In digital twin-enabled smart cities, the delivery of various urban services based on urban digital twins is expected. These services comprise various microservice-like artificial-intelligence-empowered atomic tasks that use dynamic spatial information through the digital twin. To maintain urban service quality amid fluctuating demand, effective resource utilization is essential. In this paper, we propose a system model for provisioning urban services composed of atomic tasks, along with an adaptive task and resource allocation method based on the monitor, analyze, plan, execute, and knowledge (MAPE-K) loop framework. Performance evaluation demonstrates that combining task quality and resource allocation adjustments enables effective adaptation to demand fluctuations.

There will be an increase in situations where images taken in specific locations are transmitted through networks for various services. However, this trend can lead to significant communication loads due to simultaneous transmission of images from multiple locations. Therefore, it is important to reduce the amount of network traffic in image transmission. While traditional compression methods focus on minimizing information loss in images, some applications only require the retention of semantic information, suggesting potential improvements in communication efficiency. This paper proposes an image generation-based transmission method for highly-efficient communications exploiting composition and descriptive information. The proposed method extracts specific information from an image to decrease the amount of data transmission, and reconstructs the image using an image-generative model by a receiver. In addition, image compression and reconstruction in the proposed method are demonstrated through an experiment. The experimental results indicate a need for a method to evaluate the output and a method for image reconstruction based on this evaluation.

We propose a user mobility-driven federated learning method, which integrates learning models from different regions, leveraging user mobility. This method aims to improve performance of learning models in specific regions by merging them with models from other areas. In regions with less user mobility, our method creates unique regional models, while in areas with high mobility, it integrates models for enhanced performance. Evaluation results indicate that accuracy improved with additional training, although it temporarily decreased after model integration.

Wireless point cloud streaming is a key technology in digital twin-enabled smart cities for maintaining city-scale spatial information. Various point cloud coding and compression algorithms have been studied to facilitate efficient transmission. However, real-field experiments are required to reveal the characteristics using small devices, which support the city from the edges. In this paper, we conducted field experiments to analyze the characteristics of wireless point cloud streaming. First, we proposed a middleware to mediate between the robot operating system and the network interface to easily configure the transmission protocol. Thereafter, we implemented the middleware on edge devices and analyzed the performance using IEEE 802.11ax based on the following perspectives: processing time versus transmission frequency, point cloud and packet transmission success rates versus distance between nodes, and point cloud transmission success rate versus mobility of nodes. The first experiment revealed that small edge devices with the proposed middleware could handle high-frequency streaming without exceeding the maximum time of every period. The second experiment revealed the relationship between data rate and transmission success rates of point clouds and packets. Finally, the third experiment revealed that the mobility of the nodes caused variations in the transmission success rate.

Concurrent transmission (CT) is a promising technology for achieving efficient data transmission based on constructive interference of wireless broadcasts. However, it fails to transmit packets because of the destructive interference. Opportunistic routing (OR) avoids transmission failure by using path diversity based on flexible forwarder selection. However, it fundamentally imposes transmission delay owing to the packet-level forwarder selection. In this paper, we propose backoff-based OR with CT, which efficiently forwards data packets using CT, and then supports them based on OR. Finally, we compared the performance of the proposed method with that of traditional approaches using a simple lattice topology. The simulation results revealed that the proposed method achieved both a high packet transmission success rate and a small transmission delay while suppressing an increase in the network load.

In anticipation of environments where Internet of things (IoT) is extensively utilized, various methods for understanding and managing the behavior of IoT devices based on traffic analysis have been proposed to tackle security issues such as the difficulties in replacing devices, updating their firmware, and comprehending their behavior. We have previously studied an IoT traffic analysis and device identification in smart home environments with applying two-stage clustering. This method realizes extracting time-series characteristics of IoT traffic based on unsupervised machine learning, k-means clustering, and identifying 21 IoT devices with a high accuracy. Nevertheless, we have not yet examined whether k-means clustering is appropriate to the two-stage clustering for IoT traffic analysis and device identification. In this paper, we investigate various clustering methods for two-stage clustering and then compare and discuss the characteristics of the methods for IoT traffic analysis and device identification. The results confirm that the k-means clustering or Ward’s method is more suitable for the performance of feature extraction and the computational loads.

