Surging capacity demand triggered by the increasingly mobile-oriented and exponentially growing Internet has accelerated convergence of networking technologies. In the core network side, IP and photonics have been the two key driving factors of technical innovations. Amid this technical turmoil, Generalized MPLS (GMPLS) in IETF has recently attracted sizable attentions, as it offers potential for "Grand Unification Theory" for network technology convergence. Despite its prospects, however, the proposal is still missing comparable structures in management plane, which is in dire need for carrier-class, reliable operations. Among many industry proposals and standards, TINA vision on connection management architecture (CMA) is the one offering practical and deployable architecture for the converged photonic IP network. TINA IP Control and Management (IPCM) WG was established during TINA phase II (1998-2000), to study IP control and management issues using the architecture basis of TINA-CMA. Latest activities in TINA IPCM WG, compiling experience at Sprint, Telia, Telecom Italia Lab., and Fujitsu, have resulted in a specification for connectivity provider reference points, namely ConS, ConC, and FCon. Use of TINA CMA as building blocks for the IP photonic network convergence is illustrated. An overview of a ConS reference point specification for managed IP connectivity service, named ConS-IPCM, is explained.

Recently, demand on class-of-service (CoS) has known a great increase thanks to a set of real-time applications such as Internet Telephony service. Class-Based Queuing (CBQ) is considered as an efficient queuing mechanism to guarantee CoS. ALTQ is a widely used platform for realizing CBQ. In this paper, we verify through experiments that bandwidth control of CBQ/ALTQ contains overhead for fluctuating traffic. To avoid such an overhead, we introduce dynamic bandwidth allocation scheme for real-time traffic fluctuating within fixed ranges. In the light of the limited network resources, it quickly becomes obvious that when the traffic rate exceeds the maximum available bandwidth, arriving packets will be accumulated in the router queue. As a result, the traffic delay increases and the quality of real-time applications is degraded. To cope with such a problem, we revise the RED algorithm for a large amount of traffic and propose a new packet discard algorithm that uses bandwidth as a trigger. Experiment results show that our proposal outperforms the already existing packet discard algorithms (RED, DropTail) in providing lower delay/jitter services. We show the efficiency of our proposal using a real system.

The concept of stochastic association has originally been proposed in relation to single-electron devices having stochastic behavior due to quantum effects. Stochastic association is one of the promising concepts for future VLSI systems that exceed the conventional digital systems based on deterministic operation. This paper proposes a CMOS stochastic associative processor using PWM (pulse-width modulation) chaotic signals. The processor stochastically extracts one of the stored binary patterns depending on the order of similarity to the input. We confirms stochastic associative processing operation by experiments for digit pattern association using the CMOS test chip.

A method is described that can allocate bandwidth to each user flow fairly in a scalable network architecture such as differentiated services architecture. As promising queueing techniques for providing differentiated services, class-based packet scheduling and selective packet discarding have been attracting attention. However, if we consider that bandwidth should be allocated to each flow in a weighted manner, the parameters used in these methods such as the weight assigned to each class queue should be pre-determined appropriately based on an assumption about the number of flows in each class. Thus, when the actual traffic pattern differs from the assumed one, they may not work well. Instead of assuming the traffic conditions, our method estimates the number of active flows in each class by simple traffic measurement and dynamically changes the weight assigned to each class queue based on the estimated number. Our method does not need to maintain the per-flow state, which gives it scalability. Simulation showed that this method is effective under various patterns of the number of active flows.

The IETF Differentiated Services (DiffServ) framework achieves scalability by (1) aggregating traffic flows with coarse grain QoS on the data plane, and (2) allocating network resources with a bandwidth broker (BB) on the control plane. However, there are many issues that need to be addressed under such framework. First, it has been shown that the concatenation of strict priority (SP) scheduler of class-based queues (CBQ) can cause delay jitter unbounded under certain utilization, which is not acceptable to support the premium service (PS). Furthermore, it is not clear how such a DiffServ network can support traffic flows requiring the guaranteed service (GS), which is a desirable feature of the future Internet. This paper presents architecture and mechanisms to support multiple QoS under the DiffServ paradigm. On the data plane, we present a node architecture based on the virtual time reference system (VTRS). The key building block of our node architecture is the core-stateless virtual clock (CSVC) scheduling algorithm, which, in terms of providing delay guarantee, has the same expressive power as a stateful weighted fair queueing (WFQ) scheduler. With the CSVC scheduler as our building block, we design a node architecture that is capable of supporting integrated transport of the GS, the PS, the assured service (AS), and the traditional best effort (BE) service. On the control plane, we present a BB architecture to provide flexible resource allocation and QoS provisioning. Simulation results demonstrate that our architecture and mechanisms can provide scalable and flexible transport of integrated traffic of the GS, the PS, the AS, and the BE services.

