114 results about "End-to-end delay" patented technology

The invention discloses a multi-device cooperative control method applied to deployment of an embedded real-time operating system, and relates to the field of embedded real-time operating systems, which comprises the following steps of: S1, constructing a real-time dependency relationship graph of a task link, a multi-device cooperative task is disassembled into sub-task link units including an execution path, a real-time constraint threshold and a cross-device communication dependency relationship. According to the multi-device cooperative control method applied to the deployment of the embedded real-time operating system, the problems of task priority conflict and resource utilization imbalance in a multi-device heterogeneous environment are effectively solved through a dynamic priority transmission mechanism and a distributed resource elastic perception model. The system can sense equipment load, communication link stability and task real-time requirements in real time, dynamically adjust sub-task weights and execution paths, ensure conflict-free transmission of end-to-end delay constraints, optimize the global resource utilization rate and avoid the problem of equipment overload or idle caused by traditional static scheduling.

The invention discloses a GPU (Graphics Processing Unit) and parallel IO (Input / Output) collaborative optimization method based on mode heterogeneous computing, and particularly relates to the field of heterogeneous computing optimizing.The method comprises the following steps: monitoring mode characteristics during execution of a target computing task in real time through an instruction analyzer, and synchronously collecting dynamic parameters of a heterogeneous computing environment to generate a multi-dimensional resource situation matrix; based on the mode and the resource situation matrix, collaborative optimization operation of the binding equipment dimension and the scheduling time dimension is executed; and constructing a three-level adaptive transmission mechanism and periodically implementing feedback optimization according to an execution state of the three-level adaptive transmission mechanism. According to the GPU and parallel IO collaborative optimization method based on mode heterogeneous computing, the data prefetching hit rate is increased by dynamically sensing the matching relation between a heterogeneous computing mode and an I / O access mode; end-to-end delay is reduced and task contention conflicts are reduced through a dynamic matching strategy of an equipment capability portrait and a task type; and the heterogeneous storage bandwidth peak value utilization rate and the I / O throughput are improved by differentially utilizing the hierarchical characteristics of the multi-stage storage.

The invention discloses a cloud side-end collaborative virtual power plant scheduling method and system based on dynamic task migration, relates to the technical field of virtual power plant resource scheduling, and solves the problem that a general method is difficult to meet the real-time scheduling requirement of a virtual power plant. According to the invention, task migration is realized through a three-level architecture of a cloud planning long-term strategy, regional collaboration of an edge layer and quick response of a terminal layer in combination with a dynamic load balancing algorithm; the first edge node locally processes regional collaborative tasks, reduces the communication frequency of the cloud server, reduces the end-to-end delay, and then combines federated learning to realize training parameter sharing between the edge nodes and avoid redundant data transmission, thereby effectively improving the response speed of the scheduling system, and improving the scheduling efficiency. And the real-time scheduling requirement of the virtual power plant can be better met.

The invention relates to the technical field of network flow analysis, in particular to a non-intrusive network flow real-time analysis method and system based on an eBPF, and the method comprises the steps: deploying an eBPF program in a kernel mode, and capturing a network data packet entering a kernel protocol stack in real time; identifying an application layer protocol type and analyzing a key field; the structured data is transmitted to the user mode through the annular buffer area; and the user state carries out TCP stream recombination and session-level aggregation statistics on the structured data. According to the invention, on the premise that application codes are not modified, real-time and structured protocol analysis can be carried out on various common application layer protocols in a kernel layer; end-to-end delay is accurately disassembled through a full-link delay disassembling mechanism, and a performance bottleneck link is positioned; the analysis result is output in the form of structured data, second-level performance statistics and abnormal behavior detection are supported, and the operation and maintenance analysis efficiency and the system observability are remarkably improved.

The invention discloses a data stream transmission scheduling method based on a distributed SDN satellite network, and the method comprises the following steps: building a distributed SDN satellite network model according to the application background of a large-scale satellite network, and constructing a data stream load balancing routing problem of the large-scale satellite network under the distributed SDN satellite network model; in order to solve the problem of data flow load balancing routing of a large-scale satellite network, a data transmission strategy of a distributed SDN satellite network is adopted, the load balancing effect is achieved, and end-to-end delay is reduced. According to the invention, a distributed SDN satellite network model is established according to the application background of a large-scale satellite network, a data flow load balancing routing problem of the large-scale satellite network is established under the model, and a data transmission strategy based on the distributed SDN satellite network is provided for the optimization problem. According to the strategy, a better load balancing effect and smaller end-to-end delay can be realized.