Ahstract-Various methods focusing on either spatial or temporal utilization, such as path diversity or link utilization, have been studied to realize efficient ad hoc networks. Furthermore, a method that integrates these methods has been proposed; however, it has only been evaluated theoretically and partially. Hence, achieving efficiency from both viewpoints as a protocol is challenging. This study proposes and defines block transmission-enabled opportunistic routing, which can achieve spatiotemporal efficiency as a protocol. The proposed method was evaluated by comparing its performance with that of two conventional methods that focus on either spatial or temporal efficiency, through computer simulations.

Machine learning (ML) is predominantly performed in the cloud or other computing facilities. While this computing method allows for the benefits of ML to be leveraged in urban settings, it may also lead to unfair sharing of the environment owing to the heat generated by servers in areas housing computing infrastructure. In this study, we developed a green distributed machine leaning (DML) concept-CoopL-to calculate the local consumption of computational resources based on DML, thereby mitigating the environmental burden. Moreover, we analyzed the impact of long-distance communication by incrementally raising the communication latency. Furthermore, the power consumption was examined by considering the hop count of the router.

This paper presents a flexible group voice conversation system that utilizes the location and direction of speakers to help them choose their listeners. The proposed system comprises a group construction mechanism for peer-to-peer voice conversation and a listener selection mechanism in the conversation group. To reveal the effectiveness in real environments, we implemented the proposed system as a smartphone application. Finally, we conducted a real-world experiment to perform subjective evaluations using the proposed application to capture and understand the user experience and application usability.

To collect and aggregate the enormous amounts of spatial information needed to maintain virtual three-dimensional (3D) spaces, such as digital twins, in real time, significant resources are required. Recently, light detection and ranging (LiDAR)-equipped mobile devices have become increasingly popular. Therefore, mobile sensing using such devices is a feasible solution for maintaining and updating virtual 3D city spaces by motivating people. In this study, we propose and develop a mobile sensing system based on an augmented reality coloring game using LiDAR-equipped mobile devices. Next, we verify the fundamental operation and effectiveness of the proposed system.

Backoff-based opportunistic routing can adaptively select a forwarder among receivers based on a random backoff time. However, it cannot correctly work in an environment with asymmetric links where the communication quality differs in the directions. To solve this problem, a confirmation method for link symmetry and destination reachability was previously proposed. However, the method can only confirm link symmetry with one neighbor and destination reachability for a single terminal in a single packet transmission. This paper proposes a fast confirmation method to verify the link symmetry of multiple terminals in single packet transmissions. The performance of the proposed method is clarified with simulations.

The authors have been proposed a wireless multihop routing method inspired by the behavior of the slime mold. The previous methods iteratively calculate bandwidth allocation for each link based on the maximum bandwidth, the transmission delay, and the required communication volume by the source node. However, the appropriate number of calculation iterations and the actual calculation time for bandwidth allocation vary depending on various factors. In addition, the allocation results cannot be reflected until the calculation is completed, thus the allocation results may not be suitable for the network conditions at the end of the calculation. In this paper, the authors propose an adaptive wireless multi-hop routing with reduced impact of processing delay by sequentially using values that are convergence processes as indicators for route selection.

In order to prevent potential accidents between pedestrians and vehicles, it will be effective to introduce a vehicle-to-pedestrian warning system. The system detects potential accidents between pedestrians and vehicles in advance and sends warning to pedestrians and vehicles ahead of time so that potential accidents can be avoided. The time ahead of the potential accidents is named prior notification time and it is set in advance in the system. However, there is a difference between the prior notification time and the actual notification time due to system delay and inaccuracy of GPS localization. This difference is named warning error. The location estimation error and the speed of pedestrians and vehicles will affect the appropriate estimation of prior notification time. In this paper, how warning error is changed with vehicle speed is investigated through simulation. The simulation results show that the warning error increases as the vehicle speed increases. The maximum warning error is 2 seconds regardless of the vehicle speed. Thus, notification time delay can be reduced by estimating the maximum value of the warning error and increasing the prior notification time in advance.

