerably high percentage of graphs that would fail unification is proposed. This unification-filtering process consists of comparison of signatures that correspond to each one of the graphs to be unified. Unification-filter (hereafter UF) is capable of stopping around 87% of the non-unifiable graphs before unification itself takes place. UF takes significantly less time to detect graphs that do not unify and discard them than it would take to unification to fail the attempt to unify the same graphs. As a result of using UF, unification is performed in an around 71% of the time for the fastest known unification algorithm.

This paper proposes a new approach to the management of large-scale communication networks. To manage large-scale communication networks effectively, it is essential to get a bird's-eye view of them when they are in their normal conditions. When an indication of faulty state is detected, the focus of the management is narrowed down to the faulty network elements until the appropriate granularity is reached. This management scheme is called multilevel network management in this paper, and it first explains the significance of this scheme for large-scale communication networks and presents some ideas on implementing this management scheme. It then proposes that system identification be used in multilevel network management. The system identification is used to measure transmission delays between two arbitrarily selected nodes in the networks, and multilevel network management is achieved by selecting those two nodes appropriately in accordance with the levels to be managed. Finally, it is demonstrated by computation simulation results that the proposed method is suitable for multilevel network management.

A novel total harmonic distortion (THD) measuring technique is proposed. A modified Volterra series using harmonics in place of powers of the sinusoidal input is used to identify the nonlinear models of the source and the device under test (DUT). The least-mean square (LMS) adaptive algorithm is applied for identification. While maintaining comparable speed and accuracy this technique provides a more flexible test procedure than conventional methods, in terms of the frequency resolution, the number of samples, and the sampling rate. It outperforms conventional methods when there is a bin energy leakage, which occurs in a non-coherent system. In addition, it is real-time computing while other conventional methods post process blocks of data. Simulation results collaborate the analytical results.

This paper proposes a suitable distributed computing model as the basis for building a network management system. Author has been studying a distributed reactive model, called Meta for this purpose at Cornell University. Effectiveness using Meta is to provide high level program interface for developing network management system, and programmers can achieve network management system with coding small amount of programs. It also realizes easy additions and modifications for network management application programs. To confirm the effectiveness of the proposal, the author has utilized Meta to implement an experimental network management system. The experimental system provides high level interfaces for monitoring and controlling network components. It also supports reliable communication over distributed nodes. Preliminary evaluation of the system shows that critical network management applications are provided within an appropriate response time for all applications provided by SNMP, with small development cost and easy system modification.

Trouble management is a key function in solving the problems and maintaining the high communications capability of a network when communication service network users encounter problems in the quality of services [1]. This paper proposes technologies and architecture for an intelligent management system to achieve advanced service/network trouble management. The system generates operation scenarios to find a cause to solve a reported trouble, executes them, and modifies them according to operation circumstance changes. In the scenario execution process, fault propagation simulation is used to isolate a fault in necessary cases. The evaluation of the system applied to the ISDN services shows that the proposed system can achieve high-speed, precise trouble management by the integrated cooperative work of a human (operator) and a machine (operation system).

This paper discusses Boolean functions satisfying the propagation criterion (PC) and their balancedness. Firstly, we discuss Boolean functions with n variables that satisfy the PC with respect to all but three elements in {0,1}n-{(0,...,0)}. For even n4, a necessary and sufficient condition is presented for Boolean functions with n variables to satisfy the PC with respect to all but three elements in {0,1}n-{(0,...,0)}. From this condition, it is proved that all of these Boolean functions are constructed from all perfectly nonlinear Boolean functions with n-2 variables. For odd n3, it is shown that Boolean functions with n variables satisfying the PC with respect to all but three elements in {0,1}n-{(0,...,0)} satisfy the PC with respect to all but one elements in it. Secondly, Boolean functions satisfying the PC of degree n-2 and their balancedness are considered. For even n4, it is proved that an upper bound on the degree of the PC is n-3 for balanced Boolean functions with n variables. This bound is optimal for n=4,6. It is also proved that, for odd n3, balanced Boolean functions with n variables satisfying the PC of degree n-2 satisfy the PC with respect to all but one elements in {0,1}n-{(0,...,0)}.

A highly reliable, operated efficiently, and scalable switching node is required for multivendor system management networks. This paper presents techniques for applying the Telecommunications Management Network (TMN) concept to a distributed switching node. First, we select the TMN protocol structures so as to minimize the command response time. To ensure efficient operation and flexible design of the operation, administration and maintenance (OA & M) software, we propose a command-message handling function for mapping managed objects (MOs) to the OA & M software. We have designed MO classes for switching-system-specific devices with a redundant configuration that ensures highly reliable system operation. The feasibility of these techniques have been confirmed on a prototype system.

