7 results about "Qos scheduling" patented technology

The invention discloses a data transmission system for edge network cloud, and relates to the field of cloud computing. The data transmission system for the edge network cloud comprises an edge node layer, a network transmission layer and a cloud center layer which work cooperatively through standardized interfaces in sequence, the edge node layer comprises at least one edge node, and each edge node is provided with an edge computing power scheduling unit, a data preprocessing unit and a local cache unit; the network transmission layer is configured with a dynamic routing unit, a QoS scheduling unit and an end-to-end encryption unit; and the cloud center layer is configured with a global resource scheduling unit, a data analysis unit and a security management unit. According to the data transmission system for the edge network cloud, the tasks are dynamically unloaded to the adjacent nodes or the cloud center, overload of a single node is avoided, the overall processing efficiency of the edge layer is improved, and the reasonability and timeliness of task unloading are ensured through the preset threshold value and the distance priority strategy.

The invention provides a distributed edge machine room computing power integration method and system based on a DPU, and relates to the field of processors, and the method comprises the steps: transmitting parameter copy management data and consistency control function data to a DPU layer corresponding to a DPU card deployed in an edge machine room, and obtaining a target DPU layer; constructing a hardware hierarchical communication topology combining aggregation in the machine room and cross-machine room controlled transmission in the target DPU layer to obtain a target DPU, and executing bandwidth sensing, QoS scheduling, cross-machine room parameter coordination and traffic shaping control through the target DPU; the gradient processing and parameter synchronous control logic data corresponding to the edge machine room are transmitted to a target DPU, and hardware-level unloading and cross-machine-room intelligent scheduling are carried out through the target DPU; gPU resource virtualization and network virtualization control are carried out through the target DPU, so that the multiple edge machine rooms form a unified training resource pool logically, and a distributed edge machine room computing power integration result is obtained.

The application discloses a terminal-oriented VLAN classification access and QoS scheduling method, device, equipment and medium, and relates to the technical field of wireless communication. The method comprises the following steps: acquiring full-dimension characteristic information of a terminal initiating a network access request; obtaining a corresponding level of the terminal according to the full-dimension characteristic information; determining a target VLAN partition corresponding to the terminal through a first mapping table based on the corresponding level; performing security check and quality check on the terminal according to the target VLAN partition; if the security check and the quality check are both passed, allowing the terminal to access the target VLAN partition, and allocating network access parameters of the target partition to the terminal; matching a corresponding QoS scheduling strategy from a second mapping table according to the level attribute of the target VLAN partition; and performing scheduling on data flow of the terminal in the target VLAN partition according to the corresponding QoS scheduling strategy. The method can improve the accuracy of VLAN classification access and QoS scheduling of the terminal.

This invention provides a method and system for integrating distributed edge data center computing power based on DPU, relating to the processor field. The method involves: transmitting parameter replica management data and consistency control function data to the DPU layer corresponding to the DPU card deployed within the edge data center, obtaining the target DPU layer; constructing a hardware-layered communication topology combining intra-data center aggregation and cross-data center controlled transmission within the target DPU layer, obtaining the target DPU, which performs bandwidth awareness, QoS scheduling, cross-data center parameter coordination, and traffic shaping control; transmitting gradient processing and parameter synchronization control logic data corresponding to the edge data center to the target DPU, and performing hardware-level offloading and cross-data center intelligent scheduling through the target DPU; and controlling GPU resource virtualization and network virtualization through the target DPU, so that multiple edge data centers logically form a unified training resource pool, resulting in the distributed edge data center computing power integration.

This invention discloses a QoS scheduling method, apparatus, device, and storage medium for a switch, relating to the field of network communication technology. The method includes: performing deep detection on input data packets based on a preset service feature library to obtain classified data packets with priority tags; writing the classified data packets into corresponding priority queues based on the priority tags, and reserving exclusive buffer resources for target queues in the priority queues; dynamically calculating the weights of real-time bandwidth utilization and queue waiting time to obtain a target weight sequence; performing weighted round-robin transmission on non-target queues in the priority queues according to the target weight sequence; and interrupting weighted round-robin transmission when a data packet is detected in the target queue in the priority queue and the queue waiting time exceeds a preset delay threshold, prioritizing forwarding data packets from the target queue, thus ensuring the upper limit of real-time service latency and eliminating low-priority traffic starvation under high-load burst traffic conditions.

The invention discloses a lightweight traffic forwarding path planning method and system, and belongs to the technical field of wired communication. The method comprises the following steps: assigning one device in a network as a path planner which is responsible for path planning, configuration issuing and state maintenance, and only executing a forwarding rule by other forwarding nodes; and manually assigning a traffic forwarding path through a network topology interface of the browser, or dynamically generating a forwarding path by utilizing path planning software according to a real-time bandwidth occupation condition. On the premise that existing network hardware equipment is not upgraded, accurate management and control of a flow forwarding path are realized by dividing roles of a path planner and a common forwarding switch, customizing a path priority, designing a special control message and establishing a path maintenance and switching mechanism in combination with two modes of manual configuration and automatic software planning, so that the flow forwarding path can be accurately managed and controlled, and the flow forwarding efficiency is improved. The problems that a traditional routing protocol ignores bandwidth, QoS scheduling is extensive, and SDN reconstruction cost is high are solved.

The invention provides a multi-source feature fusion QoS (Quality of Service) scheduling method for a switch port, which relates to the technical field of QoS scheduling and comprises the following steps of: filtering a time sequence of a hardware counter port of a switch chip; generating a first type of low-level quantitative feature vectors; classifying the traffic data according to a preset network behavior semantic category, and constructing an interpretable network behavior semantic feature vector through clustering analysis; generating a fusion feature vector through linear weighted combination; performing normalization processing on the fusion feature vector to generate a terminal network grading result; a switch forwarding path, a queue scheduling parameter or a forwarding priority strategy is adjusted in real time based on a grading result, and self-adaptive QoS scheduling oriented to equipment behavior characteristics is realized; the invention effectively solves the problems that the real service form is difficult to reflect by depending on a single index, the QoS grading is rough, and the priority frequently jumps under the burst flow in the prior art.