Quality of service (QoS) technology has been implemented and started to be applied to new applications on the next-generation Internet. However, as new applications have many kinds of features and requirements, some additional features should be added to current QoS control technology. For example, they require a definition of a single policy to operate QoS control in the management domain consistently and efficiently. Policy definition for transport layer in a domain and among domains is being discussed at IETF to set a standard process, however detailed policy corresponding to the application or contents information according to the application semantics has not been discussed. Therefore we developed QoS policy control mechanism using metadata which is defined as a structured data according to the application semantics. Though metadata and transport mechanism can be located into quite different positions in the concept of network layers, we made them successfully collaborated by defining meta policy. In this paper, we describe our approach to define a meta policy based on the requirements and information contents from the application as a high level layer concept to be able to classify the network behavior. Our approach enables to multiple QoS control and collaboration among domains. We also report of the activities in IETF and ITU-T.

Intensive accesses to a web server causes its response delay and/or service suspension. We propose, to solve the problem in this paper, a service model for web servers grounded upon admission control and call waiting service, which are common approaches to designing telephone systems. Software architecture which consists of a front-end server and child servers is also designed based on the service model. Through performance evaluation for an implementation of the architecture, it is shown that the overall throughput becomes much better by keeping child web servers from overload. We drew a conclusion that the architecture enhances flexibility of a resulting software with discussions from the extensibility view.

The potential network architecture of the emerging carrier class VoIP (Voice over IP) technology for NGN (Next Generation Networks) adopts distributed control architecture to take full advantage of scalability, reliability, flexibility, and interoperability. However, the design of distributed control architecture in the carrier class VoIP network is the state-of-the-art in decentralization and distribution of control. Different configurations of system elements, control scheme of inter system elements communications, signaling protocol, functional partitioning, and scheduling of jobs in call control processing may affect the system performance and QoS (Quality of Service) of MGC (Media Gateway Controller) in carrier class VoIP network. Hence, the modeling of distributed control architecture and its performance analysis are essential issues whenever optimum control architecture has to be determined to meet design requirements. Based on these reasons, this paper proposes several potential network architectures and focuses on the performance study of distributed control architecture in carrier class VoIP network. The SIGTRAN-based distributed control architecture model and the MGCP/MEGACO-based distributed control architecture model are presented. Then, we analyze the SIGTRAN-based distributed control architecture model between MGC and SG (Signaling Gateway) using WRR (Weighted Round Robin) and WF²Q (Worst-case Fair Weighted Fair Queueing) scheduling algorithms respectively. And, we analyze the MGCP/MEGACO-based distributed control architecture model between MGC and MG (Media Gateway) using M/G/1 gating service queueing model. Consequently, the results of performance analysis can be used to evaluate whether the performance of distributed control architecture model can meet the requirement of planning and design for carrier class VoIP network deployment.

Most of the QoS-based routing schemes proposed so far focus on improving the performance of individual service classes. In a multi-class network where high priority QoS traffic coexists with best-effort traffic, routing decision for QoS sessions will have an effect on lower ones. A mechanism that allows dynamic sharing of link resources among multiple classes of traffic is needed. In this paper we propose a multi-class routing algorithm based on inter-class sharing resources among multiple class of traffic. Our algorithm is based on the concept of "the virtual residual bandwidth," which is derived from the real residual bandwidth. The virtual residual bandwidth is greater than the residual bandwidth when the load of lower priority traffic is light, and smaller when the load of lower priority traffic is heavy. The idea of our approach is simple since the routing algorithm for individual traffic doesn't change, but the only change is the definition of the link cost. We demonstrate through some extensive simulations the effectiveness of our approach when the best effort distribution is uneven and when its load is heavy. Also better performance is noticed when using topology with large number of alternative paths.

