22 results about "Active queue management" patented technology

Devices, networks, systems, methods, and processes for identifying, classifying, and scheduling Low Latency, Low Loss, and Scalable throughput (LAS) traffic are described herein. A device may receive a data flow comprising one or more data packets. The device may parse the one or more data packets to check for a congestion indicator. Upon detecting the congestion indicator in any data packet, the device can classify or mark the data flow as LAS. The device may schedule the data flow by utilizing an Active Queue Management (AQM) queue for L4S data flows. The device can also mark other uplink or downlink data flows associated with the data flow as L4S. The device can also facilitate LAS capability sharing while associating with one or more wireless devices. The device may utilize a management frame with an extended capabilities field to inform the LAS capability of the device.

An RDMA queue pair multiplexing method based on a shared memory efficiently maps a large number of logical queue pairs (LQPs) to a small number of physical queue pairs (PQPs), and combines shared memory zero-copy communication and O (1) active queue management, so that ten thousand-level LQPs are supported by a small number of PQPs, the overhead of a single-connection memory is reduced by about 60%-64%, and context jitter of a network interface card (NIC) is avoided; the connection establishment time delay and the short flow end-to-end time delay are greatly reduced; through O (1) active tracking and batch processing, CPU idling and system calling overhead are greatly reduced, and high throughput and expandability are achieved; the method can be implemented in a pure user mode, a kernel or NIC does not need to be modified, and the method can be immediately deployed in an existing RDMA environment.

A method for handling data packets by a communication node in a communication network, the method comprising storing received data packets in at least two physical queues, wherein a first of said at least two physical queues is associated with low latency data packets and a second of said at least two physical queues is associated with high latency data packets, wherein each data packet is stored in one of the at least two physical queues based on a delay characteristic associated with the data packet, for each received data packet, storing an associated information record in at least two virtual queues, VQs, wherein associated information for data packets stored in said high latency physical queue is stored in a second of said at least two virtual queues and wherein associated information for data packets stored in said low latency physical queue is stored in both said first and second of said at least two virtual queues, serving data packets from the at least two physical queues, using at least two Congestion Threshold Values, CTVs, wherein a first of said at least two CTVs is applicable to data packets in said low latency physical queue and wherein both said first and second of said at least two CTVs are applicable to data packets in said low latency physical queue and data packets in said high latency physical queue, wherein said at least two CTVs are used for at least one of dropping and marking packets based on their associated information.

The application discloses a kind of based on the naming data network hop-by-hop congestion control method of forwarding path number, including router congestion detection and router congestion control;Router calculates the congestion degree of local router by means of active queue management method.Routers determine local congestion state and update the congestion information of data packet.Downstream adjacent router is based on the number of forwarding path and executes multipath forwarding strategy or interest packet sending rate adjustment strategy.When router executes multipath forwarding strategy, the router calculates the shunt proportion of interest packet flow of current interface according to the congestion information of data packet and forwards flow through current or relatively suboptimal interface.When router executes interest packet sending rate adjustment strategy, the router adjusts the interest packet sending rate of same name prefix according to the congestion information of data packet.The application proposes two router control strategies for the characteristics of multiple-source multipath transmission of named data network, to cope with single-path flow and multi-path flow in network.

A network entity (900, 1000) can determine (710) that a sequence number, SN, gap will be created. The network entity can further determine (720) a packet data convergence protocol, PDCP, SN of a first PDCP data protocol data unit, PDU, packet. The network entity can further 5 replace (730) a second PDCP data PDU packet with a PDCP control PDU including an indication of the PDCP SN of the first PDCP data PDU packet.