In mobile ad-hoc networks (MANET), multi-hop communication relies on a give-and-take relationship among users. For stimulating users to be more cooperative in MANET, an incentive mechanism that offers electric power in exchange for relaying packets by using wireless power transmission (WPT) technology has been proposed. To evaluate the availability of MANET and this incentive mechanism, it is important to reveal factors affecting to the user's willingness to relay packets. In this paper, we investigate and analyze user's willingness to relay through qualitative evaluations. On the basis of the analysis, we model user's willingness to relay packets depending on the residual battery level and user's characteristics. Finally, we also confirm that the electric incentive mechanism can stimulate reciprocal cooperation among users in MANET.

This paper proposes an adaptive peer search method that considers the dynamic users’ mobility in location-based peer-to-peer (P2P) networks. In the previous study, the neighbor peer search is periodically executed, and the search area is constantly fixed as a circle regardless of the peers’ moving speed. Therefore, it is difficult to properly acquire on neighbor peers, especially in the situation when peers are moving at high speed. In this paper, we suggest a method to determine the peer search interval and the search area size adaptively depending on the peers’ locations and moving speed. Simulation results show that the proposed method enables peers to search neighbor peers at appropriate intervals and to obtain nearby information efficiently.

The Internet of Things is now part of everyday life and there has been a wide range of novel IoT applications collecting cyber-physical data and providing information on the environment. As it is expected that the IoT traffic will count for a major part of the Internet traffic, it is essential to characterize the IoT traffic and to identify each device, and especially in the case of cyberattacks. In this paper, we present a new method to identify IoT devices based on traffic entropy. We compute the entropy values of traffic features and we rely on Machine Learning algorithms to classify the traffic. Our method succeeds in identifying devices under various network conditions with performances up to 94% in all cases. Our method is also robust to unpredictable network behavior with anomalies spreading into the network.

The Internet of Things is expected to count for a large part of the overall Internet traffic but its main characteristics are still widely unknown. In this work, we focus on IoT Traffic and we present a novel IoT Traffic Generator. As an use-case, we generate IoT traffic within a Smart-Home Environment. Our generator allows studying the IoT traffic and can serve for describing the main properties of the IoT traffic. It is a first step toward the detection of security threats and anomalies for IoT Traffic.

The Internet of Things (IoT) is expected to count for a large part of the Internet traffic and its impact on the network is still widely unknown. It is therefore essential to study the IoT Traffic in order to characterize its properties and evaluate its performances. In this paper, we propose a novel IoT traffic generator called IoTTGen. We model the IoT traffic and we generate synthetic traffic for smart home and bio-medical IoT environments. We also extracted anomalous IoT traffic from a real dataset and study the IoT traffic properties by computing the entropy value of traffic parameters. Our generator succeeds in capturing the characteristics of the IoT traffic, which can be visually observed on Behavior Shape graphs. Our generator can also serve to describe the main IoT traffic properties and also to detect IoT traffic anomalies.

This paper proposes a classification method for peer-to-peer video streaming (P2PTV) traffic using machine learning. Since the user terminals (peers) share video data in P2PTV, P2PTV traffic is difficult to control and manage statically as both the number of peers sharing the same video data and the throughput vary with respect to contents. Although there exists a conventional method to classify and model P2PTV traffic by focusing on the number of peers and throughput, problems on the classification criteria and reproducibility remain in this method. In this paper, we use a clustering method that is considered as one of the machine learning methods and try to classify P2PTV traffic data into some categories. We extracted 18 features by analyzing P2PTV traffic of popular P2PTV applications: PPStream and PPTV; and then classified the traffic. The classification results show that about 400 traffic data sets were categorized into four clusters.

In mobile ad hoc networks, radio interference and mobility of nodes may degrade the packet arrival rate due to the dynamic topological change. Then, traditional reactive routing protocols may cause huge network resource consumption due to the route request flooding for discovering the destination. To solve the above issue, flooding area restriction methods can reduce the unnecessary control messages to narrow the flooding area based on the location information. However, each node must share location information before sending route requests via control messages. Besides, it is also difficult to share the correct location information of nodes due to the mobility of nodes. This paper proposes a flooding area restriction method to reduce the unnecessary control messages to determine the flooding area by sharing the location information via a mobile network. In addition, computer simulations reveal the effectiveness of the proposed method in comparison with a traditional routing protocol.