The invention relates to the technical field of power systems, in particular to a cloud edge computing power regulation and control method and system based on edge computing, and the method comprises the following steps: S1, collecting the resource state data of a cloud data center, edge nodes and terminal equipment in real time; s2, constructing a dynamic scheduling strategy based on a deep reinforcement learning model, and generating a computing power allocation scheme according to task types, priorities, real-time resource states and historical load data; s3, performing hierarchical processing on the task; s4, autonomously executing emergency task scheduling in a network disconnection or high-delay scene through a local autonomous module of the edge node; s5, global energy consumption is optimized based on a genetic algorithm, the steps S1 to S5 cooperatively achieve dynamic regulation and control of cloud edge computing power, and accurate input is provided for dynamic scheduling by collecting multi-dimensional resource data in real time; through a task layering mechanism, it is ensured that high-real-time tasks are preferentially processed locally, and end-to-end delay is reduced.

Techniques are described for creating a network-link between a first virtual network in a first cloud environment and a second virtual network in a second cloud environment. The first virtual network in the first cloud environment is created to enable a user associated with a customer tenancy in the second cloud environment to access one or more services provided in the first cloud environment. The network-link is created based on one or more link-enabling virtual networks being deployed in the first cloud environment and the second cloud environment.

Disclosed in the present invention is a URLLC multi-connectivity transmission and resource allocation method for industrial automation. The method comprises the following steps: S1, establishing a multi-connectivity technology-based URLLC system model; S2, setting constraint conditions of end-to-end transmission on the basis of quality of service requirements for an industrial automation application; S3, reporting the length of an information byte required for transmission, and determining a bandwidth optimization problem of multi-connectivity data transmission; and S4, executing a resource allocation algorithm to calculate an optimal cross-layer resource allocation solution for the multi-connectivity data transmission. According to the present invention, a URLLC multi-connectivity transmission and resource allocation method for industrial automation is established, to avoid the problem of cellular link failure caused by dynamic changes of channels in a traditional transmission mode; and under the constraints of end-to-end delay, reliability and maximum code length, an optimal cross-layer resource allocation method is provided, to minimize a bandwidth required for multi-connectivity transmission, thereby achieving the objective of saving spectrum resources.

A routing and scheduling method based on multi-cycle CSQF mechanism and GDRL is provided. The routing and scheduling method includes the following steps: S1, initializing a cycle index detection mechanism, a queue mapping and a queue mapping constraint of a multi-cycle CSQF; S2, constructing a DFRLLS model; S3, optimizing the DFRLLS model based on GDRL; S4, off-line training a learning strategy of a GDRL model; S5, making a decision on-line based on a trained GDRL model. The routing and scheduling method based on multi-cycle CSQF mechanism and GDRL is adopted, and GCN network is used to extract topology information between networks. Compared with the method of only using reinforcement learning, the routing and scheduling method can achieve more flow scheduling, and its performance is also stable under complex network topology, multi-cycle CSQF can reduce the start-to-end delay of the flow compared to the CSQF.

The invention discloses a video stream decoding delay prediction and dynamic quality optimization method based on multi-feature fusion. The method comprises the following steps: step 1, video stream feature extraction; step 2, video stream decoding delay prediction: designing a hybrid model architecture to predict decoding delay, namely adopting a LightGBM sub-model and an LSTM residual compensation sub-model; 3, dynamically deciding the quality of the video stream: switching to Q-learning under the normal condition when the network fluctuates violently, and finally determining the sequence output of the CRF value in a rolling window; step 4, video stream smoothing processing: video playing adopts double-buffer progressive rendering and code rate smoothing filtering; and step 5, after the smooth processing of the video stream is completed, transmitting the processed video data to a display device for rendering and presentation. According to the method, quality selection is optimized and fluctuation is inhibited under end-to-end delay constraint.

