We propose an analytical model to evaluate the performance of slotted ALOHA with NOMA for mMTC under the assumption that base stations and devices are distributed according to mutually independent stationary point processes. For NOMA, we consider a simple scenario in which perfect interference cancellation can be conducted at each base station. As performance metrics, we focus on the probability of a device successfully transmitting data to the nearest base station (transmission success probability) and the expected number of devices from which a base station can correctly decode the signals (throughput of base station). We establish analytical expressions for the two performance metrics for three schemes: simple slotted ALOHA, slotted ALOHA with interference cancellation, and slotted ALOHA with power control. Through several numerical experiments, we show that the application of interference cancellation improves the throughput by up to 20%; however, the application of interference cancellation does not solve the near-far problem. We also show that the application of power control solves the near-far problem but significantly degrades the performance of slotted ALOHA.

Accurate sensor localization in wireless sensor networks (WSNs) is essential for context-aware applications, such as vehicle detection and environmental monitoring in smart cities. In densely deployed WSNs, spatial proximal sensor measurements are known to be highly correlated. Herein, we propose a sensor localization method based on the similarity of sensor measurements that simultaneously updates locations and intersensor distances by leveraging this similarity–distance relationship. The intersensor distance is modeled as a similarity-conditioned Gaussian distribution. Given an initial set of sensor locations, observed pairwise distances are used to update distribution parameters using maximum likelihood estimation and curve fitting. Subsequently, sensor locations are estimated by solving a nonlinear optimization problem. The proposed method establishes a self-consistent iterative process in which sensor locations and distance distributions are jointly updated using a statistical model. The simulation results indicate that the proposed method reduces the average localization error by up to 86% compared with that of initial locations when the sensor detection radius is 3.0–4.0 m. Even if the distance estimation contains errors, the iterative process improves localization accuracy by leveraging similarity information that does not directly represent distance. The proposed method eliminates the need for explicit ranging or threshold tuning, making it well-suited for dense sensor deployments. As a result, it expands the practical potential of densely deployed low-cost WSNs for large-scale environmental monitoring applications.

We propose an analytical model to evaluate the performance of NOMA-based slotted ALOHA for mMTC under the assumption that base stations and devices are distributed according to mutually independent stationary point processes. For the NOMA, we consider a simple scenario in which perfect interference cancellation can be conducted at each base station. As performance metrics, we focus on the probability for a device successfully transmitting data to the nearest base station (transmission success probability) and the expected number of devices from which a base station can correctly decode the signals (throughput of a base station). We establish a universal relationship between these two performance metrics that holds for general stationary point processes, including Poisson point processes. Using this universal relationship, we derive analytical expressions for the two metrics for the case in which the base stations and devices are distributed according to mutually independent Poisson point processes. Numerical experiments using simulations show that throughput can be improved by up to 15% compared to the case where interference cancellation is not applied.

It is known that a Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA)-based wireless local area network (WLAN) may exhibit bistable behavior in a scenario where wireless stations send frames at a constant rate to the access point and the total offered load provided by the wireless stations is approximately equal to the WLAN capacity. This paper proposes an analytical framework for examining such bistable behavior. In the model, frames are generated according to a stationary arrival process and are queued in the buffer inside each station. Each station sends the frames in its buffer to the access point via a common channel. The interaction between the frame queues of the stations causes an interdependence of the frame processing time among stations, which in turn causes bistable behavior of the WLAN. The bistability of WLANs analytically derived using the proposed framework is consistent with that obtained in computer simulations.

COVID-19 digital contact tracing applications for smartphones have become popular worldwide to reduce the effects of the pandemic. We considered that contact information between smartphones used in these applications can be used for the indoor localization of pedestrians. In this paper, we propose two indoor pedestrian localization methods based on contact information obtained from Bluetooth low energy (BLE) beacons installed in pedestrian’s smartphones. Proposed method 1 is multilateration, and proposed method 2 solves a nonlinear optimization problem to further improve the accuracy of method 1. These two proposed methods comprise three steps: (1) the smartphones and anchor nodes recognize the proximity relationship with neighbor nodes using BLE signals transmitted from other smartphones and anchor nodes. The recognized proximity relationship is sent to a server. (2) The server estimates the distance between each node (smartphone or anchor node) from the proximity relationship. (3) The positions of smartphones are estimated based on the distance between nodes estimated by the server. We verified the localization accuracy of the proposed methods through simulation experiments. In an indoor area of 15 m × 30 m, the average localization error of the proposed method 2 was 0.74 m when the pedestrian density was 0.5 /m2.