Differentiated Service (DiffServ) is a technology designed to provide Quality of Service (QoS) in the Internet, and is superior to Integrated Service (IntServ) technology with respect to the simplicity of its architecture and the scalability of networks. Although various simulation studies and estimations over testbeds have investigated the QoS that is offered via the DiffServ framework, almost all of them focused on the characteristics in a single DiffServ domain. However, the Internet is actually composed of a large number of AS domains, and thus packets are very likely to arrive at their destinations after going through many different domains. Therefore, we have analyzed the QoS performance in a model consisting of multiple DiffServ domains, and focused especially on the quality provided by Assured Forwarding Service (AF) to achieve statistical bandwidth allocation with AF-PHB. Our simulation results show some throughput characteristics of flows over multiple Diffserv domains, which clarify how network configurations impact the QoS over multiple DiffServ domains.

This paper presents the basic design and principles of an FQDN-based Internet Architecture, where a host is identified in the Internet only by its FQDN (Fully Qualified Domain Name). The address shortage problem for the IPv4 paradigm has been marginally solved by introducing private addresses within the intranet and also by DHCP for public Internet access services. These two approaches have been independently developed. Considering more elaborate peer-to-peer communications such as voice over IP (VoIP) in the future Internet, the shortage of address space will become a serious problem for a call or session to be established. To alleviate this, this paper proposes an FQDN-based Internet architecture, assuming that a global endpoint identifier (EID) in the network is FQDN instead of an IP address. Since a countably infinite number of addresses can be logically produced by FQDN, this system resolves the address shortage problem assuming a caller uses the FQDN of the called terminal or host. This requires a dynamic address loading system from the FQDN of the called terminal into one of the available public IP addresses. After proposing a possible address loading system, some mathematical results on the required number of public addresses for VoIP traffic, the address space size of the derived Internet, etc. are also presented.

As the number of Internet applications in critical situations increases, the quality assurance of the network infrastructure becomes more and more important. The operators of such networks have to maintain the network as reliable and available with sufficient performance. Measurement technology is the key to maintaining the network condition and provides the network as being the important infrastructure of the network society. To meet these requirements, a high-speed and accurate IP meter, HIM, has been developed. Specific hardware has been made to realize the necessary specifications that enable data capturing with 20-µsecond order timing information from a gigabit class network. The experimental results with HIM show that: 1) software-based equipment cannot handle 100-µsecond order timing information, 2) the current computer system can transmit a high-speed (30 Mpbs) multimedia stream in accurate timing (jitter less than 20 µseconds) with UDP, 3) but TCP's retransmission mechanism makes the jitter worse (up to 264 mil seconds) and, 4) WAN's one-way traffic exhibits some strange behavior, which requires further investigation.

Providing a fair allocation of bandwidth among different connections over the Internet without affecting link efficiency has been a challenging issue in the area of network performance improvement. Congestion signaling is essential for the purpose. Conventional TCP uses packet loss as an implicit indication of congestion. Several enhancements to TCP have been proposed for faster congestion recovery and thus to improve the network performance. However, packet loss reduces TCP goodput and adds large delay. Also the variance in the share of bandwidth obtained by each connection may become unaccepatbly high. To the contrary, Explicit Congestion Notification (ECN) indicates a congestion explicitly before it actually occurs. Therefore, ECN facilitates a faster congestion detection and contributes to the network performance improvement. In this paper, we consider the performance implications of employing different ECN strategies along with several TCP enhancement schemes. We also introduce a new ECN packet marking strategy FIM and evaluate its relative performance and suitability for deployment along with different TCP enhancements. Simulation results show the superiority of FIM over other existing marking strategies with different TCP enhancement schemes by providing the best fairness without hampering link efficiency. We also observe FIM maintains a more consistent delay bound than other strategies and as such, is more suitable for application in practical purposes.

In this paper, we implemented practical resource adaptation mechanism for broadband application environment especially suitable for portable computers (note PCs) with limited power supply. Continuous development of portability enhancement and increasing computation with less power consumption for battery operating environment in note PC equips with intelligent resource controlling mechanism embedded inside the hardware. In such environment, where occasional environmental changes take in place rapidly, adaptive and collaborative controls of resources are required. Optimization of resource based on power supply management is dedicated to the operating environment. Bi-directional, user, application, operating system and device aware interface for resource configuration/management is extended to the current Unix operating system by implementing application programming interface of Advanced Configuration and Power Interface (ACPI). Traditional operating system conceals the active device probing and controlling method autonomously. Implementation of "State handler daemon" interface enables applications and application users to monitor and adapt the resources available during the operation. Thick collaboration between application and devices that are reserved in its limited environments economizes the consuming device utilization resulting: for example, life extension of battery life and effective network bandwidth adaptation. We focused on DVTS (Digital Video Transport System) as a typical broadband application. Evaluation shows 125 minutes of continuous battery operation with battery life extension compared to 70 minutes with traditional systems. Collaboration with network device management enables 100% packet transfer compared to 89% without any Resource adaptations.

