136 results about "Inter-server" patented technology

The invention relates to the technical field of server scheduling, and discloses a large model reasoning scheduling method based on an off-network computing power server, which comprises the following steps of: collecting hardware information of the off-network computing power server, and constructing a server resource vector; decomposing a large model reasoning task into a plurality of sub-tasks, and constructing task modeling; according to the hardware information of the off-network server and the task dependency relationship, performing task division by using an optimization method; performing dynamic task scheduling on the sub-tasks based on the current load state of the server; according to the real-time calculation load of the server, dynamically migrating part of tasks to other servers to ensure the load balance among the servers; and executing the distributed task, monitoring the running state of the server and the task completion condition, and dynamically adjusting a task distribution strategy according to feedback data. Through modeling task requirements and server resources, in combination with real-time monitoring and optimal scheduling, efficient calculation and resource utilization maximization of large model reasoning in an off-network environment are realized.

A resource set which includes multiple servers with a respective plurality of training computing devices is identified for training a machine learning model. The resource set is subdivided into partition groups, such that each partition group can store a respective replica of state information of the model. The model is trained using the partition groups. The training comprises a multi-stage gathering of a portion of the state information at training computing devices of a particular partition group. Different types of communication channels between training computing devices are used in respective stages of the gathering, including inter-server communication channels in one stage and an intra-server communication channel during another stage. A trained version of the model is stored.

The invention discloses a cluster job planning method and system based on a dual time window mechanism. The method comprises the steps that jobs are divided into a multi-level structure with an execution sequence, and a directed acyclic graph is constructed to form a task topological structure; constructing a dual time window mechanism; designing a multi-objective optimization function; modeling cluster job scheduling into a multi-agent Markov decision process, setting a hierarchical action space, and designing a reward function reflecting scheduling balance; the method comprises the following steps: constructing a hybrid neural network architecture, extracting a time sequence correlation feature of a task, encoding a heterogeneous feature between the task and a server by using a heterogeneous graph neural network, and generating an intelligent matching score between the task and the server; a multi-agent near-end strategy optimization algorithm is applied, and agent collaborative optimization is achieved through a collaborative mechanism updated by a centralized value function and a distributed strategy. According to the method, fine modeling and state dynamic perception of complex operation topology are realized, and scientificity and controllability of scheduling decision are improved.

The invention provides a physical guidance multi-level federated learning energy storage unit degradation estimation method for heterogeneous terminal equipment, and mainly relates to the field of artificial intelligence and energy storage health estimation. The method mainly comprises the following steps: taking power battery data of various devices such as an electric vehicle, an unmanned aerial vehicle and a robot as local data of a client, and performing data preprocessing; establishing a lightweight model for each client, and embedding a battery aging physical mechanism into a neural network model; a multi-layer federated learning framework is established, the first layer is a client formed by all devices, the second layer is local servers of multiple types of devices (electric vehicles, unmanned aerial vehicles and robots), and the third layer is a central server; designing a model screening strategy and an error perception weighted aggregation strategy between the client and the local server, and designing a model compression strategy between the local server and the central server; and obtaining a global model through multiple rounds of training, and estimating the health state of the power battery of the target equipment. In order to solve the problem that generalization of power battery health estimation in multiple heterogeneous devices is weak, a multi-level federated learning framework is designed, meanwhile, physical mechanisms and data modeling are combined, the limitation that the accuracy of battery health state estimation is insufficient is overcome, model estimation errors are remarkably reduced, and therefore the safety of an energy storage system is improved.

The invention belongs to the field of server clusters, and particularly relates to a dynamic resource scheduling-based server cluster energy efficiency optimization method, which comprises the following steps of: obtaining a resource load capacity of a server in a historical time period and a total resource load capacity of a server cluster, and training a load prediction model based on data; predicting the total resource load amount in the future time period through the load prediction model, and generating a load curve of the corresponding time; dynamically making a plurality of resource scheduling decisions according to the load curve, wherein the decisions comprise task migration, dormancy awakening and dynamic frequency modulation among the servers; and simulating and executing an energy consumption curve of each resource scheduling decision, wherein the decision with a low fluctuation index of the energy consumption curve is a target resource scheduling decision. The total resource load amount of the server cluster is predicted, a plurality of energy consumption curves conforming to the predicted load curve are generated and compared, the energy consumption curve with a low fluctuation index is selected as a target resource scheduling decision, and the influence of fluctuation on server hardware is solved while energy efficiency optimization is performed.