We propose a method for localizing sensors based solely on the non-geotagged data that is collected by the sensors themselves. In the proposed method, the localization is performed in two steps. In the first step, proximity relationships among sensors are identified based on the sensor-collected data. In this step, we use the correlation coefficient between the data collected by two sensors as an indicator of the closeness of those sensors. In the second step, the sensors are localized based on identified proximity information by applying a proximity-based localization method [1]. To investigate the effectiveness of the proposed method, we focus on a specific situation in which sensors for detecting target objects are deployed within an observation area. Rather than location data, each sensor transmits an identifier, such as its media access control (MAC) address, when it detects an object nearby. We also report on the results of simulation experiments that demonstrate that the proposed method can estimate sensor locations accurately in the envisioned situation and show that area information, such as the number of target objects, can be obtained based on the estimated locations of sensors responding to those objects.

We investigate the analytical expressions of typical performance metrics, namely message reception probability, message reception interval, and information dissemination rate, for vehicle-to-everything (V2X) safety communication in urban scenarios. Each of the performance metrics can be expressed in a simple closed form under some moderate assumptions such as vehicles being positioned according to a Poisson process in a two-dimensional plane. Based on the obtained expressions, we analyze the dependence of these performance metrics on the message transmission interval, which is a typical parameter of congestion control of V2X communication. The results show that the optimal message transmission interval that minimizes the average message reception interval can be derived in a simple closed form. The information dissemination rate is insensitive to the message transmission interval.

In image-based ray tracing, the possible permutations of reflectors (walls, doors, and sides of pieces of furniture) are all chosen to exhaustively generate images of a signal source and, for each generated image, the existence of an actual ray path that leaves the image and reaches the destination is investigated. In a typical indoor environment, where the reflectors are arranged parallel or perpendicular to each other, permutations of reflectors generate many image duplicates; that is, different permutations of reflectors generate the same image. Focusing on this fact, this paper proposes a novel technique for accelerating the image-based ray tracing. The key idea of the proposal is to choose a limited set of the permutations of reflectors to generate necessary and sufficient images without duplication. The proposed technique avoids the redundant ray-path search to reduces the run time of ray tracing without degrading the accuracy. The simulation experiments verify that the proposed technique makes the ray tracing much less demanding of computation.

We propose a light-weight performance analysis of Wi-Fi offload using a simple Markov model. Although the proposed model does not describe the behavior of the contention window of each station in detail, it allows us to analyze the non-saturation throughput while considering interaction among interfering stations and the queueing behavior of each station. In addition to this, the proposed model does not need to use the assumption that frames arrive at relay station according to a Poisson process. In this paper, we show how to apply the mean field approximation to the model to significantly reduce the amount of computation. We show that the proposed model accurately estimates the throughput performance of the Wi-Fi offload, and that the mean-field approximation is still valid in the scenario of the Wi-Fi offload.

This paper proposes an algorithm for counting the number of distinct targets using binary proximity sensors randomly deployed for target detection. This paper focuses on the fact that sensors simultaneously detecting a target should be in close proximity to each other. Thus, if multiple targets exist in a region of watch, sensors detecting them at a specific time are divided into several clusters, and each cluster represents one target. Based on such consideration, this paper uses the number of clusters of target-detecting sensors as an estimator of the number of distinct targets. To find clusters of sensors, the notion of mutual nearest neighbor (MNN) is employed. The MNN is a set of target-detecting sensors, where the largest distance within the set is smaller than the distance to any target-detecting sensor outside the set. The notion of the MNN yields a polynomial-time algorithm for partitioning a set of target-detecting sensors into several clusters. Simulation experiments verify that the proposed algorithm gives very precise estimates of the number of distinct targets if sensors are densely deployed.

We propose an algorithm for the connectivity-based sensor localization. We formulate the sensor localization into a variant of the classical MDS, in which the squared error between the measured inter-sensor distances and the calculated ones based on estimated sensor locations is used as a stress function. The proposed formulation is applicable even if the distances of several sensor pairs are unknown, and thus it does not require the minimum-hop-count assumption to estimate all inter-sensor distances. This feature is suitable especially for anisotropic networks, in which the minimum-hop-count assumption does not hold in general. We propose a technique, referred to as stress relaxation in this paper, which allows us to get a good solution while avoiding local minimums of the stress function. Simulation experiments verify that the proposed algorithm significantly outperforms the localization scheme based on the classical MDS.

