129 results about "Virtual computing" patented technology

The present invention provides a fuzz testing system and method based on a virtual computer model. Traditional fuzz testing approaches primarily modify external inputs to detect security vulnerabilities, making it difficult to monitor internal state changes in real-time. This invention constructs a test environment using a virtual computer model and incorporates automatic assertion generation and detection functions, enabling precise analysis of both internal and external signals of the system. After executing fuzz testing, assertion data is analyzed to assess test coverage and identify untested code regions, allowing for an optimized testing strategy. This improves the detection rate of security vulnerabilities, enhances test automation and reliability, and can be applied in various fields such as networks, automotive ECUs, IoT, finance, aerospace, and semiconductors.

Systems and methods are provided for managing computing services for an application comprising a plurality of virtual computing components executing on one or more host computing devices, wherein a service virtual computing component is to perform application functionality, and wherein a system computing component is to perform system functionality including management of the application virtual computing component; determining the service virtual computing component is to execute using a first access credential to provide a first computing service to the application virtual computing component, and the service virtual computing component is to execute using a second access credential to provide a second computing service to the system computing component, wherein the first access credential is assigned a different set of computing resource access permissions than the second access credential.

The invention discloses a bypass laser scanning path planning method for additive manufacturing, and relates to the technical field of additive manufacturing, a printing area and an initial position are determined, the whole printing path is divided into a straight line section and an arc section, an area assistant is distributed for each divided area, a communication link is established, and the communication link is connected with the straight line section and the arc section; the regional assistant is provided with an independent virtual computing resource space; self-adaptive segmentation is executed through each regional assistant, communication with adjacent regional assistants is carried out by using a communication link when local control is executed, and a segmentation state and a result are synchronized; generating a virtual adhesive to carry out smooth transition and error compensation on a splicing gap; and after splicing is completed, submitting the whole path after splicing to a printing control system to execute printing. According to the method, adaptive path planning, multi-region parallel path segmentation task processing and segmented region high-precision splicing according to the local features of the printing region are realized, and efficient and high-quality additive manufacturing printing is achieved.

The invention discloses a six-degree-of-freedom multi-spacecraft control experiment platform and method based on semi-physics-virtuality, belongs to the technical field of spacecraft system and control system simulation, and constructs a multi-spacecraft system composed of semi-physics spacecrafts and pure virtual spacecrafts. The attitude of the semi-physical spacecraft is simulated by a simulator mounted on a three-dimensional attitude air floating table, is acquired by a sensor and is converted into a rotation matrix form, and a translation part of the semi-physical spacecraft is virtually calculated through a simulation integrator; the pose of the virtual spacecraft is virtually simulated by an SE (3) integrator with high-guarantee geometric characteristics; the system takes the ROS as a communication framework, and realizes state and instruction release by adopting a custom message; and ensuring that the time system of each spacecraft is consistent with the real time through a time synchronization mechanism. The platform supports cooperative operation of a semi-physical spacecraft and a pure virtual spacecraft, and has the characteristics of flexible structure, high simulation precision and good expandability.

Techniques for executing multiple large language models (LLMs) on a single physical computing device, including receiving a task to be performed by at least a first LLM and a second LLM, creating a plurality of virtual compute nodes on the single physical computing device for processing at least some of first set of sub-tasks to be performed by the first LLM and second set of sub-tasks to be performed by the second LLM, wherein each one of the plurality of virtual compute nodes includes one or more graphics processing unit (GPU) cores allocated for performing a sub-task of the first set of sub-tasks or the second set of sub-tasks, and executing at least some of the first set of sub-tasks and the second set of sub-tasks substantially in parallel across the plurality of virtual compute nodes using shared memory. Inferences between the first and second LLMs may be shared.