The present application provides a software development resource scheduling method and system based on big data. Among them, the method includes: mapping the heterogeneous configuration feature graph to the hyperbolic embedding space, combining the curvature adaptive algorithm, measuring the environmental configuration similarity value between the client server and the existing business servers in the pre-built resource pool, and the resource pool is constructed based on historical big data; according to the resource type required for the software development task provided by the client server, screening the existing business servers with configuration similarity values ​​greater than the preset similarity threshold and matching resource types from the resource pool; sorting the called business servers in descending order according to the similarity value, and elastically scheduling the software development resources in the called business servers based on the sorting results. The present application utilizes a mechanism for dynamically evaluating the environmental configuration similarity value, and combines it with a priority scheduling strategy to achieve zero-dependency conflict, high-precision, and highly effective resource scheduling, as well as the elastic reuse capability of resources.

The invention discloses a server cluster asset management method, a server and a client, and relates to the technical field of server clusters. The method comprises the steps that heartbeat information of a client node is received, and the heartbeat information is used for representing the state of the client node; constructing a node tree based on the heartbeat information; and managing the server cluster according to the node tree, thereby solving the technical problems of increased management difficulty and non-uniform resource distribution caused by the complex network environment and the affiliation relationship between the servers.

Methods and apparatus are described for enhanced node communication within a wireless node network having nodes and a server. The method begins with a first node associating with a second node, and the first node capturing relevant node information. The first node transmits the relevant node information to the server via the second node operating as an intermediary for indirect communication with the server. A third node captures relevant node information and transmits the relevant node information to the server without using the second node as an intermediary or bridge to the server. Each of the first node, the second node, and the third node may be wireless transceiver-based nodes and respectively implemented as integrated circuits.

The invention discloses a heterogeneous cellular network task unloading method oriented to location privacy protection, and belongs to the technical field of communication. The invention provides a heterogeneous cellular network task unloading method oriented to location privacy protection, aiming at solving the problem that location privacy leakage is possibly caused when a mobile user unloads a task to a micro base station edge server to which the mobile user belongs in heterogeneous cellular network edge calculation. According to the method, the privacy leakage risk of task unloading is evaluated according to user position distribution, the number of unit tasks to be unloaded and the resource state of each edge server, the position correlation between users and target servers is weakened by using a relay forwarding mechanism of a macro base station, and a deep reinforcement learning method is adopted, so that the task unloading efficiency is improved. And the position privacy risk and the task unloading strategy of the user are dynamically learned and optimized, so that the position privacy protection of the user is effectively realized and the average unit task energy consumption is minimized under the constraint of average unit task tolerance time delay.

The invention provides a satellite communication method and device, computer equipment and a storage medium, the method is applied to a satellite communication system, the satellite communication system comprises user equipment, a first modem, a second modem and a target server, and the method comprises the following steps: responding to a communication connection request of the user equipment to the target server; creating a satellite communication link between the user equipment and the target server through the first modem and the second modem; and performing data transmission control on the satellite communication link based on a dynamic bandwidth delay product window model and a BBRv2 congestion control algorithm. Based on the scheme of the invention, the transmission performance in a high-delay and high-error-code environment is remarkably improved while the transparent proxy of a satellite link is realized.

An embodiment of the present invention provides a method, device, server, medium, and product for updating process orchestration. The method includes: determining a process node to be opened in a target process body; generating corresponding foreground access specification interface data according to the process node to be opened, where the foreground access specification interface data includes a parameter of a foreground service identity; if a data acquisition request sent by a foreground server is received, sending the foreground access specification interface data to the foreground server, where the foreground access specification interface data is used to instruct the foreground server to generate a corresponding development data packet to be updated, and the development data packet to be updated includes a foreground service identity; if it is monitored that a development data packet update condition for the process node to be opened is satisfied, updating the development data packet of the process node to be opened according to the development data packet to be updated and the corresponding foreground service identity. The decoupling between the middleware server and the foreground server is realized, and the efficiency of meeting business requirements is effectively improved.