We propose a method for estimating the size, perimeter length, and location of a time-variant target object moving in a monitored area. This method uses only binary information from sensors of which locations are unknown, where the binary information includes whether each sensor has detected the target object or not. Analysis based on integral geometry provides the relationship between the number of sensors detecting the target object and the target object parameters to be estimated. Because this relationship is linear, a linear filter, such as the Kalman filter, is applicable to estimate parameters if we can assume that the dynamics of the parameters are linear. As a concrete example, the size, shape, and location of an active thunder area is estimated. The model discussed in this paper is applicable as a model of participatory sensing.

Random walks are the first choice for gathering random samples from communication networks. In its simplest form, which we will call classical random-walk (RW) sampling, random walkers select an outgoing link with equal probability to visit an adjacent node. Although classical RW sampling is preferable because of its simplicity, it does not perform uniform sampling when the degree of the network nodes is not uniformly distributed, in that case, it introduces a large amount of bias to the nodes of large degree. We propose a technique that uses an idea from probability theory, change of measure, to correct the bias in the data obtained by classical RW sampling. For this purpose, we introduce two probability measures defined on a common space: one is a probability measure that characterizes the observation by using a uniformly sampled node, and the other characterizes the observation by using a node sampled by classical RW sampling. Using the relationship between the two probability measures, the statistics for the data from the uniform sampling can be inferred from the statistics for the data from the classical RW sampling. We present a mathematical framework for the above-mentioned approach and derive several formulas for transforming between the two types of statistics. Simulation experiments based on the data of real networks verify the effectiveness of this proposal.

This paper proposes an algorithm for counting the number of distinct targets based on the responses of binary proximity sensors. A basic idea of the proposal is that sensors detecting targets at a specific time should be partitioned into several clusters, where each cluster is composed of sensors detecting a common target. Thus, the number of clusters would be a natural choice of the estimator for the target counting. This paper presents a mathematical framework and a simple algorithm for partitioning a set of target-detecting sensors into several clusters. Simulation experiments verify that the proposed algorithm gives precise estimates of the number of distinct targets.

Theoretical analysis of IEEE802.11e enhanced distributed channel access (EDCA) mechanism is now a mature research theme, receiving considerable attention. However, most of these works depend on per-slot status, namely probability of transmission attempt and collisions at each slot. Tinnirello et al. [1] present an analytic model which does not rely on per-slot status under a saturated traffic assumption, whereas typical network conditions are non-saturated and heterogeneous. We present an extension of their model to a non-saturated environment. Through the comparison with the simulation results by ns2, we confirm that the proposed model accurately estimate the throughput performance of the EDCA-based WLANs where voice and best effort traffic are multiplexed together.

Spatial relations between sensors placed for target detection can be inferred from the responses of individual sensors to target objects. Using this fact, we propose a sensor localization method to estimate the relative locations of sensors based on the responses of the sensors to moving objects. We let several objects with different sizes move along various paths one by one within the field of watch, where sensors are deployed. When two sensors concurrently detect each of the objects, we assume that the distance between the two sensors is equal to a given constant. Based on the distances between sensors estimated from sensors' responses to the objects, the relative locations of sensors are estimated by a technique of the cooperative localization. Through several simulation experiments, we find that the proposed algorithm yields surprisingly accurate estimate of the relative locations of sensors.

Cooperative localization, where sensors exchange the information with each other to determine their locations, has received considerable attention. In this work, we study the cooperative localization in order to investigate several fundamental properties that have not been well addressed so far. We formulate the cooperative localization in a general setting, where a relative or absolute location map is obtained, depending on the number of anchors. The (relative or absolute) location map is the output of an optimization problem, where the objective function is given as a norm of a space where a vector composed of distances between sensors is defined. We show that several error bounds and the estimation bias of the cooperative localization can be obtained by simple arguments (e.g. by using triangle inequality) without specifying the detail of the objective function. Next, we theoretically and numerically verify that the cooperative localization has a preferable scaling property such that the estimation becomes more accurate as sensors are more densely deployed. Finally, we consider the problem that the objective functions used in the cooperative localization are usually multimodal and have a number of local optima and saddle points. We show that the gradient descent algorithm starting from a random prior (initial estimates) often fails to find the optimal solution when the distance measurements between some pair of sensors are not available. We propose a new prior, called shortest-path-distance-based prior, which is very powerful for obtaining accurate estimates even when the distances between some sensor pairs are not measurable.