The invention relates to the field of audio and video communication, and particularly discloses an audio and video communication demonstration method and demonstration device under an electronic countermeasure condition, and the method comprises the following steps: 1, carrying out the coding compression of an original audio and video signal through a perception compression technology, the perception compression technology is combined with a deep learning algorithm, ROI region detection and protection and coding parameter prediction based on content perception; through combination of deep learning, ROI region protection and content awareness coding, on the premise of keeping the visual image quality of human eyes, the 1080P video code rate is compressed from 2Mbps of traditional H.265 to below 400Kbps, the code rate is reduced to 1 / 20 of the original code rate, the channel bandwidth is remarkably saved, and the communication cost is reduced; by adopting adaptive multi-protocol hybrid transmission and forward error correction coding, the audio and video availability can still be maintained when the bidirectional packet loss rate is as high as 70%, the end-to-end delay is reduced to be within 500ms, and the inherent defects of a traditional single protocol are overcome.

The invention relates to a digital twin-driven Internet of Things user behavior prediction and resource allocation method, and belongs to the field of mobile communication. Aiming at the problems of service quality reduction, load imbalance, low utilization rate and the like caused by resource allocation lag in the traditional method, the invention provides the following scheme: constructing a three-level system model of a physical infrastructure layer, a digital twin network layer and a control layer; an SA-TCN model combining a self-attention mechanism and a time convolution network is designed, and the prediction precision in a dynamic environment is improved through meta learning; establishing an end-to-end delay, DT credibility and system cost model, and taking maximization of system utility as a target; and dynamically optimizing a resource allocation strategy by adopting a PER-MATD3 algorithm. According to the method, high-precision user behavior prediction is realized, the real-time performance of resource scheduling is improved, the resource utilization rate, the load balance and the comprehensive effectiveness of the system are remarkably improved under the condition that end-to-end delay and credibility constraints are guaranteed, and the method is suitable for a high-mobility Internet of Things scene.

The invention relates to a scale out network backup path fast switching notification system and method, and aims to solve the problem that the existing VRRP / BFD protocol switching time cannot meet the high real-time requirement of an AI server scale out network. The scheme is realized through a collaborative architecture of a CPU (Central Processing Unit), an inspection FPGA (Field Programmable Gate Array) and power supply monitoring: the CPU runs a VRRP / BFD (Bidirectional Forwarding Detection) protocol stack and maintains a state table, and the FPGA is responsible for detecting a power supply fault, a port fault and GPU alarm in real time, and generating and sending a pre-installed message in one main frequency period; and the power supply monitoring module ensures that power supply is maintained for more than or equal to 5ms after sudden power failure through the energy storage capacitor, so that complete transmission of fault messages is ensured. The system has the advantages that the system supports multi-scale out interface redundancy, the end-to-end delay of fault detection and backup path switching is smaller than or equal to 1ms by adopting a low-power-consumption switching chip and independent minimum power domain design, the fault response speed and reliability of an AI server scale out network are remarkably improved, and the system is suitable for guaranteeing the network continuity of a high-density AI training cluster.

The invention relates to the technical field of Internet of Things and water conservancy informatization, and discloses a water conservancy Internet of Things system and a control method thereof, and the system comprises an Internet of Things management platform, a business system, a water conservancy intelligent device, and a message queue service, a real-time database, a real-time monitoring system and an Internet of Things gateway which are in communication connection with the Internet of Things management platform. The business system is in communication connection with the real-time monitoring system and the message queue service, the real-time monitoring system is in communication connection with the message queue service, and the water conservancy intelligent equipment is in communication connection with the Internet of Things gateway. The problems of poor compatibility and the like in the prior art are solved. Through the cooperation of the protocol conversion service and the Internet of Things gateway, the plug-and-play of more than 90% of industrial equipment in an irrigation area is supported; the message queue service is combined with the MQTT protocol, so that the end-to-end delay is reduced to be less than 50ms; based on strategy optimization of data driving, manual intervention is reduced, and the scheduling efficiency is improved by more than 30%.

The invention provides a call center system and a multi-language intelligent voice translation method thereof. The method belongs to the technical field of artificial intelligence and natural language processing. The method comprises the following steps: carrying out multi-language interaction scene division on a call center system to generate scene classification data; a streaming processing assembly line framework is constructed; the continuous voice is dynamically fragmented into microsecond-level fragments, an ASR module is decoded in real time through an edge computing node, and initial text data is generated; and matching a context memory bank, extracting historical semantic vectors of the first five rounds, and generating a context enhanced text by joint coding. Obvious technical breakthrough and benefit improvement are realized through the technologies of streaming processing assembly line, context enhanced translation, seat correction instant learning and the like; the end-to-end delay is smaller than or equal to 1.5 s (P95), the delay problem in a traditional serial processing mode is greatly reduced, and the strict requirement for real-time interaction is met.