The invention relates to the technical field of scheduling between servers, in particular to a batch task running system for scheduling between servers. The system comprises a task configuration module, a task deployment module and a task operation kernel module. According to the invention, a batch task processing strategy is formulated through the task configuration module, and the execution process of the batch tasks is managed and monitored through the task deployment module, so that the batch tasks are processed in batches in a process manner, the processing work of each batch task is systematically managed, the processing logic of the batch tasks is improved, and the processing efficiency is improved. Meanwhile, the task operation kernel module responds to the task instruction of the task configuration module, the corresponding task models are controlled to execute and process the batch tasks, the task instructions are automatically adapted through processing of the corresponding batch tasks in an adaptive mode through the multiple different types of task models, the task operation processing in different scenes is met, and the task operation efficiency is improved. The flexibility of the whole system is improved.

Methods and apparatus are described for enhanced node communication within a wireless node network having nodes and a server. The method begins with a first node associating with a second node, and the first node capturing relevant node information including shared data from another of the nodes. The first node transmits the relevant node information to the server via the second node operating as an intermediary for indirect communication with the server. A third node captures relevant node information and transmits the relevant node information to the server without using the second node as an intermediary or bridge to the server. Each of the first node, the second node, and the third node may be wireless transceiver-based nodes and respectively implemented as integrated circuits.

The invention discloses a complex application-oriented extensible server energy efficiency modeling method and device, and the method comprises the steps: carrying out the energy efficiency modeling of a baseline server, and carrying out the extensible modeling of a target server: 1), carrying out the energy efficiency modeling of the baseline server: carrying out the independent energy efficiency modeling of hardware and software, the comprehensive performance and power consumption of the hardware are obtained based on software and hardware collaborative analysis, and finally the multi-hardware weight is dynamically adjusted to obtain the system-level energy efficiency of the server; and 2) extensible modeling of the target server: updating the trained energy efficiency model to migrate the business energy efficiency assessment from the baseline server to the target server to be tested by combining the operation energy efficiency of the unknown application to be tested under the baseline server and the hardware architecture difference between the target server. By means of the hardware decoupling method, the energy efficiency of the application load on the non-running target server can be predicted, complexity of real machine energy efficiency evaluation is avoided, extensible server energy efficiency modeling is achieved, and the generalization problem of an energy efficiency model in a complex software and hardware combination environment is solved.

The present invention relates to a multifunctional drone-assisted asynchronous cluster personalized federated learning method, which belongs to the field of wireless communication technology. The method comprises the following steps: S1: establishing a drone-assisted FL network system model; S2: establishing a drone logistics transportation and assisted FL training model; S3: establishing a communication delay, model training time, and uplink delay model between the drone and the sub-server; S4: establishing a personalized federated learning mechanism, solving the personalized model in the inner layer and the global model in the outer layer, and training the personalized model through asynchronous two-layer parallel optimization and intra-group synchronous federated averaging mechanism; S5: optimizing the drone flight path using the group model staleness as the optimization variable; S6: adjusting the optimization variable according to the relationship between the model staleness and the number of user device training times, and proposing a drone path dynamic optimization algorithm based on deep reinforcement learning. The multifunctional drone path optimization algorithm provided by the present invention significantly improves the overall communication efficiency.

The present invention discloses a data exchange method and model for government hall equipment, belonging to the field of electronic government information technology. The method processes data notification based on the middleware of the MQTT protocol, uses the publish / subscribe mode of its message queue, and quickly notifies the relevant equipment of data changes; the equipment initialization data is first read from the server side, and the network between the equipment and the server is connected after the equipment software is installed. The equipment software configuration only needs to configure the address of the server, and then pull other basic configuration data from the server. The present invention can adapt to the complex network environment of the government hall, the situation where the equipment is diverse and the model is not fixed, reduce the difficulty of development and maintenance, reduce the workload of deployment and implementation, and can be applicable to the agile development requirements with rapidly changing requirements.

The application provides a CPU load balancing-based inter-server communication scheduling method and system, and relates to the technical field of CPU load balancing, wherein the method comprises: performing quality evaluation on network state data to generate strategy data, and meanwhile converting electromagnetic intensity signals and temperature distribution signals in a computer room into corresponding digital signals to form an environment data set; fusing the strategy data and the environment state data set through a dynamic weight adjustment algorithm to generate a frequency modulation scheme and a channel switching instruction; screening a set of available CPU cores according to load data; and finally distributing data packets to target CPU cores for processing and transmission according to the frequency modulation scheme and the channel switching instruction, establishing multiple concurrent connection channels, and thereby realizing load balancing and high stability transmission between servers. The application improves the communication efficiency and transmission reliability of a server cluster.