Some applications of sensor networks do not require the knowledge of absolute locations of sensors; knowledge of relative locations of sensors is sufficient. We study the problem of estimating relative locations of sensors, which we refer to as relative localization, under the assumption that each sensor can measure the distances between neighbor sensors. We show that the relative localization can be performed by setting up a non-linear optimization problem and solving it by standard optimization techniques such as the steepest descent method. We theoretically and numerically investigate the characteristics of the proposed relative localization and obtain several findings. For example, the relative locations of sensors can be accurately estimated even if the distance measurements include large errors. Thus, primitive distance measurement techniques, including RSSI-based measurements, are applicable for the relative localization. We also find that the relative localization has a preferable scaling property; it performs better as the number of sensors increases.

IEEE 802.11e Enhanced Distributed Channel Access (EDCA) is a MAC protocol for WLAN to realize Quality of Service (QoS) control. EDCA offers the possibility of defining four different classes of service at the MAC layer so that QoS requirements can be supported. Service between classes is differentiated by different sets of channel contention parameters. However, it is difficult to guarantee QoS of the high-priority stations in EDCA where both EDCA stations and Distributed Coordination Function (DCF) stations are connected to an access point (AP), because quality control is not considered in DCF. This paper proposes a novel scheme to guarantee the QoS requirement of the high-priority stations in EDCA in such a mixed environment. In this scheme, Receiving-Opportunity Control in MAC Frame (ROC) proposed by the authors is employed. ROC probabilistically refuses data frames from low-priority stations. This paper analyzes the appropriate probability which satisfies the requirement of high-priority stations in EDCA. The effectiveness of the proposed scheme is validated through simulation.

We propose an algorithm for shape recognition of the target object using the notion of the inner distance. The inner distance of the object going across the field, where binary sensors are densely deployed, are obtained from the distance between two binary sensors concurrently detecting the object. To estimate the position of each binary sensor, we apply a range-free localization using a moving landmark, which wanders around the field of watch. We numerically study the accuracy of the shape recognition based on the inner distance when it is combinedly used with the range-free sensor localization based on the moving landmark.

We propose to use a buffer management scheme, called adaptive pushout, in the transmission buffer of the access point (AP) in order to solve unfairness problems between competing TCP flows over wireless LANs. The adaptive pushout comes from the observation that multiple ACK segments in a TCP connection are redundant because of the cumulative acknowledgment mechanism of the TCP. Under the adaptive pushout, an arriving packet encountering the full buffer removes one of redundant ACKs from the transmission buffer of the AP to enter there. When the buffer is full and has no redundant ACK, the arriving packet adaptively chooses one of packets in the buffer for pushing it out from the buffer. The choice of the packet to be pushed out depends on the number of uplink and downlink TCP flows, which are continuously monitored in the proposal. We have conducted simulation experiments to show that the adaptive pushout greatly improves the fairness compared with existing proposals.

We propose a method for estimating parameters of multiple target objects by using networked binary sensors whose locations are unknown. These target objects may have different parameters, such as size and perimeter length. Each sensors, which is incapable of monitoring the target object's parameters, sends only binary data describing whether or not it detects target objects coming into, moving around, or leaving the sensing area at every moment. We previously developed a parameter estimation method for a single target object. However, a straight-forward extension of this method is not applicable for estimating multiple heterogeneous target objects. This is because a networked binary sensor at an unknown location cannot provide information that distinguishes individual target objects, but it can provide information on the total perimeter length and size of multiple target objects. Therefore, we propose composite sensor nodes with multiple sensors in a predetermined layout for obtaining additional information for estimating the parameter of each target object. As an example of a composite sensor node, we consider a two-sensor composite sensor node, which consists of two sensors, one at each of the two end points of a line segment of known length. For the two-sensor composite sensor node, measures are derived such as the two sensors detecting target objects. These derived measures are the basis for identifying the shape of each target object among a given set of categories (for example, disks and rectangles) and estimating parameters such as the radius and lengths of two sides of each target object. Numerical examples demonstrate that networked composite sensor nodes consisting of two binary sensors enable us to estimate the parameters of target objects.