The invention discloses a PCIe-based CPU, NPU and GPU interconnection cluster management method, and belongs to the technical field of interconnection cluster management. According to the method, an interconnection link is established through a PCIe switch, and bandwidth is dynamically allocated by means of the link to adapt to task requirements; multi-priority data transmission is supported, and bandwidth resources are distributed in order according to priorities; a multi-element operation mode is configured in advance, and physical and virtual computing power resources are distributed according to needs; and an equipment state monitoring mechanism is established, idle equipment enters a sleep mode, and meanwhile, all transmission links are combed and an optimal path is selected. According to the method, fine scheduling of cluster resources is realized, the bandwidth and computing power resource utilization rate is improved, the adaptation capability to different tasks is enhanced, the orderliness and stability of data transmission are guaranteed, the cluster operation energy consumption is reduced, and efficient and stable operation of the interconnected cluster is facilitated.

Methods and apparatuses for provisioning a database management platform in a cloud computing environment include a server that reserves virtual computing resources in the cloud environment. The server provisions a database management platform using the reserved virtual computing resources. The database management platform includes primary and secondary database instances, a database observer instance, and a platform monitor agent. The server configures a database observer instance to monitor availability of database instances and to route traffic to other database instances. The server integrates the database management platform with an identity authentication service, monitors operational status of the database management platform using a monitoring service, and refreshes the reserved virtual computing resources in the database management platform using a rehydration service.

In some examples, a system receives, from a requester, a request to perform a first workload in a virtual computing environment, and determines a type of the requester, the determined type being one of a plurality of different requester types. The system receives metrics relating to resource usage in the virtual computing environment, and determines, using a machine learning model, an allocation of resources to the first workload based on the determined type of the requester and the metrics. The machine learning model adjusts the allocation of resources to the first workload based on further collected metrics relating to resource usage by the workload and based on a detected behavior of the requester while the first workload is performed in the virtual computing environment.

The invention discloses a heterogeneous AI computing power resource segmentation and scheduling method and system, and the method comprises the steps: recognizing the type of a heterogeneous AI computing power card accessed to a current server and hardware resource parameters of the heterogeneous AI computing power card, and carrying out the abstract modeling to generate a logic resource model; dividing a computing core and a video memory of a single heterogeneous AI computing power card into at least one independent virtual computing power instance through a resource virtualization technology on the basis of the constraint of a logic resource model; and in response to a computing task request, mounting the virtual computing power instance to a computing task in a containerization form, and dynamically scheduling the virtual computing power instance based on the demand of the computing task. Standardized logic resource units are generated through the minimum segmentation unit, and the utilization rate and scheduling flexibility of heterogeneous AI computing power resources are remarkably improved.

Methods and apparatuses are described for automatically provisioning and managing secure database resources in a cloud computing environment. A server provisions virtual computing resources in the cloud computing environment using resource templates and creates a database instance in each of the virtual computing resources. The server integrates the database instances and the virtual computing resources in the cloud computing environment with an identity authentication service. The server configures the database instances in two or more of the virtual computing resources to synchronize data. The server monitors each of the plurality of virtual computing resources using a monitoring service. The server updates the database instances by detaching a physical disk drive from the database instance, deleting the virtual computing resource that contains the database instance, provisioning a new virtual computing resource, creating a new database instance in the new virtual computing resource, and re-attaching the physical disk drive.

The invention discloses a parallel computing and distributed network access risk management and control device and a management and control method, and relates to the technical field of risk management and control. The system comprises three layers, namely network element equipment, a network controller and a security management and control platform, the security management and control platform integrates a PKI-based password infrastructure, a tool engine and an AI engine and uniformly distributes security management and control related strategies, and the network controller is used as an edge execution component and is connected with various network element equipment of a managed network, so that the security management and control related strategies can be uniformly distributed. The risk management and control device is constructed on a full-switching parallel computing architecture which provides a virtual computing power pool. Under the support of high-performance computing power provided by full-switching parallel computing, a safety management and control platform based on a distributed safety management and control system is constructed, unified state monitoring, risk analysis and strategy issuing are carried out on network access risks, and the safety of network access is improved through distributed network event and performance management and control. And network access risk real-time management and control are carried out.

Systems and methods are provided for managing computing services for an application, the application comprising a plurality of virtual computing components executing on one or more host computing devices, wherein service virtual computing components are for performing application functions, and wherein system computing components are for performing system functions, the system functions comprising management of the application virtual computing components; determining that a service virtual computing component is for performing using a first access credential to provide a first computing service to an application virtual computing component, and that the service virtual computing component is for performing using a second access credential to provide a second computing service to a system computing component, wherein the first access credential is assigned a different set of computing resource access permissions than the second access credential.