The present invention relates to the technical field of server scheduling, and discloses a large model reasoning scheduling method based on an off-grid computing server, comprising the following steps: collecting hardware information of the off-grid computing server and constructing a server resource vector; decomposing the large model reasoning task into multiple subtasks and constructing task modeling; dividing the tasks using an optimization method based on the hardware information and task dependencies of the off-grid server; dynamically scheduling the subtasks based on the current load status of the server; dynamically migrating some tasks to other servers based on the real-time computing load of the server to ensure load balance between servers; executing the assigned tasks, monitoring the operating status of the server and the completion of the tasks, and dynamically adjusting the task allocation strategy based on the feedback data. By modeling task requirements and server resources, combined with real-time monitoring and optimized scheduling, efficient computing and maximizing resource utilization of large model reasoning in an off-grid environment are achieved.

The invention provides an offshore floating unit remote control method and system, and the method comprises the steps: S1, an SCADA server generates an instruction signal, and transmits the instruction signal to a command and control server, and the command and control server intermittently transmits a corresponding instruction signal to a Beidou satellite; s2, the Beidou satellite forwards the received instruction signal to a unit, the unit starts a backup power supply or sends unit state information to the Beidou satellite according to the instruction signal, and then the unit closes the backup power supply; s3, the Beidou satellite forwards the unit state information to the command and control server, the command and control server intermittently receives the state information according to the transmission capacity and forwards the state information to the SCADA server, and then the steps S1 to S3 are repeated until the number of times of starting the backup power supply reaches a preset value; and S4, keeping the backup power supply in a starting state, and executing a preset typhoon-resisting strategy by the unit. According to the invention, the problem that the communication between the floating unit and the land control center is difficult due to the fact that the traditional mobile communication network cannot cover in the prior art is solved.

The invention discloses a DNS cache snooping attack protection method and device based on an onion network. The method comprises the following steps: receiving a DNS query request sent by a client; a DNS pseudo resource record type is packaged in the DNS query request, and the DNS pseudo resource record type comprises an onion field; extracting an onion field in the DNS pseudo resource record type, and searching a DNS record corresponding to the request domain name, the request type and the onion field in a cache of the DNS pseudo resource record; if the corresponding DNS record is found in the own cache, a refusal response result is returned to the client; and if the corresponding DNS record is not found in the cache, triggering recursive query. According to the DNS cache snooping attack protection method based on the onion network provided by the invention, cache snooping protection is provided for DNS query by utilizing the idea of onion routing, and meanwhile, the decreasing effect of ttl is considered, so that the problem of influencing the performance or the consistency between data and an authoritative server is avoided.

The embodiment of the invention provides a federal large model learning method based on confrontation feature aggregation and a multi-party computing system, and relates to the field of artificial intelligence. The method is applied to a multi-party computing system, the multi-party computing system comprises a plurality of participants and a central service party, and the participants are in communication connection with the central service party. The method comprises the following steps: for any one of a plurality of participants, the participants perform adversarial feature generation processing on local data based on a noise generator to obtain adversarial features, and the noise generator is used for generating noise data; uploading the confrontation features to the central service party; and the central service party performs model training on a specific task through an adversarial feature cluster to obtain a target model, and the adversarial feature cluster is obtained by aggregating the adversarial features issued by the plurality of participants. According to the method provided by the invention, the communication overhead between each participant and the server is reduced, and the privacy protection of local data is improved.

The application provides a book management system data real-time synchronization processing method for dynamic updating of a book inventory, relates to the technical field of book management, and comprises the following steps: real-time monitoring of book inventory data change events of each book management terminal and a background server end, if a change is monitored, a change trigger signal carrying a change end identifier and a timestamp is generated; based on the change trigger signal, incremental change data is captured to analyze a book identifier, change content and an operation type, and a plurality of key frame identifier nodes corresponding to a transmission protocol encapsulation layer of the incremental data are extracted as dynamic anchor points to form an anchor point coordinate set; the anchor point coordinate set is bound with the incremental data packet and is synchronized to all other terminals and servers, so that each receiver simultaneously obtains the incremental data to be processed and the corresponding anchor point coordinate set. The application can realize real-time and strongly consistent synchronization of book inventory data between multiple terminals and a background server, and avoids inventory data errors caused by concurrent operations.