This paper proposes an ant-inspred backoff-based opportunistic routing protocol which optimizes the forwarding area based on the concept of ant colony optimization. In the traditional backoff-based opportunistic routing protocols, they may increase unnecessary packet forwarding to expand the forwarding area since a receiver of data packet autonomously participates the packet forwarding. Hence, to deal with the forwarding area expansion, the proposed method suppresses the potential forwarders to adjust the backoff time based on their pheromone levels. Finally, this paper clarifies the performance of the proposed method in comparison with the conventional method using computer simulations. The results show that the proposed method can reduce the network resource consumptions and the end-to-end delay while keeping higher packet transmission success rate.

After being the support of the data and voice convergence, the Internet has become one of the main video providers such as TV-stream. As an alternative to limited or expensive technologies, P2PTV has turned out to be a promising support for such applications. This infrastructure strongly relies on the overlay composed by the peers that consume and diffuse video contents at the same time. Understanding the dynamical properties of this overlay, and in particular how the users switch from one overlay to another, appears to be a key aspect if one wants to improve the quality of P2PTV. In this paper, we investigate the question of relying on non-invasive measurement techniques to track the presence of users on several channels of P2PTV. Using two datasets obtained by using network measurement on P2PTV infrastructure, we show that such an approach contains sufficient information to track the presence of users on several channels. Besides, exploiting the view provided by sliding time windows, we are able to refine the analysis and track users that switch from one channel to another, leading to the detection of super-peers and providing explanations of the different roles they can play in the infrastructure. In addition, by comparing the results obtained on the two datasets, we show how such analyses can shed some light on the evolution of the infrastructure policy.

Owing to proliferation of smart phones, communication services such as a video streaming service are common in a mobile situation. For these services, quality evaluation and communication control based on Quality of Experience (QoE), which is the degree of a user's subjective satisfaction, is very important because the final goal of delivering high-quality service is improving user's satisfaction. QoE tends to be affected by several factors including Quality of Service (QoS). Therefore, collecting QoS data from the users who use the mobile application has become one of promising schemes to meet the QoE, which is called the crowd sourcing data. The crowd sourcing data are, however, apt to be affected by sensing errors and low accuracy. In this study, we propose to estimate some densities just from the records of application use and its location information to remove the sensing errors and low accuracy. The Kernel density estimator is used to derive a continuous density function from discrete distributed sample data such as the collected QoS data. From the viewpoint of QoS, it is important to extract high-density fields of the application use to find out QoS degradation. After estimating the Kernel density estimator, we determine the borderline between the high-density field and the other field by using the reference value that can be determined from the observed data. Simulated experiments verify effectiveness of the proposed method.

In ad hoc networks, backoff-based OR (opportunistic routing) protocols, which autonomously selects a forwarder among the potential forwarders based on a random backoff time, have been proposed. However, each potential forwarder must wait for the backoff time to avoid packet collisions among receivers. In addition, the terminal density strongly affects the performance since the backoff-based OR protocols improve the performance by using multiple terminals. In this paper, we propose a novel OR protocol called PRIOR (prioritized forwarding for opportunistic routing). In PRIOR, a terminal called prioritized forwarder, which forwards packets without using a backoff time, is selected among neighbours. In addition, we propose a hop-by-hop retransmission control that performs the retransmission terminal selection on the basis of a neighbour relation with the PFs. Moreover, we introduce an explicit acknowledgement mechanism on the basis of the difference of hop counts to alleviate the bad effect of the retransmission control in dense environments. Finally, we evaluate PRIOR in comparison with conventional backoff-based OR protocols in computer simulation.

Network providers are nowadays facing with a big question in allocating network resources due to the constantly increasing of Internet services. While the network resource is not unlimited, users always desire the best quality of experience (QoE) with the huge information exchange. Therefore, finding a justice of network resource allocation based on the user experience is mandatory. In previous studies, network resources were allocated to all users by using a specific utility function without considering the user characteristics. In fact, the network resource consumptions are very different among individual users and directly depend on users' behavior. For instance, the network demands of relax users are usually lower than those of busy users. Thus, allocating the same amount of resources to all users might not meet their expectations. In this paper, we propose two bandwidth allocation methods by classifying users into different groups based on their characteristics such as relax, short-time busy, and long-time busy users. To allocate the bandwidth in each group, specific utility functions are applied. By using these methods, the obtained results show that users get the different allocated bandwidth while they still experience the same level of QoE. On the other hand, the allocation considers users' behavior to allocate suitable bandwidth based on the real resource consumption of users. As a result from these analyses, our proposal tends toward a fair allocation as well as an efficient management of the network resources.