A control device (10) is provided with a scaling determination unit (142), an image selection unit (143), and a reconfiguration unit (144). When a time constraint corresponding to each of at least one of one or more manufacturing facilities (40) has been changed, the scaling determination unit (142) determines, on the basis of the time constraint corresponding to each of the one or more manufacturing facilities (40) after the change, whether to execute scaling processing for changing a calculation resource to be allocated to a virtual calculation environment. When it is determined that the scaling processing is to be executed, the image selection unit (143) selects, as the selection image, an image of the virtual calculation environment that satisfies the time constraint corresponding to each of the one or more manufacturing facilities (40) after the change, said selection being made on the basis of the amount of the calculation resources allocated to the virtual calculation environment. The reconfiguration unit (144) reconfigures the virtual calculation environment on the basis of the selection image.

The invention provides a kidney needle biopsy quantitative teaching evaluation system and method based on mixed reality, and the system comprises a virtual-real space registration unit which is used for building the space mapping between a virtual kidney model containing a renal cortex layered anatomical structure and a simulation entity training model body, and completing the space registration of the virtual model and the entity model body; the operation data acquisition module is used for acquiring three-dimensional motion trail data of the puncture needle in the puncture operation process; according to the quantitative scoring module, the multi-dimensional weighted scoring model takes the cortical ratio as a core evaluation index, and the scoring dimension corresponding to the cortical ratio has the highest weight in the model; the quantitative scoring module calculates a cortical ratio based on the acquired three-dimensional motion trail data, and outputs an operation quantitative score through a multi-dimensional weighted scoring model; and the closed-loop verification unit is used for acquiring a real puncture result on the simulation entity training die body and comparing the real puncture result with the puncture track subjected to virtual calculation so as to calibrate the operation quantitative score.

A seat test system is disclosed, the seat test system comprising: at least one of an interface and a control module configured to receive an input; a dummy positioning module configured to: i) position an anthropomorphic test device (ATD) in a seat system to be tested in a virtual computer-aided engineering environment, ii) access a target parameter comprising the input, and iii) determine a parameter indicative of the positioning of the ATD in the seat system based on the target parameter; and a simulation evaluation module configured to i) operate the ATD and the seating system by simulation of a collision in a virtual computer-aided engineering environment based on the parameters, and ii) generate a rating of the seating system based on a result of the simulation.

Systems, computer program products, and methods are described herein for generating an interactive 3D environment using spatial computing to generate historical digital components. The present invention is configured to identify a user account(s), wherein the user account(s) is associated with a resource account(s); identify a resource transmission(s) associated with the resource account(s), wherein the resource transmission(s) is associated with a geolocation coordinate(s) or an entity identifier(s); generate a virtual computing environment based on the at least one resource transmission and geolocation coordinate(s) or the entity identifier(s); generate a virtual computing environment avatar based on the user account, the resource transmission(s) and the geolocation coordinate(s) or the entity identifier(s); and generate a resource transmission movie in the virtual computing environment, wherein the resource transmission movie comprises at least one virtual computing environment component based on the resource transmission(s).

A first satellite includes a first host-side bridge and a first host, connected with a network via the first host-side bridge, mounted thereon. A second satellite includes a first I / O-side bridge and a first I / O device, connected with the network via the first I / O-side bridge, mounted thereon. A management device selects the first host mounted on at least one of the first satellites and the first I / O device mounted on at least one of the second satellites, according to a configuration request, and forms an information processing device in which the first host-side bridge connected with the selected first host and the first I / O-side bridge connected with the selected first I / O device constitute a virtual computer internal bus.