The invention relates to the technical field of computer networks, and discloses a DPDK-based high-speed data isolation method, which comprises the following steps of: S1, directly taking over a data packet transceiving function of a network card through a DPDK driver, bypassing a kernel protocol stack of an operating system, and establishing a user mode data channel; s2, based on a multi-core CPU architecture, distributing the received data packets to independent processing cores according to a preset rule, and binding each processing core to an exclusive memory pool to realize physical-level data isolation; and S3, directly transmitting the isolated data to a target application module through a zero copy technology in the user mode, and verifying the data integrity in real time through a hardware verification module. According to the method, the network card is directly taken over through the DPDK driver, the user state data channel is established, an operating system kernel protocol stack is bypassed, the zero copy technology is combined, data does not need to be copied in the whole process from receiving, processing to transmission, the end-to-end delay is reduced to 50 microseconds or below, and the method is suitable for a high throughput scene.

The embodiment of the invention provides a server-free function deployment method for an air-space-ground integrated network, which is applied to the technical field of edge computing, and comprises the following steps: constructing a three-layer network model of the air-space-ground integrated network, and determining information of nodes and links thereof; establishing a function dependency graph of a server-free application and generating a task request queue; modeling a server-free function placement problem in the space-air-ground integrated network as a weighted sum of minimum request execution cost and end-to-end delay, and designing a constraint condition and a target function; establishing a reward function required by reinforcement learning according to the target function; and obtaining an optimal server-free function placement strategy through a deep reinforcement learning algorithm based on graph perception embedding. According to the method, the end-to-end execution cost and time delay are reduced, adaptive optimization can be carried out for the dynamic nature of the network and the heterogeneity of node resources, and it is ensured that a delay sensitive task obtains stable and high-quality services in a complex scene.

The invention discloses a GB / T28181 monitoring video stream efficient transmission and decoding method based on HTTP3 (Hyper Text Transfer Protocol 3) and electronic equipment. According to the method, the problems of high delay and high resource consumption of a traditional GB / T 28181 video monitoring Web end playing scheme based on WebSocket are solved by utilizing a new generation Web API (Application Program Interface) of WebTransport and WebCodec. According to the invention, through the UDP transmission capability of WebTransport based on HTTP / 3, the direct transmission of the RTP data packet from the server to the Web client is realized, and protocol conversion does not need to be carried out; the WebCodes is adopted to realize efficient hardware acceleration decoding of the client; and meanwhile, a multiplexing real-time monitoring and self-adaptive video quality control mechanism is provided. Compared with the traditional scheme, the method has the advantages that the end-to-end delay is remarkably reduced, the resource consumption of the server is reduced, the concurrency capability is improved, the video parameters can be dynamically adjusted according to the network condition, and the method is suitable for video monitoring application in various network environments.

The invention relates to the technical field of communication, in particular to a communication system fusing a Wi-Fi network and a TSN and a control method, and a wireless access point in the communication system is configured to obtain an actually measured signal-to-noise ratio, an MCS table and preset parameters; when the packet error rate is smaller than or equal to the target packet error rate, determining a target MCS in the MCS table; cr (MCSi) is obtained; obtaining Nmin (r); acquiring a Nud; if Nmin (r) is smaller than or equal to Nud, Qi is obtained; determining Nmin (r) first target time slots in the current communication period; and sending data to the gateway through the Nmin (r) first target time slots, the target MCS and the Ru of the ith specification. According to the scheme, jitter occurring at the moment when the data arrives at the gateway can be effectively reduced, so that the situation that subsequent data is blocked due to the fact that the cache space of the gateway is occupied is avoided, time delay generated by data transmission is avoided, and end-to-end delay is reduced.