To generalize characteristics of a received signal level distribution from narrow- to wide-bands in a mobile radio channel, a new propagation parameter called equivalent received bandwidth (2ΔfΔLmax) has been proposed. The distributions are discussed mainly with computer simulation results. The simulation results shows the level distribution depends on 2ΔfΔLmax and power ratio a of direct to indirect waves, and the value of 2ΔfΔLmax classifies the radio channel as narrow- or wide-bands transmission. To confirm these simulated results, a field test was performed with a 3.35 GHz radio wave. This paper describes that the field test demonstrated the simulation results. It is concluded that the equation representing received signal level in the computer simulation is valid. And the fading depth depends directly on 2ΔfΔLmax, and the 2ΔfΔLmax is effective for generalizing the received signal level distribution. Furthermore, a method for calculating the power ratio was found to be better for a peak level model.

The efficiency-fractional bandwidth product (EB), which is expressed as a ratio of the radiation resistance to the absolute value of the input reactance of an antenna, is used as a performance criterion for small dielectric loaded monopole antennas (DLMAs). The dependence of the EB on the permittivity of the dielectric loading (i.e., the electrical volume) is experimentally and numerically investigated for the first time in antenna research. As a result, it is found that the EBs of the some DLMAs are enhanced over a bare monopole antenna and an EB characteristic curve has a maximum point. This result suggests the presence of the optimum electrical volume for the dielectric loading in order to obtain the best EB performance. A general reason for the existence of the peak value is also explained using a mathematical deduction. Finally the system EB, which is an efficiency-fractional bandwidth product of the DLMA with a practical matching circuit, is defined and its dependence on the relative permittivity is illustrated. Consequently, the existence of the peak value is also confirmed for the system EBs. In addition, it is demonstrated that the enhancement of the system EB is mainly due to the enhancement in the efficiency of the antenna system.

In this paper we study the scalability issue in the design of a centralized bandwidth broker model for dynamic control and management of QoS provisioning. We propose and develop a path-oriented, quota-based dynamic bandwidth allocation mechanism for efficient admission control operations under the centralized bandwidth broker model. We demonstrate that this dynamic bandwidth allocation mechanism can significantly reduce the overall number of QoS state accesses/updates, thereby increasing the overall call processing capability of the bandwidth broker. Based on the proposed dynamic bandwidth allocation mechanism, we also extend the centralized architecture with a single bandwidth broker to a hierarchically distributed architecture with multiple bandwidth brokers to further improve its scalability. Our study demonstrates that the bandwidth broker architecture can be designed in such a manner that it scales with the increase in the network capacity.

AAL2 technology, which will be used in 3rd generation mobile communications systems, can be used to efficiently transmit low-bit-rate traffic. Because user connections are multiplexed at virtual-channel connections in AAL2 networks, conventional ATM QoS management, which operates in units of VC connections, may be inadequate for managing the QoS of AAL2 connections. In this paper, we argue that for efficient utilization of network resources, it is advantageous to accommodate AAL2 connections with different QoS conditions in the same VC connection. We present a multiplexing configuration that enables QoS to be controlled at the AAL2-connection level. It works by matching the AAL2-packet-multiplexing timing to the cell-transmission timing. We also address the issue of AAL2-bandwidth management. To calculate the bandwidth, we evaluate the characteristics of multiplexing CPS packets into the ATM cell payload.

Network dimensioning is an important issue to provide stable and QoS-rich communication services. A reliable estimation of bandwidths of links between the end-to-end path is a first step towards the network dimensioning. Pathchar is one of such tools for the bandwidth estimation for every link between two end hosts. However, pathchar still has several problems. If unexpectedly large errors are included or if route alternation is present during the measurement, the obtained estimation is much far from the correct one. We investigate the method to eliminate those errors in estimating the bandwidth. To increase the reliability on the estimation, the confidence interval for the estimated bandwidth is important. For this purpose, two approaches, parametric and nonparametric approaches, are investigated to add the confidence intervals. Another important issue is the method for controlling the measurement period to eliminate the measurement overheads. In this paper, we propose a measurement method to adaptively control the number of measurement data sets. Through experimental results, we show that our statistical approaches can provide the robust estimation regardless of the network conditions.