The Fat-Btree which is a new parallel B-tree structure has been proposed to improve the access performance of shared-nothing parallel database systems. Since the Fat-Btree has only a part of index nodes on each processing element, it can reduce the synchronization cost in update operations. For these reasons, both retrieval and update operations can be processed at high throughput compared to previously proposed parallel B-tree structures for shared-nothing computers. Though we tried to apply some conventional concurrency control methods to the Fat-Btree, e.g., B-OPT and ARIES/IM, which were designed for shared-everything machines, we found that these methods are not always appropriate for the Fat-Btree. In this paper, it is shown that the conventional methods are not suitable for the Fat-Btree and other parallel B-trees. We propose a new deadlock free concurrency control protocol, named INC-OPT, to improve the performance of the Fat-Btree more effectively than the B-OPT and ARIES/IM. Furthermore, in order to prove that the Fat-Btree provides the impact on the performance of shared-nothing parallel databases, we compare the real performance of three types of parallel B-tree structures, Fat-Btree, Copy-Whole-Btree, and Single-Index-Btree, on an nCUBE3 machine where the INC-OPT is applied.

Caching of frequently accessed data has been shown to be a useful technique for reducing congestion on the narrow bandwidth of wireless channels. However, traditional client/server strategies for supporting transactional cache consistency, which require extensive communications between a client and a server, are not appropriate in a wireless mobile database. This paper proposes two, simple but effective, transactional cache consistency protocols for mobile read-only transactions by utilizing the broadcast-based solutions for the problem of invalidating caches. The novelty of our approach is that the consistency check on accessed data and the commitment protocol are implemented in a truly distributed fashion as an integral part of cache invalidation process. The applicability of proposed techniques is also examined by an analytical study.

In this paper, a new voice personality transformation algorithm which uses the vocal tract characteristics and pitch period as feature parameters is proposed. The vocal tract transfer function is divided into time-invariant and time-varying parts. Conversion rules for the time-varying part are constructed by the classified-linear transformation matrix based on soft-clustering techniques for LPC cepstrum expressed in KL (Karhunen-Loève) coefficients. An excitation signal containing prosodic information is transformed by average pitch ratio. In order to improve the naturalness, transformation on the excitation signal is separately applied to voiced and unvoiced bands to preserve the overall spectral structure. Objective tests show that the distance between the LPC cepstrum of a target speaker and that of the speech synthesized using the proposed method is reduced by about 70% compared with the distance between the target speaker's LPC cepstrum and the source speaker's. Also, subjective listening tests show that 60-70% of listeners identify the transformed speech as the target speaker's.

This paper first presents robust algorithms to extract invariants of the linear Lie algebra model from 3D objects. In particular, an extended 3D Hough transform is presented to extract accurate estimates of the normal vectors. The Least square fitting is used to find normal vectors and representation matrices. Then an algorithm of segmentation for 3D objects is shown using the invariants of the linear Lie algebra. Distributions of invariants, both in the invariant space and on the object surface, are used for clustering and edge detection.

