9 results about "Queries per second" patented technology

The invention discloses a table information updating method and device, equipment, a medium and a product. The method comprises the steps of receiving a first editing operation for a first table, and obtaining a data adjustment type; wherein the data adjustment type is related to the first editing operation, and the first table is a table which can be collaboratively edited by at least two collaborative users; obtaining a data synchronization attribute of at least one online client based on the first query rate per second and the number of clients; and based on the data synchronization attribute of the at least one online client, performing content updating and displaying on the first table displayed by the at least one online client. According to the mode, the collaborative request for table synchronization is orderly processed, so that the effect of effective table synchronization is achieved.

This application discloses a method, apparatus, device, and storage medium for resource scheduling. The specific scheme includes: obtaining a first query per second (QPS) of a target service to be predicted within a preset first time period; determining a second QPS within a preset prediction time period based on the first QPS using a preset load prediction model; determining a first number of Pods corresponding to the second QPS using a preset capacity model, wherein the preset capacity model is trained and determined based on sample data of the service; and scheduling resources for the target service based on the first number of Pods and preset resource usage conditions. According to the embodiments of this application, resource scheduling can be made more reasonable and convenient, improving resource scheduling efficiency and thus enhancing the overall resource utilization of the service.

Aspects of the disclosure are directed to enhancing RAG through a structured schema for RAG that captures a wider range of RAG techniques to serve as a foundation for improving performance and an optimization framework for efficiently serving the wider range of RAG techniques based on significant performance variability across RAG workloads. The optimization framework may achieve up to a 2× increase or more in query per second (QPS) per chip compared to alternative RAG frameworks built on LLM-system extensions.

The invention discloses a cache dynamic preheating method, device and equipment based on service peak prediction and a medium, which are used for solving the technical problems that the prior art depends on artificial experience, is lagged in preheating, cannot avoid risks in advance, is low in prediction precision and weak in anti-interference capability, and a cache server is easy to overload. The method comprises the steps that when preheating prediction is triggered, service feature data and system resource data in a preset time period are collected; extracting a time feature, a service trend feature and an activity influence feature from the service feature data and the system resource data; inputting the time feature, the business trend feature and the activity influence feature into a pre-trained multi-dimensional peak value prediction model to obtain a predicted peak value time period, a predicted peak value per second query rate and a predicted hotspot key list; acquiring real-time system resources; and performing gradient preheating on the predicted hotspot key list according to the real-time system resources, the predicted peak value time period and the predicted peak value per second query rate.

This application provides a rate limiting method, apparatus, computer device, computer-readable storage medium, and computer program product, belonging to the field of data processing. The rate limiting method includes: receiving rate limiting parameters broadcast by a master node of a distributed system, the rate limiting parameters including at least a global start time; determining the current warming-up phase of the node based on the global start time and the current time, the warming-up phase corresponding to several phases with different request rates, where each subsequent phase has a higher request rate than the previous phase, the request rate being the rate at which a target service is called, and the node being any execution node of the distributed system; and dynamically adjusting the request rate of the node calling the target service based on the warming-up phase. The technical solution of this application embodiment can achieve a gradual warming-up of the QPS (Queries Per Second) of a distributed system calling a target service, improving the stability of the target service calls and the overall success rate of the task.

The embodiment of the invention discloses a log delivery system and method, and the method comprises the steps: obtaining a log from a source Kafka cluster through starting a consumer thread, distributing the log to a corresponding buffer queue according to a target cluster identifier in the log, and carrying out the log delivery through a producer thread in a thread pool corresponding to the buffer queue. The log located at the queue head of the buffer queue is delivered to the corresponding user Kafka cluster, and the capacity of the thread pool is dynamically adjusted according to the query number per second and / or the data backlog amount of the buffer queue, so that the log delivery from the source Kafka cluster to the user Kafka cluster can be realized through the multi-thread service, the data transmission path is simplified, and the data transmission efficiency is improved. And the delay of the log delivery process is effectively reduced.

Embodiments of the present application provide a user matching method and device, equipment and storage medium. In the embodiments of the present application, in response to a matching request of a target user, a matching grade of the target user is determined based on feature information of the target user; based on a query rate per second of the distributed cluster server at the time of the matching request and a query rate per second threshold of an ordered queue, a plurality of ordered queues are divided from a plurality of users to be matched in the distributed cluster server, and one ordered queue corresponds to one matching grade interval; from the plurality of ordered queues, a target ordered queue matched with the matching grade of the target user is determined; and the target user is matched with the user to be matched in the target ordered queue.

The invention relates to a fragment dynamic adjustment method and device, and the method comprises the steps: obtaining a configuration data fragment set corresponding to a region node, wherein the configuration data fragment set comprises at least one configuration data fragment; for any configuration data fragment in the configuration data fragment set, configuration data stored in the configuration data fragment is determined, and the service priority of the configuration data is searched; and adjusting the configuration data fragments according to the query rate per second and the service priority of the configuration data fragments. By acquiring the configuration data fragment set of the regional node and dynamically adjusting the fragment granularity in combination with the per-second query rate of each configuration data fragment and the service priority of the corresponding configuration data, the dynamic change of the data access heat can be accurately adapted, the problem of fragment node overload caused by centralized access of hotspot data is effectively avoided, and the data access efficiency is improved. The hotspot data processing delay is obviously reduced, and the system response efficiency is improved.

The invention provides an abnormal traffic processing method and device, equipment, a medium and a product, and relates to the technical field of Internet of Things security. The method comprises the steps of obtaining a target side traffic feature, and obtaining first target traffic according to the target side traffic feature and a preset feature database; according to the first target traffic and a pre-training traffic detection model, obtaining a probability that the first target traffic is abnormal traffic; when it is detected that the probability that the first target traffic is the abnormal traffic is greater than a first preset threshold value, determining the first target traffic as the abnormal traffic; acquiring the bit rate of the abnormal traffic; and when it is detected that the bit rate of the abnormal traffic is greater than or equal to a second preset threshold value, obtaining the number of queries per second of the abnormal traffic, and executing a first cleaning strategy on the abnormal traffic according to the number of queries per second of the abnormal traffic to obtain second target traffic. The abnormal traffic is identified and intercepted before the attack traffic reaches the target server, so that the protection delay problem is relieved, and the calculation and bandwidth pressure of the target server can be effectively relieved.