The invention provides a multi-source data fusion state estimation method, device and equipment, and relates to the technical field of state estimation, and the method comprises the steps: S1, collecting original observation data from N heterogeneous sensors, writing a local hardware timestamp for each piece of original observation data, and converting the timestamp into a second timestamp under a unified reference time base through master-slave clock bidirectional synchronization to obtain an original data stream with a unified time scale. According to the multi-source data fusion state estimation method, device and equipment, sensor time delay is used as an explicit variable for online estimation, and late data is digested by using an independent compensation channel. The main filter always runs according to a high-frequency beat without pause and rollback, and the end-to-end delay of the system is less than 5 milliseconds; the sliding window can be dynamically expanded, the delay error is converged to a sampling interval through correlation sampling interpolation, and the state precision is improved by about 30% compared with that of a traditional loose coupling method.

The invention provides a perception adaptive video stream code rate decision-making method and system based on a neural symbol reinforcement learning buffer area, and belongs to the technical field of computer network and multimedia transmission. According to the method, a neural symbol reinforcement learning agent is constructed, a comprehensive state vector is generated by collecting a playing state, video content information and user preference, and double decisions of code rate selection and buffer area adjustment are jointly output; security rules are formalized in a first-order logic mode, security exploration is achieved through a security monitor and a security layer strategy, and illegal risks such as underflow of a buffer area are avoided; the bandwidth prediction precision is improved by adopting HTTP / 3 stable connection sampling and a hybrid prediction model; and designing a multi-target reward function to balance user experience quality and buffer occupation optimization. According to the method, high-video-quality, low-lagging and low-delay transmission can be stably realized in a low-speed to 5G high-speed network, and unnecessary buffer accumulation can be remarkably reduced on the premise of ensuring smooth playing, so that end-to-end delay is reduced.

Disclosed is a fifth generation (5G) or sixth generation (6G) communication system for supporting higher data transmission rate. A method for providing an ultra-reliable and low-latency communication (URLLC) service in an ultra-reliable and low-latency communication function (URLLCF) of a mobile communication system is provided. The method includes receiving an end-to-end latency request message from an application function (AF) device, the latency request message including a latency requirement and at least one of generic public subscription identifier (GPSI) information of a specific user equipment (UE), data network name (DNN), or single-network slice selection assistance (S-NSSAI), performing a subscription procedure of URLLC service condition with a unified data management (UDM) / user data repository (UDR) device, obtaining location information of the UE, identifying whether the location information conforms to a range required by the URLLC service, configuring a policy and association satisfying a latency for the URLLC service to the UE, and providing a URLLC service notification to the AF based on the configuration.

The invention relates to a semantic-driven collaborative edge computing double-time-domain caching and computing power collaboration method and system, and belongs to the technical field of mobile communication and edge computing crossing. Aiming at the problems of black box of semantic fidelity, insufficient space coupling among multi-edge servers, difficulty in dual-time scale resource collaboration and the like in the collaborative multi-antenna semantic perception MEC system, the method comprises the following steps of: establishing an uplink MEC system model, and defining semantic perception QoE indexes of coupling semantic similarity and end-to-end delay; according to the method, an optimization problem taking long-term reward maximization as a target is constructed, an SCALE framework is designed, long-term cache placement is optimized by combining a DDPG agent enhanced by a long short-term memory (LSTM) network, and short-term semantic compression, task unloading, server allocation and joint power computing resource optimization are processed by a real-time PPO integrated decision RAPID algorithm. Simulation results show that the method is superior to an existing baseline scheme in the aspects of convergence speed, long-term reward, average QoE and cache switching overhead, and the system performance and the resource utilization rate are improved.

A channel access process to meet increasingly stringent low-latency WLAN requirements by mitigating channel access contention, through a cooperative TXOP sharing process between stations, while maintaining channel access flexibility for non-AP stations. When a non-AP STA obtains a Transmit Opportunity (TXOP) as 'TXOP holder', its TXOP can be shared in part with other non-AP STAs, as shared non-AP STAs; allowing the shared non-AP STAs to more efficiently transmit and utilize the channel during the shared TXOP without contending with each other for the channel. Simulations show the process provides significant mitigation of End-to-End (E2E) delay and allows improved spectrum utilization by reducing overall network congestion.

The application provides a target perception-based adaptive cutting transmission method for satellite-borne remote sensing images, and relates to the technical field of satellite remote sensing data acquisition and processing. The application changes the full-width transmission mode to only transmit the region of interest, greatly reduces the invalid data transmission amount, and effectively alleviates the bandwidth bottleneck of the satellite-ground transmission link. The end-to-end delay from the user's demand to the receipt of the target region data is significantly shortened. The traditional "full-width shooting -> storage -> transit transmission -> ground extraction" mode is changed to the "shooting and identification -> immediate cutting -> preferential transmission" mode, and the response speed in the emergency scene is improved.