This paper introduces an automatic and multi-instrument snowfall observation system and proposes techniques that could be used in the estimation of snowfall characteristics. The instruments used in this study include two microwave radars, an optical lidar, a CCD camera based imaging system and high-accuracy electrical balances for reference data. The emphasis has been on obtaining good temporal resolution and synchronization accuracy of separate datasets. In most research done so far, this has not been a principal point, either because only very long snowfall events have been measured, or wide area estimates were desired, or due to limitations in manual sampling methods and other technical issues. The measurements were also contained in a small area to make sure that all instruments record data from the same target. One radar and the optical lidar recorded an atmospheric profile up to 6000 m, while the other radar, the imaging system and the two balances recorded snowfall on the ground level. The combination of optical, microwave and direct visual observations of snowfall show that a change in cloud conditions can result in snowfall having different characteristics. The lidar backscatter was used as main indicator of transitions in cloud conditions. A direct visual evaluation of snowflake size distribution using a CCD camera shows that it is extremely helpful in order to interpret radar data. The camera observed velocity distribution showed no large variations between snowfall events, however, it could be useful in detecting graupel and hail precipitations which have much faster terminal velocities. This paper will conclude with a discussion on further elaborating the use of lidar and visual data to complement radar observations of snowfall.

A quasi-periodic signal is a periodic signal with period and amplitude variations. Several physiological signals, including the electrocardiogram (ECG), can be treated as quasi-periodic. Vector quantization (VQ) is a valuable and universal tool for signal compression. However, compressing quasi-periodic signals using VQ presents several problems. First, a pre-trained codebook has little adaptation to signal variations, resulting in no quality control of reconstructed signals. Secondly, the periodicity of the signal causes data redundancy in the codebook, where many codevectors are highly correlated. These two problems are solved by the proposed codebook replenishment VQ (CRVQ) scheme based on a bar-shaped (BS) codebook structure. In the CRVQ, codevectors can be updated online according to signal variations, and the quality of reconstructed signals can be specified. With the BS codebook structure, the codebook redundancy is reduced significantly and great codebook storage space is saved; moreover variable-dimension (VD) codevectors can be used to minimize the coding bit rate subject to a distortion constraint. The theoretic rationale and implementation scheme of the VD-CRVQ is given. The ECG data from the MIT/BIH arrhythmic database are tested, and the result is substantially better than that of using other VQ compression methods.

Recently there has been increased attention to the causality among biomedical signals. The causality between brain structures involved in the generation of alpha activity is examined based on EEG signals acquired simultaneously in the frontal and occipital regions of the scalp. The concept of directed coherence (DC) is introduced as a means of resolving two-signal observations into the constituent components of original signals, the interaction between signals and the influence of one signal source on the other, through autoregressive modeling. The technique was applied to EEG recorded from 11 normal subjects with eyes closed. Through an analysis of the directed coherence, it was found that in both the left and right hemispheres, alpha rhythms with relatively low frequency had a significantly higher correlation in the frontal-occipital direction than in the opposite direction. In the upper alpha frequency band, a significantly higher DC was observed in the occipital-frontal direction, and the right-left DC in the occipital area was consistently higher. The activity of rhythms near 10 Hz was widespread. These results suggest that there is a difference in the genesis and the structure of information transmission in the lower and upper band, and for 10-Hz alpha waves.

The format conversion operations between a floating-point number and an integer number and a round operation are the important standard floating-point operations. In most cases, these operations are implemented by adding additional hardware to the floating-point adder. The SR (simultaneous rounding) method, one of the techniques used to improve the performance of the floating-point adder, can perform addition and rounding operations at the same stage and is an efficient method with respect to the silicon area and its performance. In this paper, a hardware model to execute CRops (conversion and rounding operations) for the SR floating-point adder is presented and CRops are analyzed on the proposed hardware model. Implementation details are also discussed. The proposed scheme can maintain the advantages of the SR method and can perform each CRop with three pipeline stages.

This letter addresses a neural network (NN)-based predictor for the LP (Linear Prediction) residual. A new NN predictor takes into consideration not only prediction error but also quantization effects. To increase robustness against the quantization noise of the nonlinear prediction residual, a constrained back propagation learning algorithm, which satisfies a Kuhn-Tucker inequality condition is proposed. Preliminary results indicate that the prediction gain of the proposed NN predictor was not seriously decreased even when the constrained optimization algorithm was employed.

This paper proposes a new method of retrieving images from large image databases. The method is based on VQ (Vector Quantization) of local texture features at interest points automatically detected in an image. The texture features are extracted by Gabor wavelet filter bank, and rearranged for rotation. These features are classified by VQ and then construct a pattern histogram. Retrievals are performed by just comparing pattern histograms between images.