The purpose of our study is to guarantee the required throughput of high-priority (UDP) flows and to achieve the stable throughput for each flow in a wireless LAN in which both high-priority and low-priority (TCP) flows exist. The authors previously proposed a QoS control scheme called Receiving Opportunity Control in MAC Frame (ROC), in which an Access Point (AP) intentionally does not return an ACK for low-priority data frames. In this paper, we propose a novel QoS control scheme using Token Bucket Filter (TBF) based ROC, which achieves the above function. The TBF mechanism is known as a basic token based traffic control. The simulation results show the effectiveness of the proposed scheme.

We propose a random-walk-based file search for unstructured P2P networks. In the proposal, each node keeps two pieces of information, one is on the hop-limited shortest path tree rooted at itself and the other is on the indexes of files owned by neighbor nodes, referred to as the file list. A random-walk search is conducted along the concatenation of hop limited shortest path trees. To find a file, a node first checks its file list. If the requested file is found in the list, the node sends the file request message to the file owner, otherwise, it sends a file-search message to a randomly-selected leaf node on the hop-limited shortest path trees. Numerical examples show that our proposal is much more efficient than the normal random-walk search, while it waists much less network bandwidth than the flooding (network-broadcast) based search.

An estimation method for the shape and size of a target object by using networked binary sensors whose locations are unknown is proposed. Each of those sensors, which individually are incapable of monitoring the target object's shape and size, sends only binary data describing whether or not it detects at every moment the target object. By using these data and the explicit formulas derived in this paper, we can estimate the size and the perimeter length of the target object even without sensor location information. An additional parameter, which is a function of the shape of the target object, can also be estimated when the target object is non-convex.

In this article, we present a mathematical model for evaluating the throughput of IEEE 802.11 MAC layer as well as the network-layer performance such as queueing delay or packet loss at each station. Our proposal combines a Markov chain model for evaluating the throughput of the IEEE 802.11 DCF and a queueing model for analyzing the network-layer performance of each station. Through extensive simulations using the simulator ns2, we show that our model can accurately predict how many voice conversations can be multiplexed over a WLAN without loss of any packets. Our model allows us to evaluate the IEEE 802.11 DCF with contention window (CW) differentiation, which is a service differentiation scheme provided by EDCA. In this article, we also investigate how much the CW differentiation could improve the bandwidth utilization by making contention window of the access point smaller than wireless stations.

We analyze the path coverage property of a sensor network, where a large number of sensors are randomly deployed. We characterize the path coverage in terms of three metrics:fractionofcoverage,probabilityofcompletecoverage,andprobabilityofpartialcoverage. We derive the expressions for the three coverage metrics as functions of the sensor density, the sensing range of a sensor, and the communication range of a sensor. Based on the derived expressions, we study the impact of the communication range on the path coverage and, for example, show that a sensor should have a long wireless-communication range to obtain large detectability of intruding objects (partial coverage probability of the trajectories of objects) with modest density of sensors.

We study the fundamental properties of the hybrid combination of optical circuit switching (OCS) and electrical packet switching (EPS) used in an ultra-high-speed local area network (LAN) of the future, where chunks of data such as those containing non-compressed super-high-definition images are frequently transmitted as well as text based data. We assume the logical topology of the OCS network in such a LAN to be single star in which the end-to-end connections are provided through a non-blocking optical matrix switch located at the center, while a conventional modest-speed EPS based network such as Ethernet is co-installed to function as another data plane as well as the control plane of the OCS network. We refer to queueing theory in order to understand the fundamental performances of OCS and packet switching, and show that the hybrid use of OCS and EPS is very preferable especially when the variance of file size is very large. Simulation results show that the hybrid configuration can significantly outperform a network based on OCS only.