The invention provides an FPGA-based multi-queue scheduling system and a congestion avoiding method. The FPGA-based multi-queue scheduling system comprises a data packet queue analysis module, an active queue management module, an on-chip queue module, an off-chip queue module, a DEMUX module and an MUX module. The threshold value and the depth of each queue are established and maintained on a chip, the cache states of the on-chip and off-chip queues are combined as congestion information, input data packets are judged and distributed, the data cache requirement is met under the condition of relieving congestion, and meanwhile the transmission efficiency is considered. The storage spaces of the on-chip queue and the off-chip queue are configurable, and the flexibility and the programmability are good; the active queue management performance depends on the receiving performance of the on-chip cache module and the off-chip cache module, the judgment process does not affect the pipeline performance, and the overall output efficiency during data stream merging output can be considered while the data cache requirement is met.

A queue management system (including but not limited to active queue management systems, traditional networks, and / or software-defined networking), method, computer program product, or integrated circuit that selectively manages flows based on methods for how the data of the flows have been encoded (such as FEC and / or compression) and / or characteristics of the flows (including but not limited to loss-recovery capabilities). Similarly, an encoder system, method, computer program product, or integrated circuit can make encoding decisions based on known and / or anticipated queue management policy (e.g., an FEC encoder may be configured to make frame splitting, parity symbol allocation, and / or packetization decisions based on known and / or anticipated queue management policy, and a data compressor may be configured to make data compression decisions based on known and / or anticipated queue management policy).

Described herein are devices, networks, systems, methods, and procedures for identifying, classifying, and scheduling low-latency, low-loss, and scalable throughput (L4S) traffic. A device can receive a data flow comprising one or more data packets. The device can parse the one or more data packets to check for a congestion indicator. Upon detecting the congestion indicator in any of the data packets, the device can classify or mark the data flow as L4S. The device can schedule the data flow by utilizing an active queue management (AQM) queue for L4S data flows. The device can also mark other uplink or downlink data flows associated with the data flow as L4S. The device can also facilitate L4S capability sharing while associated with one or more wireless devices. The device can utilize a management frame with an extended capabilities field to advertise the L4S capability of the device.

The present application discloses an active queue management method for a network device and a network device, which can be applied to the field of data transmission, including: the network device calculates the idle forwarding capacity (such as a letter of credit counter CC) of a target outgoing queue or a target outgoing port on the network device, and the idle forwarding capacity is used to characterize the difference between the upper limit of the forwarding data volume of the target outgoing queue or the target outgoing port and the current actual forwarding volume. After that, the network device obtains a first message carrying an adjustment request, and when the first message does not have a preset tag, obtains a decision result based on the idle forwarding capacity and the adjustment request, and carries the decision result in the first message to obtain a second message. In the embodiment of the present application, the sending end device adjusts the sending window / rate in different ways according to different decision results, and the new AQM algorithm proposed in the present application can realize the decoupling of the queue depth on the network device and the number of flows on the bottleneck link, so that the queue depth on the network device can be customized.

The application provides a kind of active queue management method based on dynamic bandwidth perception, comprising: S1, real-time response data packet arrival, obtain the available bandwidth value of current network link and the length of data packet queue;S2, according to the available bandwidth value of current network link and the preset target forwarding delay, calculate the current congestion threshold;S3, compare the length of current data packet queue and the size of current congestion threshold, when the length of current data packet queue is greater than or equal to current congestion threshold, it is congestion state;When in congestion state, further judge whether the data stream to which the current arrived data packet belongs is aggressive flow, if it is aggressive flow, then according to the preset packet loss method, the current arrived data packet is randomly discarded.The application maps the changing link available bandwidth into dynamic threshold to manage queue, so that network device can efficiently utilize network bandwidth;Through dynamic identification of aggressive flow and targeted setting of packet loss strategy, network device can provide forwarding service more fairly.