In ad hoc networks, radio interference and terminal' mobility can cause frame losses. When the frame losses occur continuously, general datalink layer protocols regard it as a link failure and then notify it to network layer protocols. In conventional routing protocols, a source terminal initiates route discovery process to recover from the link failures on receiving the notifications even if the source terminal still has a valid route to a destination terminal. However, the route discovery requires a route request flooding and the flooding may consume limited network resources of ad hoc networks and that also degrades transmission efficiency. For solving the problems, several retransmission and relay control methods have been proposed. In these methods, neighbour terminals of a transmission route overhear transmitted frames and retransmit lost frames autonomously by using the overheard frames. However, the conventional retransmission methods have a disadvantage that they degrade the transmission efficiency under bursty frame loss interval. Moreover, conventional relay control methods suffer from a processing overhead and a slow adaptation problem under a sparse frame loss. In this paper, we propose an efficient and lightweight method that neighbour terminals adaptively retransmit lost frames considering a frequency of the losses. Computer simulations clarify the performances and the effectiveness of the proposed method.

Recently, most peer-to-peer (P2P) applications ignore traffic costs at internet service providers (ISPs), and thus generate a large amount of traffic crossing ISPs or autonomous systems (ASes) on the Internet. To reduce the cross-ISP/AS traffic, the existing approaches introduce network-aware strategies in which a lot of modifications of P2P systems are required. In this paper, we propose a router-aided method for localizing P2P traffic hierarchically with multiple levels. By intentionally limiting the bandwidth of each connection path between peers based on geographical location of the peers' destinations, the traffic can be localized. Compared to the existing locality-enhancing approaches, our proposed method does not require dedicated servers, cooperation between ISPs and users, or any modification of existing P2P application software. Therefore, the proposal can be easily utilized by all types of P2P applications. The experimental results obtained with P2P streaming applications show that our proposed method not only successfully realizes traffic localization but also maintains a good performance of P2P applications.

Notion of the user-centered design (UCD) is requisite to improve user satisfaction. Matching the concept of UCD, Quality of Experience (QoE) is expected to become a paradigm shift from the existing operator-initiative network services. There are, however, several difficulties to realize practical communication services based on QoE. One difficulty is how to bridge QoE and Quality of Service (QoS) parameters relevant to system performance. Moreover, since QoE is based on the end-user subjective evaluation, flexible modeling corresponding to the end-user psychological status is needed. In this paper, an interactive Web service is considered and a QoE model with end-user psychological effect is derived by use of the Weber-Fechner law. Data of QoE and QoS parameter are collected through experiments and the mapping functions between QoE and QoS are estimated from the data.

Recently, peer-to-peer (P2P) traffic is increasing rapidly in volume day by day. One of the main causes is that most of P2P applications including file sharing and streaming applications often form overlay networks for exchanging data that are oblivious to the underlay network topology. As a result, they generate a large amount of inter-domain traffic causing higher cost for Internet service providers (ISPs). This raises a problem of traffic localization. To optimize the cross-ISP/AS traffic, existing approaches focus on solving the problem on the application layer where each P2P application must be equipped with an additional protocol to obtain underlay network information from an “Oracle” server or an additional locality-aware procedure to estimate location by itself. Therefore some modifications of application software are required for these approaches. In this paper, we propose a novel solution for addressing the problem, called PLS, forcing packet loss to each P2P packet based on geographical location of each destination at the network layer. Since PLS is implemented at a network router, no software modification is required. This proposal can be applied to all types of P2P applications in order to localize the traffic. The experiment in this study evaluated on popular P2P streaming applications show that our proposed method significantly improves the performance of traffic localization problem.

Recently, peer-to-peer has been becoming a promising architecture for developing video streaming services because it can reduce the bottlenecks and the load at the server side. However, P2P systems implement their own routing protocols on overlay networks, which are largely independent of the Internet routing. As a result, the traffic generated by P2P systems is immense and unpredictable. Controlling the P2P traffic is therefore becoming a big challenge for internet service providers (ISPs). Considering the P2P traffic localization is one of the most efficient solutions to optimize the traffic. The existing approaches, however, require several modifications of the application software. In this paper, we propose a novel traffic localization method for P2P streaming application (P2PTV) by deep packet inspection. Based on the peer list and the geographical location of the listed peers, we propose two mechanisms for localizing the traffic: the modification of the peer list packets and the redirection of the video request packets at network routers. Since our method is implemented on the network routers, it can be applied for all P2P applications without any software modifications. The experiments evaluated on a popular P2PTV, namely PPStream, prove the effectiveness of our proposed method on the problem of traffic localization.