We study the connectivity in vehicular ad-hoc networks, where the motion of vehicles is constrained on a lattice-shaped road network. First, we theoretically investigate the connectivity under the Poisson-positioning assumption, where vehicles are positioned according to a Poisson process on each road at any arbitrary instants. We find that the Poisson-positioning assumption allows the existence of the finite critical-vehicle density; that is, if (and only if) the density of vehicles is greater than the finite critical density, then there exists a large (theoretically infinite) cluster of vehicles and an arbitrary pair of vehicles in the set is connected in single or multiple hops. We obtain an analytical expression for the critical density as a function of the transmission range of each vehicle and the distance between intersections. Next, we consider the connectivity under more realistic movement patterns of vehicles where the Poisson-positioning assumption does not hold. We numerically find that, even in non-Poisson-positioning cases, there exists the critical vehicle density. The critical density in non-Poisson-positioning cases is, however, larger than the one under the Poisson-positioning assumption. We also gain some insight on the efficiency of roadside-relay-station deployment to provide better connectivity between vehicles.

This paper discusses networked sensors that are randomly distributed to monitor and detect a target object moving on a trajectory. On the basis of analysis results, we propose a method for estimating the shape and size of a target object by using sensors whose locations are unknown. It is surprising that sensors that individually are incapable of monitoring the target object's shape can be used to estimate it without sensor location information. An example shows that a combination of sensors with different sensing areas can increase the shape estimation performance. This estimation method can work for any trajectory and with sensor areas of arbitrary shape and size.

We propose a new finite population model for cellular mobile systems with traveling users. In this model, mobile users arrive according to a Poisson process from outside the system, independently travel in the system, and leave the system in due time. A mobile user is either in call (active) or out of call (inactive). We find that if two minor modifications are made to the model, the joint distribution of the number of calls in progress in each cell has the product form. Making the two modifications is referred to as product-form approximation in this paper. Under the product-form approximation, the call blocking probability or the hand-off blocking probability is given by the Erlang-loss formula. We evaluate the accuracy of the product-form approximation through several simulation experiments, and find that the Erlang-loss formula gives accurate estimates of call-blocking probability in most of cases. Our work would give theoretical justification for applying the Erlang-loss formula to the performance analysis or channel provisioning of cellular mobile systems.

This work proposes a network model to understand the scale-free property of directed networks. The proposed model assigns two intrinsic variables (incoming and outgoing weights) to every node. A directed link is established from node ;' to node j if the sum of outgoing weight of node ;' and the incoming weight of node j exceeds a predetermined threshold. The proposed model allows us to know the exact analytical expressions for degree distributions and clustering. For example, the in-degree and out-degree distributions have power-law tails and their scaling exponents are controllable within the range (1,infin). The average clustering coefficient of nodes with out-degree (or in-degree) n also has a power-low tail as a function of n. We also find that the scaling exponent of the clustering coefficient depends on the correlation between incoming and outgoing weights.

We propose a packet sampling strategy called fixed-period sampling, which selects at most one packet in every fixed-length period. Under the fixed-period sampling, the number of flow-cache lookups during a unit of time or the number of entries in a flow cache is bounded by a constant, which is simply expressed by a few tuning parameters. As an application of the fixed-period sampling, we also focus on the flow-rate estimation from fixed-period sampled packet streams. In particular, we propose a simple estimator based solely on the sampling frequency overestimates flow rates of detected flows. We have conducted simulation experiments using two real traces to show basic characteristics of the fixed-period sampling for the comparison with the fixed-period sampling. We also show the accuracy of the proposed flow-rate estimator through simulations.

We propose a technique for reducing the number of message duplicates during message flooding in wired networks. The key feature of our proposal is that each node keeps the partial information of shortest path trees whose roots are in its neighborhood. When receiving the flooding message, each node generates its duplicates and forwards them to a subset of neighbors, which are on the partial shortest path tree rooted at the message source. The partial information on shortest path trees is stored in message forwarding table of each node. We show that the partial shortest path tree can be constructed in a fully-distributed manner by simply using dummy message flooding. Our proposal can largely reduce the number of message duplicates while it guarantees the full reachability and keeps the time to reach the same as that in the full flooding. Duplicate reduction effect of our proposal is theoretically evaluated and numerically examined by simulation experiments.