The invention discloses an active queue management method based on a deep Q network and adopting multi-GPU parallelization, and the method comprises the following steps: determining a data scale F and a GPU number G, and starting an MGPUSimm kernel; extracting playback data from the experience pool and loading the playback data to a global memory of the MGPUSimm; loading playback data from a global memory of the MGPUSimm to a multi-level cache, segmenting and issuing the playback data to each GPU; the GPUs complete forward propagation and reverse propagation simultaneously, independently and in parallel; calculating results of the GPUs are output to a global memory to complete integration, and DQN model parameters are uniformly updated; and identifying the performance bottleneck and verifying the improvement. The invention aims to provide an active queue management method based on a deep Q network by adopting multi-GPU parallelization, and aims to solve the current situation that a current active queue management method based on a DQN model is too slow in training speed and too low in parallelization degree. The training process of the original serial processing DQN model is expanded into highly parallel calculation tasks on the data level, so that the training efficiency of the DQN model is remarkably improved, and the purpose of solving the current situation that the training efficiency is low in the prior art is achieved.

A queue management system (including but not limited to active queue management systems, traditional networks, and / or software-defined networking), method, computer program product, or integrated circuit that selectively manages flows based on methods for how the data of the flows have been encoded (such as FEC and / or compression) and / or characteristics of the flows (including but not limited to loss-recovery capabilities). Similarly, an encoder system, method, computer program product, or integrated circuit can make encoding decisions based on known and / or anticipated queue management policy (e.g., an FEC encoder may be configured to make frame splitting, parity symbol allocation, and / or packetization decisions based on known and / or anticipated queue management policy, and a data compressor may be configured to make data compression decisions based on known and / or anticipated queue management policy).

The application discloses a naming data network hybrid congestion control method based on data packet queuing delay, which comprises router congestion information updating, first, the congestion degree of a local router is calculated by means of an active queue management method; the router discriminates the local congestion state according to the congestion degree and updates the congestion information of the data packet; in the router congestion control, the residual flow is forwarded through the current forwarding interface or a relatively suboptimal interface according to the forwarding proportion or the local congestion degree, the data packet queuing delay and the consumer congestion window adjustment value are updated into the data packet through interest packet sending rate control; the pre-control before the internal congestion of the network and the control after the congestion are realized through the consumer end control. The application cooperates the consumer and the router control, can cope with the single path flow and the multi-path flow in the network, and realizes the congestion avoidance and the congestion control through the consumer end control.

In one aspect, an electronic active queue management system (AQMS) may be used to identify processing capacity of entry points and quantifiable amounts of congestion at a land port of entry (LPOE) based on one or more real-time measurements of traffic flow. Based on the amount(s) of congestion, the AQMS may dynamically manage arrival times at the LPOE through one or more lanes of the LPOE, controlling wait times in the queue system. As part of managing arrival times at the LPOE, the AQMS can schedule arrival times for vehicles and / or people at the dedicated lane of the LPOE. The arrivals may be scheduled through a dedicated AQMS app, through a web-based interface, and / or through telephone text messaging. Present principles may also be used for queue management at airports, concerts, sporting events, etc.

In one aspect, an electronic active queue management system (AQMS) may be used to identify processing capacity of entry points and quantifiable amounts of congestion at a land port of entry (LPOE) based on one or more real-time measurements of traffic flow. Based on the amount(s) of congestion, the AQMS may dynamically manage arrival times at the LPOE through one or more lanes of the LPOE, controlling wait times in the queue system. As part of managing arrival times at the LPOE, the AQMS can schedule arrival times for vehicles and / or people at the dedicated lane of the LPOE. The arrivals may be scheduled through a dedicated AQMS app, through a web-based interface, and / or through telephone text messaging. Present principles may also be used for queue management at airports, concerts, sporting events, etc.

The invention provides a dynamic priority dual-channel active queue management method, which comprises the following steps that: data packet forwarding is carried out based on a configured dual-channel forwarding queue, the dual-channel forwarding queue comprises a main queue and an auxiliary queue, and the forwarding mode comprises the following steps: judging whether the main queue is in a congestion state or not in real time; when the main queue is in a congestion state, judging whether each newly arrived data packet needs to be jumped, and when the newly arrived data packet needs to be jumped, putting the newly arrived data packet into an auxiliary queue so as to preferentially forward the newly arrived data packet before all the data packets of the main queue are forwarded; and when the main queue is in a non-congestion state or queue jumping is not needed, putting the newly arrived data packet into the main queue for queuing and forwarding. According to the technical scheme of the invention, aiming at a congestion scene, a part of newly arrived data packets can be temporarily guided to the auxiliary queue, so that the congestion pressure is relieved, and key services are ensured not to be influenced by common service congestion.