The differentiated services (diffserv) architecture has been proposed for implementing scalable service differentiation in the Internet. Expedited forwarding and assured forwarding have been standardized as Per-Hop Behaviors (PHB) in diffserv. Assured forwarding can be utilized to realize the service, which provides each user with a minimum guaranteed rate and a fair share of the residual bandwidth. We call it guaranteed rate (GR) service. With GR service, each packet for flow i is marked in or out based on comparison between the sending rate and the minimum guaranteed rate. When congestion occurs in networks, out packets are dropped more aggressively than in packets. Recently, several fair queuing schemes have been proposed for core stateless networks. They can achieve fairer bandwidth allocation than random early detection (RED). However, there have not been any studies that consider in/out bit usage to support GR service. This paper proposes how to extend the schemes that have been proposed for core stateless networks to allow the support of in/out bit usage. We present the performance of one of the extended schemes and compare the scheme to RED with in/out bit (RIO) in terms of fair bandwidth allocation.

Recently, traffic of real-time communication such as video or IP telephony is increasing in Internet. Low delay and low jitter communication is required to perform these kinds of services smoothly. CoS (Class of Services) scheme which controls traffic by the class unit such as application is currently being paid attention. It is considered to be efficient control mechanism which utilizes characteristics of Internet traffic. One of effective methods for queueing control to realize CoS is CBQ (Class Based Queueing). Efficient and dynamic bandwidth allocation can be realized by using CBQ. However, CBQ mechanism is not necessarily efficient for real-time communication. In this paper, We clarify the effect of dynamic CBQ control scheme on real-time communication by experiment. In addition, we show that low delay and low jitter communication is possible by reducing unnecessary control overhead which is caused by traffic change. We propose new mechanism which changes bandwidth allocation of CBQ dynamically suppressing non real-time traffic and giving priority to real-time traffic on using bandwidth. We also show its efficiency by using real system.

This paper proposes a bandwidth allocation scheme which improves degradation of communication quality due to handoffs in mobile multimedia networks. In general, a multimedia call consists of several component calls. For example, a video phone call consists of a voice call and a video call. In realistic environments, each component call included in one multimedia call may have different requirements for quality-of-service (QoS) from each other, and priorities among these component calls often exist with respect to importance for communications. When the available bandwidth is not enough for a handoff call, the proposed scheme eliminates a low priority component call and defers bandwidth allocation for a component call whose delay related QoS is not strict. Moreover, in the allocation, the scheme gives priority to new calls and handoff calls over a deferred call and also performs bandwidth reallocation to eliminated component calls. By computer simulation, we evaluate the performance such as call dropping probability and show effectiveness of the proposed scheme.

Based on filtering ground clutter power directly in the frequency domain, a new non-coefficient Adaptive MTI (AMTI) scheme is presented in this letter. The results of simulation example show that this scheme has smaller signal-to-noise ratio loss than the classical AMTI based on spectral estimation, as well as high improvement factor.

The recent trend towards highly parallel on-chip data processing, as e.g. in single-chip processors with parallel execution capability of multiple instructions, leads to the requirement of on-chip data storage with high random-access bandwidth, parallel access capability and large capacity. The first two requirements call for the application of multi-ported memories. However, the conventional architecture, based on multi-port storage cells for each bit, cannot efficiently realize the large storage capacity, because cell area explodes due to a quadratic increase with port number (N). A promising method for obtaining area efficiency is to increase the size of the smallest unit with N-port capability, e.g. by introducing N-port capability on the level of blocks of 1-port cells and not for each cell. We report a quantitative analysis of this method for the SRAM case, which is based on design data in a 0.5 µm, 2-metal CMOS technology. Achievable area-reduction magnitudes in comparison to the conventional architecture are found to be enormous and to accelerate as a function of N. Reduction factors to areas < 1/2, < 1/5, < 1/14 and < 1/30 are estimated for 4, 8, 16 and 32 ports, respectively. Since the demerit of the proposed approach is an increased access-rejection probability, a trade-off between area reduction and allowable access-rejection probability is always necessary for practical applications. This is discussed for the application of multi-port cache memories.

In this paper, we investigate the performance of Fast Reservation Protocols (FRP) for burst-level bandwidth allocation in ATM networks. FRP schemes can be classified into delayed transmission (DT) and immediate transmission (IT) methods according to reservation procedure. Moreover, according to the responsibility for negative acknowledgment (NAK) cell generations when burst blocking occurs, FRP schemes can be further classified into blocking node NAK (BNAK) and destination node NAK (DNAK) schemes. We analize the FRP schemes with different reservation and NAK methods for single node and multihop network models, respectively. We then discuss the dependence of performance for each FRP scheme on propagation delay, peak rate, and the number of hops.