We propose an RSA-type scheme over the nonsingular part of a singular cubic curve En (0,b) : y2x3+bx2 (mod n), where n is a product of form-free primes p and q. Our new scheme encrypts/decrypts messages of 2 log n bits by operations of the x and y coordinates. The decryption is carried out over Fp or a subgroup of a quadratic extension of Fp, depending on quadratic residuosity of message-dependent parameter b. The decryption speed in our new scheme is about 4.6 and 5.8 times faster than that in the KMOV scheme and the Demytko scheme, respectively. We prove that if b is a quadratic residue in Zn, breaking our new scheme over En(0,b) is not easier than breaking the RSA scheme.

This paper compares three typical system-sharing configurations for FTTH networks that provide narrowband and video distribution services and proposes a remote node locating strategy for each configuration. Two new evaluation factors, required land space and service provisioning effort, are included in the calculation, in addition to facility cost and maintenance effort. By considering these factors together, the total network cost is calculated and the sensitivity to the number of remote nodes is evaluated. Finally, the most economical system-sharing configuration is identified on the basis of the evaluations for two typical service areas in Japan, for both present and future cost environments.

In this paper, chaos synchronization in coupled discrete-time dynamical systems is studied. Computer results display the interesting synchronization behaviors in the mutually coupled systems. As possible applications of chaos synchronization, parameter estimations and secure communications are proposed. Furthermore, a modified OGY method is given, which converts a chaotic motion into a periodic motion.

A new approach to implement queues for controlling ATM switch LSI is presented. In many conventional architecture, external FIFOs are provided for each output link and used to manage the address of the buffer in an ATM switch. We reduce the number of FIFOs by using a self-timed queue with a search circuit that finds the earliest entry for each output link. Using this architecture, number of the FIFOs is reduced to 1/N, where N is the switch size. Delay priority and multicasting can be supported without doubling the number of the queues. This new queue can also be utilized as an ATM switch by itself. Evaluation chip was fabricated using 0.5-µm CMOS process technology. Inter-stage transfer speed over 500 MHz and cycle time over 125 MHz was obtained. This performance is enough for a 622-Mbps 1616 ATM Switch.

This paper presents the results of a study made to determine the line length coverage of the high-bit-rate digital subscriber line (HDSL) present in NTT's local networks. The HDSL carries one bi-directional 784 kbit/s channel per pair and supports the digital interface at 1544kbit/s by using two cable pairs. The primary purpose of this study is to estimate the range limits for candidate transmission schemes considering line installation conditions, and to determine the most promising transmission scheme and its feasibility given the environment of NTT's local networks. Pulse amplitude modulation (PAM) and quadrature amplitude modulation (QAM) transmission schemes are compared for HDSL implementation. It is shown that 2B1Q-PAM and 16-QAM generally achieve better performance than the more complicated PAM and QAM given the presence intra-system crosstalk interference (interference between identical transmission systems). The range limits determined by inter-system crosstalk interference (interference between different transmission systems) with basic rate access (BRA) implementing a burst-mode transmission method are also estimated. This paper concludes that 2B1Q-PAM achieves the best overall performance in NTT's local networks. A feasibility study of 192-6144 kbit/s transmission is also described.

An Ultra-high-speed (higher than 60 MHz) Viterbi decoder VLSIC with coding rates from one-half to fifteen-sixteenth and a constraint length of seven for forward error correction (FEC) has been developed using 0.8-µm semicustom CMOS LSIC technology and a newly developed high-speed ACS circuit. To reduce power consumption of the one-chip Viterbi decoder, a universal-coding-rate scarce-state-transition (SST) Viterbi decoding scheme and low-power-consumption burst-mode-selection (BMS) path memory have been proposed and employed to the developed VLSIC. In addition, a new maximum-likelihood-decision (MLD) circuit for the SST Viterbi decoder has been developed. The total power consumption of the developed chip is reduced to 75% of the conventional one and the developed Viterbi decodar VLSIC achieves a maximum operation speed of 60 MHz. It achieves near theoretical net coding-gain performance for various coding rates.

This paper presents an alternative traffic model for an ATM multiplexer providing video, voice, image, and data services. The traffic model classifies the input traffic into two types: real-time and non-real-time. The input process for realtime traffic is periodic and correlated, while that for non-realtime traffic is batch Poisson and independent. This multiplexer is assumed to be a priority queueing system with synchronous servers operating on time-frame basis and with separate finite buffers for each type of traffic. State probabilities and performance measures are successfully obtained using a Markov analysis technique and an application of the residue theorem in complex variable. The results can be applied in the design of an ATM multiplexer.