A wireless IEEE 802.11 queuing architecture supporting Low Latency Low Loss Scalable throughput (L4S) and Dual Queue Active Queue Management (AQM) in which queues for both L4S streams and classic streams are contained in the medium access control (MAC) layer, and one or more of these queues are configured with more than one internal queue, to which different L4S packets and other packet types are retained in different internal queues within the MAC transmit queues. The MAC queues are configured for MAC layer buffering of L4S packets and classical packets in separate queues and / or queues having internal queues for buffering different packet types, classifications, latency sensitivity, QoS considerations, and so forth. The architecture also allows for multiple packet queues in the upper layers above the MAC layer.

Described herein are solutions for reducing packet loss for low latency, low loss, scalable throughput (LAS) traffic. A user equipment (UE) can determine whether a base station is capable of supporting LAS traffic and can implement rate adaptation in different ways based on the LAS capabilities of the base station and active queue management (AQM) policies of the base station. A base station can establish protocol data unit (PUD) sessions between the UE and different instances of a network slice based on whether the base station receives packets from applications capable of LS4 rate adaptation. These and many other features and examples are described in additional detail.

A wireless IEEE 802.11 queuing architecture supporting Low Latency Low Loss Scalable throughput (L4S) and Dual Queue Active Queue Management (AQM) in which queues for both L4S streams and classic streams are contained in the medium access control (MAC) layer, and one or more of these queues are configured with more than one internal queue, to which different L4S packets and other packet types are retained in different internal queues within the MAC transmit queues. The MAC queues are configured for MAC layer buffering of L4S packets and classical packets in separate queues and / or queues having internal queues for buffering different packet types, classifications, latency sensitivity, QoS considerations, and so forth. The architecture also allows for multiple packet queues in the upper layers above the MAC layer.

The invention relates to an apparatus and method for queue management. In accordance with an example aspect of the present disclosure, there is provided an apparatus configured to receive first configuration information from a network device, the first configuration information indicating a first threshold for marking uplink packets for explicit congestion notification, and to use the first threshold in an active queue management mechanism of the apparatus, to control uplink packet marking for the explicit congestion notification.

The invention relates to the technical field of low-altitude unmanned aerial vehicle communication, in particular to an intelligent active queue management method for low-altitude unmanned aerial vehicle relay transmission, and aims to overcome the defect that traditional queue management is one-cut in a dynamic environment and a multi-service scene. The invention provides an intelligent active control scheme based on a SAC (Soft Actor-Critic) deep reinforcement learning framework. An intelligent agent senses transmission queue characteristics and channel quality in real time and outputs continuous actions to finely schedule transmission opportunities of data packets with different priorities; the multi-target reward function synchronization guarantees the low-delay demand of the high-priority data packet and punishes low-efficiency transmission, so that the energy consumption is reduced; after knowledge distillation compression, the data are embedded into unmanned aerial vehicle edge hardware, millisecond-level real-time decision and low-power-consumption operation are realized, collaborative optimization of communication efficiency and energy efficiency is achieved, and a sustainable intelligent solution is provided for low-altitude relay.

Described herein are solutions for reducing packet loss for low latency, low loss, scalable throughput (LAS) traffic. A user equipment (UE) can determine whether a base station is capable of supporting LAS traffic and can implement rate adaptation in different ways based on the LAS capabilities of the base station and active queue management (AQM) policies of the base station. A base station can establish protocol data unit (PUD) sessions between the UE and different instances of a network slice based on whether the base station receives packets from applications capable of LS4 rate adaptation. These and many other features and examples are described in additional detail.