It is generally believed that IntServ architecture is more suitable for strict QoS guarantees than DiffServ architecture because the former has per-flow-base traffic handling mechanisms but the latter does not. This general belief, however, has not been fully verified yet. In this paper, we theoretically and numerically compare IntServ and DiffServ architectures in terms of x-percentile end-to-end queueing delay as well as the maximum end-to-end queueing delay. We find that, in some network topology, DiffServ architecture makes the maximum queueing delay smaller than IntServ architecture does. We also find that in most cases x-percentile queueing delay in DiffServ-based networks is much smaller than that in IntServ-based networks

We study the problem of estimating a source-destination (SD) traffic matrix of an IP-based VPN from measurements of aggregated traffic along all access links. We derive two different estimators by formulating the SD-traffic-matrix estimation as a deterministic or stochastic optimization problem. One estimator does not rely on any stochastic information about the SD traffic and the other estimator assumes the linear relationship between the average and variance of the SD traffic. The proposed estimators have expressions much simpler than those of standard SD-traffic-matrix estimators; one estimator has an explicit representation that does not involve matrix inversion. The other is given by a simple matrix formula and does not need any prior SD-traffic matrices. Through simulation experiments, we show that these proposals are superior or comparable to the standard SD-traffic-matrix estimator in term of the estimation accuracy.

We propose an admission control for the IP telephony service over DiffServ-based IP networks. The proposal has three important features. First, the proposed control is fully distributed; each edge router autonomously admits or rejects connection-set-up requests based on the stored information of resource availability. Secondly, the resource availability is expressed in terms of the maximum number of multiplexable IP telephony sources in advance. This feature allows making connection acceptance judgment in real time without measuring IP-telephony traffic. Finally, our proposal can specify QoS performance targets as an arbitrary x-percentile queueing delay. This feature allows the full receiving of the benefit of statistical multiplexing gain.

We propose a new technique for measuring the bottleneck bandwidth of a network path between end hosts. Our proposal is based on the packet-pair technique and uses a simple statistical procedure for filtering out the cross traffic effect. We explain the basic mechanism of our proposal and its implementation in two different types of tools, single end-point and double end-point tools. The former runs in one end-point of the path, assuming that the other end-point or intermediate routers should reply to each received packet using ICMP (Internet control message protocol) packets. The latter works under the cooperation of both source and destination hosts. We have extensively conducted several experiments on nationwide and international networks to show that our proposal outperforms existing tools, such as Pathchar and Nettimer, in terms of accuracy and robustness of measurement.

We propose an estimator of the effective bandwidth, called Gaussian estimator using buffer measurement, for superposition of sources in IP networks. Our estimator does not rely on a specific traffic model and works only based on a small set of measurements of the workload in the buffer of a router. We conducted simulation experiments by using real traces, which were collected on a LAN at Chiba University and collected at an US-Japan link of WIDE backbone, and show that our proposed estimator performs better than other existing estimators that rely only on the measurement of traffic amount from sources.

The asymptotics of cell-loss ratio (CLR) in the regime of large buffers are characterized by two parameters, the asymptotic constant and asymptotic decay rate. Both parameters can be expanded in powers of one minus the link utilization (heavy traffic expansion). Thus, once the coefficients in this expansion are obtained, the CLR can be very easily estimated from the measured link utilization by using CLR asymptotics. This paper proposes an algorithm for estimating these coefficients in real time. For this purpose, the notion of state-space representation for a single-server queue is introduced: the elements of the state vector are the coefficients in the expansion of the asymptotic constant and asymptotic decay rate. Bayesian regression analysis is applied to estimate the state vector based on the buffer measurement. Our approach does not require any models describing the statistics of the traffic other than the asymptotic behavior of the CLR in the regime of large buffers. In addition, it allows us to estimate the effective bandwidth of the aggregated process easily, so it is applicable to a wide range of ATM traffic control methods, such as connection admission control and VP bandwidth control. We describe how this method of CLR estimation will be applied to connection admission control and VP bandwidth control by using results from simulation experiments.

The paper presents a centralized virtual path bandwidth (VPB) control scheme for ATM networks which satisfies mainly three specifications: a) optimality during a medium-term control interval, b) fast time response so as to absorb the medium-term traffic fluctuations, and c) easy implementation. The paper mainly points at the impact of direct, on-line traffic measurements on bandwidth control. The control objective is to rearrange the installed bandwidth of the virtual paths according to the offered traffic so as to minimize the maximum call blocking probability of the whole network. Network simulation shows that a sophisticated VPB controller which relies on simple measurements of the offered traffic can substantially improve the performance of an ATM network. The necessary bandwidth rearrangement time is also examined by simulation.>