Both of Quality of Service (QoS) and Quality of Experience (QoE) are defined to specify the degree of service quality. Although they are dealt with in different layers in multi-layered models, collaboration of these is necessary to improve the user satisfaction for telecommunication services. In this paper, after sorting out the concepts and specification of QoS and QoE, a service quality coordination model combining these is proposed. The model is applied to a video-sharing service and its coordination model is derived based on subjective experiments. The structural equation modeling is used to compute the user satisfaction from QoS and QoE.

This paper proposes a video compression algorithm that uses the edges of frames to estimate and compensate for motions. Based on the algorithm, we propose two schemes that balance energy consumption among nodes in a cluster on a wireless video sensor network. In these schemes, the compression process is divided into several small processing components, which are then distributed to multiple nodes while considering their residual energy. We conducted extensive computational simulations to verify our methods and found that the proposed schemes not only solve the energy balance problem by coordination of the processing tasks but also increase the quality of decoded video.

In ad hoc networks, wireless links can be unidirectional due to the difference of transmission range of each node. Particularly for temporally formed ad hoc networks, routing protocols should support route construction using unidirectional links to provide high connectivity in various network configurations. Although existing DSR have a mode of two way flooding to support route construction using unidirectional links, it involves substantial control traffic and should be avoided in the high node density environment. This paper proposes a novel routing protocol supporting unidirectional links called FOCUS2 (flooding based ad hoc routing protocol controlled with overheard control information to support unidirectional links). FOCUS2 employs an area-controlled two way flooding that adaptively limiting flooding area. Performance evaluation by simulation shows that FOCUS2 reduces the number of control packets by approximately 50 % as compared with existing two way flooding DSR and that the degradation of route construction performance is small

In ad hoc networks, unidirectional links between nodes exist due to differing propagation patterns or wireless interferences. Routing protocols for bidirectional links must cause the decline of connectivity. In recent years, some routing protocols considering unidirectional links have been proposed. However, these protocols have a problem: the broadcast traffic for backward path discovery increases. This paper proposes a novel routing protocol for ad hoc networks that positively uses unidirectional links and that drastically reduces the number of control packets. Our protocol improves network connectivity and reduces the number of control traffic.

In ad hoc networks where mobile nodes form networks autonomously, many unidirectional links can be generated due to the diversity of transmission range of each node In order to achieve high connectivity in such networks, routing protocols should construct routes by aggressively using unidirectional links. Typical existing routing protocols such as DSR (dynamic source routing) are basically designed assuming links are bidirectional, and they generate substantial control traffic in the unidirectional routing since their optimization schemes are not effective for unidirectional links. This paper presents a novel on-demand ad hoc routing method called FOCUS (controlled-flooding-based ad hoc routing protocol using overheard hop count information to support unidirectional links). FOCUS reduces control traffic of route reconstruction as compared with conventional method by issuing route request in the restricted region around the broken route. The results of performance evaluation by simulation show FOCUS reduces the number of RREQ (route request) packets by about 60 % and reduces packet transmission delay by about 30% compared to DSR.

This paper investigates topological designs for multicast networks using simulated annealing (SA) and proposes a new method for finding an effective initial solution to the problem of reducing the computational time of SA. Multicast communications can decrease network traffic due to branch connections. Accordingly, multicast services will be predominant in future traffic. Since the characteristics of multicast communications are quite different from those of unicast communications, it is necessary to find a suitable topology for future networks that accommodates both multicast and unicast services. Finding the optimal topology is generally known as an NP-hard problem, and so heuristic algorithms have been used in many cases. This paper optimizes the multicast network topology by simulated annealing, a well-known powerful heuristic. Even if SA were employed, however, more additional inventions would be required to compute and find the solution with the smallest possible number of iterations. Therefore, this paper proposes a method for finding an effective initial solution for SA. The results show that the computational time needed to reach the final solution becomes shorter if our proposed method is applied.