A new approach for bandwidth allocation of heterogeneous regulated traffics is proposed for cases of lossless multiplexing and statistical multiplexing under the constraint of maximum delay. Minimum bandwidth required for lossless multiplexing can be expressed as 1-dimensional unconstrained-optimization problem. The corresponding optimality condition gives an optimal bandwidth and buffer requirements. This optimality condition is extended to the case of statistical multiplexing. In case of statistical multiplexing with independent, periodic on-off fluid sources, two random variables representing stochastic worst cases for the amount of traffics generated in an arbitrary time interval are introduced and these are combined optimally. This combined model guarantees the worst case bound for the regulated traffics. Using the proposed approach, bandwidth and buffer requirements of each virtual circuit are obtained simultaneously. As these values are sensitive to multiplexing environments, an effective bandwidth considering trade-off between bandwidth and buffer requirements is introduced. The proposed effective bandwidth can be matched directly with the boundary of the acceptance region, which is less sensitive to mutiplexing environments.

This paper investigates the effectiveness of a network that uses unspecified bit rate (UBR) for a virtual channel (VC) accommodating AAL2 connections. AAL2 is a new ATM adaptation layer that has recently been standardized. Since it is designed to carry low-bit-rate voice signals efficiently, it should be used in the ATM backbone for mobile networks, especially in the IMT-2000 network (International Mobile Telecommunication in the Year 2000 network). Normally, constant-bit-rate (CBR) VCs or variable-bit-rate (VBR) VCs are used to accommodate AAL2 connections. In our previous work, however, we showed that using UBR VCs (equivalent to no VC-level bandwidth management) to accommodate AAL2 connections needs much less VP bandwidth than using CBR or VBR VCs. In this paper, the previous results are extended to the network and the network bandwidth reduction is shown to be larger than that of the virtual path. In addition, the bandwidth reduction achieved by using UBR VCs is comparable with that achieved by introducing AAL2 switching nodes. Based on these results, the core network of the IMT-2000 is discussed.

This paper presents a new design of spread spectrum signals with the minimally sufficient dimension from the view point of frequency diversity. Letting the signature signal duration and the bandwidth be denoted by T and B, respectively, we can nominally represent a signature signal of either Direct Sequence (DS) or MultiCarrier (MC) spread spectrum system as the sum of N=BT sinusoidal signal units with their frequencies separated by 1/T or its multiples. In our design,assuming the maximum expected channel delay spread σd « T as usual, one signature signal viewed in the frequency domain is made up of the minimum number K 2πσdB of sinusoidal signal units which are arranged so as there is placed at least one unit in coherence bandwidth 1/(2πσd) in which the fading channel transfer function has strong correlation. Although the signature signal does not make use of all the units in the given frequency domain as in the ordinary spread spectrum systems, but uses only skipped units, it can be shown that almost the same frequency diversity effect is attained. And it is also shown that the immunity to the external interfering signals is by no means inferior. If every L=N/K T/(2πσd) consecutive sinusoidal signal units are assigned to the K signal units of a signature signal, L different signature signals are simultaneously available mutually orthogonal when the synchronous demodulation is performed in the same T period. We call each of the orthogonal sinusoidal signal sets a Frequency Devision (FD) signal set. Now, CDMA can be independently realized on each of the L FD signal sets provided the operation is synchronous or quasi-synchronous with respect to the symbol demodulation (or signature) period. Partitioning the simultaneous users among the FD sets, it is possible to decrease the number of CDMA users to be processed, retaining the total number of simultaneous users. Owing to this effect, the multiple access performance for the FD/CDMA system is shown to be superior to that of the ordinary DS or MC/CDMA system, assuming matched filter reception based on the complete estimation of the channel characteristics for the both cases. The decrease of the number of CDMA users per FD set makes it practical for the receiver to employ multiple access interference cancellation and even the maximum likelihood detection. Curiously, any FD signal set can be represented in the time domain as L repetition of a sequence with its period equal to K in the number of 1/B duration time chips.

This paper proposes an on-chip memory-path architecture employing the dynamically variable line-size (D-VLS) cache for high performance and low energy consumption. The D-VLS cache exploits the high on-chip memory bandwidth attainable on merged DRAM/logic LSIs by replacing a whole large cache line in one cycle. At the same time, it attempts to avoid frequent evictions by decreasing the cache-line size when programs have poor spatial locality. Activating only on-chip DRAM subarrays corresponding to a replaced cache-line size produces a significant energy reduction. In our simulation, it is observed that our proposed on-chip memory-path architecture, which employs a direct-mapped D-VLS cache, improves the ED (Energy Delay) product by more than 75% over a conventional memory-path model.