The invention discloses a low-earth-orbit satellite communication network routing method based on reinforcement learning, and relates to the field of low-earth-orbit satellite communication networks in 5G NTN, in particular to the low-earth-orbit satellite communication network routing method based on reinforcement learning. The objective of the invention is to solve the problem that existing reinforcement learning Q-Learning applications are mostly focused on a ground network, and the unique dynamic topology and three-dimensional space routing problems of a satellite network are not solved. The invention provides a lightweight and adaptive low earth orbit satellite communication network routing optimization method based on reinforcement learning. The method is realized through distributed Q-Learning: dynamically adapting to topological change, dynamically adapting to three-dimensional space routing, balancing link load, minimizing end-to-end delay, and reducing on-satellite computing overhead.

The invention discloses a real-time audio and video end-to-end delay statistical method based on two-way time synchronization, and particularly relates to the technical field of real-time audio and video communication and network performance measurement, which comprises the following steps: S1, time synchronization: a plug flow client and a pull flow client respectively calculate clock skew with an SFU server through two-way message exchange; the method comprises the following steps: S1, a flow pushing end converts an acquisition timestamp and deviation into SFU reference time, and the SFU reference time and stage time consumption metadata are embedded into SEI information to be sent along with a video stream, S3, an SFU server forwards the video stream containing the SEI information, and S4, delay calculation and decomposition: a flow pulling end analyzes the SEI information, calculates total end-to-end delay in combination with the deviation of the flow pulling end, and decomposes the time consumption of each stage. Through high-precision time synchronization and time data losslessly carried by SEI information, non-intrusive, automatic, full-link-covered and decomposable end-to-end delay accurate statistics is realized, and the bottleneck problems that a traditional method is high in invasiveness, low in precision and incapable of positioning are effectively solved.

A receiver is configured is configured to receive a data signal, the data signal having at least one data signal block, the data signal block having a predefined duration in the time domain, a number of symbols in the time domain and a bandwidth in the frequency domain, wherein the data signal block is received over a plurality of transmission intervals, wherein a length of one transmission interval is shorter than the duration of the data signal block so that a subset of the symbols of the data signal block is received over one transmission interval, and wherein the receiver includes a signal processing device configured to process the symbols of the data signal block which have been received over one transmission interval.

The invention discloses a video frame transmission method, a video frame packaging method, a video frame packaging system and a storage medium based on a private protocol and packet loss compensation, which are mainly applied to the field of professional video transmission with strict requirements on delay and reliability. The core of the method is as follows: a sending end encapsulates each frame of video into a UDP data packet sequence carrying frame-level global information and packet-level sequence information according to a private protocol; a receiving end carries out real-time packet loss compensation and frame recombination through CRC (Cyclic Redundancy Check) and an intra-frame adjacent effective packet data replication mechanism, and traditional retransmission causing high delay is replaced; the system adapts to a wireless network and a wired network at the same time, and supports dynamic MTU adjustment; according to the invention, the sub-frame level transmission with the end-to-end delay stability lower than 16.7 milliseconds is realized; the broadcast-level visual lossless quality can still be maintained in a 5% random packet loss environment; and through deep integration of a security mechanism and a transmission process, copyright protection requirements of HDCP2.2 and the like are met, and system processing delay and implementation complexity are remarkably reduced.

Disclosed is a network device (1) for packet switching in accordance with a bounded end-to-end delay. The network device comprises a plurality (11) of n first-in first-out, FIFO, queues, being servable in accordance with a round-robin based service policy and a fixed packet processing time (T) . A respective queue of the plurality is associated with a bounded delay (D max) depending on the service policy of the plurality of queues and being a function of an adaptable buffer capacity (B e) of the respective queue. The network device further comprises a processor (12) , being configured to determine a threshold crossing of an extent of reservation of the buffer capacity of the respective queue; and to adapt the buffer capacity of the respective queue in accordance with the determined threshold crossing. This enables network devices of a network to trade off QoS and capacity locally in a distributed manner, without a centralized controller.