36results about How to "Improve resource scheduling efficiency" patented technology

The present invention provides a resource scheduling method based on the Kubernetes container cluster management system, comprising the following steps S1, traversing all available Nodes, and calculating the computing resource score of each available Node; S2, collecting the network of each available Node through a log system Store the average transmission rate and the network storage average transmission rate sum of all available Nodes; S3, calculate the network storage average transmission rate score of each available Node; S4, according to the calculation resource score of each available Node in step S1 and step S3 The network stores the average transfer rate score, calculates the priority of available Nodes, and creates and runs Pods on the Node with the highest priority. In the present invention, the scheduler adds the average transmission rate score of the network storage, and calculates the priority of the available Node in combination with the computing resource score of the available Node, and schedules the new Pod to the available Node with the highest priority, thus improving the efficiency of new Pods. Pod speed and improve the efficiency and stability of container operation.

The invention discloses a method, device and electronic equipment for dynamic scheduling of system data resources. The method includes counting all request data received within a first preset time length, and generating a unique call based on each request data information corresponding to each request data. String; select the target request data one by one in each request data group in turn, and after constructing the request data sequence table based on the selection sequence of each target request data, divide the request data sequence table in sequence based on random numbers; receive each of the request processing nodes Based on the request processing results fed back, the node processing time curve is fitted for each request data group based on the request processing results; The allocation request data is allocated to the request processing node with the shortest estimated processing time. The invention improves the efficiency of resource scheduling and ensures the optimal solution of global resource scheduling.

The present application relates to a resource scheduling method, device, computer equipment and storage medium. The method relates to the field of cloud technology, and the method includes: receiving a service association request, where the service association request includes a service identifier and a policy identifier; responding to the service association request, associating a target service corresponding to the service identifier with a target policy object corresponding to the policy identifier; Execute the scheduling policy associated with the target policy object for the target service; receive the resource scheduling request for the target service; respond to the resource scheduling request, and schedule the resources that meet the scheduling policy to run the target service. Adopting this method can ensure the two-way selection between resources and services, reduce the workload and difficulty of operation and maintenance personnel, and improve the efficiency of resource scheduling.

The invention discloses a MU-MIMO-based proportional fair scheduling method for flow distribution, which mainly solves the problem of proportional fair scheduling ignoring user service quality in the prior art. The implementation steps are: 1. Set the corresponding parameters according to the actual scheduling problem; 2. Calculate the user weight in the utility function according to the mean value of the emission probability satisfying the normal distribution; 3. Generate the sample value of the emission probability and calculate it according to the user weight The utility function value of the sample; 4. Filter out the sample value with good performance to update the mean value of the normal distribution; 5. Record the change of the mean value and switch the variance of different stages according to the corresponding rules; 6. End the iterative process according to the termination condition, A scheduling result that satisfies proportional fairness is obtained. The present invention can obtain satisfactory dispatching results under both sufficient and insufficient available wireless resources, has low computational complexity and fewer iterations, improves dispatching efficiency of base stations, and can be used in a single base station system using MIMO technology.

A CQI (Channel Quality Indicator) adjusting method based on position information and a base station belong to the technical field of wireless communication. The base station partitions the covering area firstly, determines the reference CQI corresponding to each partitioned small area according to specific rules, and saves each small area and the reference CQI corresponding to each small area; then position information of a mobile station is predicted in scheduling time according to the obtained position information of the mobile station, the small area where the mobile station is positioned in the time can be determined according to position information of the mobile station in scheduling time, and the reference CQI corresponding to small area that the mobile station is positioned in scheduling time is determined by finding each saved small area and the reference CQI corresponding to each saved small area; and finally, the reference CQI corresponding to small area that the mobile station is positioned in scheduling time can be adjusted according to the number of acknowledgment signals or disacknowledgment signals. The feedback overhead and the complexity of the mobile station are reduced, and the matching with a real channel environment can be ensured at the same time.

The present application provides a resource scheduling method, a terminal device and a network side device. The method includes: the terminal device sends first user plane data to the network side device, and the first user plane data carries application state information of the terminal device; the application state information is used to characterize the application status of the terminal device Running state: the terminal device obtains from the network side device the time-frequency resources scheduled by the network side device for the terminal device according to the application state information. With the method of the present application, the application state information does not need to be transmitted across the RRC layer, which improves the resource scheduling efficiency of the network side device, and also does not require the development of the terminal operating system or the baseband chip in the terminal device and the network side device from the operating system to the RRC The version of the transmission channel or transmission interface of the layer reduces the development cost of the terminal equipment.

The invention provides a Kubernetes scheduling method, device and storage medium, which filters out Node nodes that do not meet the requirements according to a pre-selection algorithm; sets queues for each Node node according to resource types; assigns priorities to Pods to be scheduled; analyzes Pods to be scheduled Status; when the cluster resources are satisfied, the Pod calculates the similarity according to the first rule, and searches for the first Node node for scheduling; when the cluster resources are not satisfied, it enters preemptive scheduling, sets the first rule to calculate the similarity and the second rule priority preemption rule, and searches for The Node node performs preemptive scheduling; set the maximum number of cycles for Pods that cannot be scheduled, and when the maximum is reached, priority is given to scheduling. The invention solves the problem that the priority is set regardless of the actual situation of the cluster resources, the resources cannot be reasonably arranged during the preemption process, and the scheduling cannot be completed all the time. The invention improves the integrity of task operation and the load balance of the entire cluster, reduces the generation of resource fragments, improves the timeliness of scheduling, and enhances the efficiency of resource scheduling in the cluster management system.