The invention relates to the technical field of computer fault early warning, in particular to a server fault early warning system and method for a computer. Comprising a multi-source signal acquisition module, a data preprocessing module, a cross-domain signal correlation analysis module, a nonlinear dynamic modeling module, a frequency domain feature extraction module, an intermittent fault prediction model and a multi-level fault early warning module. According to the method, through technologies of multi-source signal fusion acquisition, nonlinear dynamic modeling, frequency domain feature extraction, Bayesian reasoning and the like, accurate prediction and multi-level early warning of intermittent faults of the server are realized, and the accuracy and real-time response capability of fault detection are improved, so that stable operation of the server is effectively guaranteed.

Server to server direct integration is provided. A system constructs an array of power values from time series data collected by a power sensor coupled with a television. The system inputs the array of power values into a model trained with machine learning based on log files of data collected from a plurality of power sensors coupled with a plurality of televisions. The system determines, based on output from the model generated with the array of power values, the television is active. The system executes, responsive to the determination that the television is active, an action to evaluate a performance of content rendered by the television.

The embodiment of the invention provides a method and device for sending a penetrating handshake packet, and the method comprises the steps that an equipment end receives a target connection establishment request from a platform server, and the target connection establishment request carries first smooth round-trip time between a client and first network address translation equipment and second smooth round-trip time between the client and the platform server; the device end receives a time delay acquisition response from the platform server, and the time delay acquisition response carries third smooth round-trip time between the platform server and the second network address translation device; and the device end determines a target moment for sending the penetrating handshake packet to the client according to the first smooth round-trip time, the second smooth round-trip time and the third smooth round-trip time, and sends the penetrating handshake packet to the client at the target moment. According to the method and the device, the problem of NAT penetration failure caused by time sequence difference when the connectivity detection message reaches the opposite-end outlet NAT in the related technology is solved.

This application relates to a communication method and apparatus, to implement a full interconnection between servers by connecting a computing device in a server included in a computing cluster to an optical switch. In this application, a management device obtains a quantity M of servers in the computing cluster and a quantity N of computing devices in each server; determines, based on the quantity M of the servers, a minimum quantity n of computing devices required by each server to implement a full interconnection between the M servers, where M, N, and n are positive integers; when n is less than or equal to N, determines a connection policy between the servers when the M servers are fully interconnected; and connects the M servers according to the determined connection policy to enable the M servers to be fully interconnected.

This invention relates to an inter-domain data transmission device for industrial field instruments and a DCS monitoring system. The device includes: a device management IO data processor, a device management IO data encoder, a memory, and a device management IO data decoder. The device management IO data processor acquires data from industrial field instruments transmitted by all DCS controllers within a specified range and encodes all the data into character sequences. When data from different character sequences are aggregated, the data is processed according to the information encoding hierarchy, and the data is classified and merged to obtain information items for categorized storage in the memory. The device management IO data encoder, based on the domain to which the stored information item belongs, sends the information item to the DCS monitoring server of the corresponding domain according to the output data stream format. This invention solves the problems of slow data transmission speed and cumbersome configuration of structured data of field instrument devices between different servers.

The present disclosure provides an SCADA system that can equalize the time to which alarm information is given between two redundant SCADA servers. A SCADA system is provided with a first SCADA server and a second SCADA server configured so as to be able to receive a signal relating to an alarm from a PLC. When each of the first SCADA server and the second SCADA server receives a signal related to an alarm from the PLC, the first SCADA server and the second SCADA server generates alarm data in which the signal related to the alarm is endowed with the current time, and transmits the generated local alarm data to the SCADA server of the opposite side. In addition, remote alarm data is received from the counterpart SCADA server. In addition, when at least one of the local alarm data and the remote alarm data is acquired, any one of the local alarm data and the remote alarm data is selected according to a predetermined rule.

The embodiment of the present application provides a method and device for sending a penetration handshake packet, comprising: receiving a target connection request from a platform server by a device end, wherein the target connection request carries a first smooth round trip time between a client and a first network address translation device, and a second smooth round trip time between the client and the platform server; receiving a time delay acquisition response from the platform server by the device end, wherein the time delay acquisition response carries a third smooth round trip time between the platform server and a second network address translation device; determining a target time for sending the penetration handshake packet to the client according to the first smooth round trip time and the second smooth round trip time, and the third smooth round trip time by the device end, and sending the penetration handshake packet to the client at the target time. Through the present application, the problem of NAT penetration failure caused by the time sequence difference of the connectivity detection message reaching the exit NAT of the opposite end in the related art is solved.