The invention relates to the technical field of vehicle intelligent management, and discloses a vehicle thermal management method and device, and the method comprises the steps: obtaining first temperature data of a range extender unit and second temperature data of a heated unit; wherein the first temperature data refers to the temperature of cooling water at an outlet of a heat dissipation loop when the range extender unit works; the second temperature data is the temperature of a virtual convergence point in the heat exchange loop of the heated unit; the virtual convergence point is any point between the heat exchange outlet of the heated unit and the inlet of the flow control unit; and the target opening degree of the flow control unit is determined based on the first temperature data and the second temperature data, and the range extender unit is used for conducting heat management on the heated unit. By obtaining the direct measurement temperature of the range extender unit and the virtual calculation temperature of the heated unit, introduction of waste heat of the range extender is accurately controlled. And through closed-loop control logic, the control precision and responsiveness of the system are improved, so that the energy efficiency level of thermal management of the whole vehicle is improved.

The invention relates to the technical field of computers, in particular to a heterogeneous computing power perception scheduling system and method for large model services, and the method comprises the following steps: receiving an application of a working load Pod for a virtual computing power unit; sensing fragmented heterogeneous GPU resources in the cluster and a physical topological structure thereof; dynamically generating an optimal scheduling scheme for integrating a plurality of fragment GPUs into a virtual aggregation GPU based on communication cost evaluation; scheduling the Pod according to the scheduling scheme, and starting a communication optimization agent when needed to optimize cross-node communication; the method has the advantages that original idle leftover material resources are converted into effective computing power capable of being used by a large model through the virtual aggregation technology of converting zero into integer, the average GPU utilization rate of a cluster can be increased by dozens of percentage points, and idle waste of the computing power resources is greatly reduced.

A virtual computer management system is configured including an active system management server and a backup system management server each configured to receive an operation request for operating a virtual computer. The virtual computer management system includes: a request generation unit that, when the result of a first operation request to the active system management server is a normal termination, converts the first operation request into a second operation request to the backup system management server; and a queue management unit that enqueues the second operation request in a queue and, when the result of a third operation request to the backup management server, located at the head of the queue, is a normal termination, deletes the third operation request from the queue and transmits an operation request newly located at the head of the queue to the backup system management server.

The invention relates to the technical field of material life cycle management, and discloses a material life cycle management-oriented adaptive privacy protection federated learning method, which comprises a privacy protection federated learning system, and is characterized in that the privacy protection federated learning system comprises a data management module, a virtual computing layer, a privacy collaboration layer and an application service layer; by introducing a dynamic privacy budget allocation mechanism and based on an arithmetic progression dynamic allocation strategy, it is ensured that privacy budget allocation in each round of communication is scientific and reasonable, the privacy disclosure risk is effectively prevented, privacy budgets are allocated according to an arithmetic progression formula in each round of communication, and the corresponding Gaussian noise standard deviation is calculated, so that the privacy disclosure risk is effectively prevented. The method is used for noise injection in model training, maximizes the privacy protection effect while ensuring the model performance, then introduces a dynamic regularization mechanism, and adjusts the model training process according to the loss change trend of the last round of training by defining a total loss function including mean square error loss and a dynamic regularization item and adaptively updating a regularization coefficient.

To provide a security system, a security method, and a program capable of leading an attack to an environment where an attack by an attacker can be observed while preventing the attack by the attacker.SOLUTION: In a security system including an information processing device, a verification device, a communication control device, an authentication device, and a restoration support device, the information processing device includes: a device verification unit that verifies device authenticity, which is the authenticity of hardware, by using verification information generated from startup information of the information processing device; and a control instruction unit that, if the device authenticity is not verified, instructs the communication control device that controls communication from a virtual computer to set a communication partner of the virtual computer to a decoy network that imitates a connection destination of the virtual computer.SELECTED DRAWING: Figure 4

Systems and methods for facilitating communication between multiple private networks with conflicting IP addresses are provided. The system comprises a virtual computing device configured to allocate blocks of shadow IP addresses to first and second private networks. A first bridge device is connected to the first private network and configured to receive a packet from a first host in the first private network. The packet comprises a destination address corresponding to a second host in the second private network. The first bridge device is configured to route the packet to the virtual computing device if the destination address is in a block of shadow IP addresses. A second bridge device is connected to the second private network and configured to receive the packet from the virtual computing device and translate the destination address to a corresponding native IP address.

This disclosure relates to the field of artificial intelligence chip technology, specifically to a method, architecture, and chip for dynamic collaborative resource scheduling of heterogeneous multi-core processors. The method includes: processing the computation graph using a trained graph convolutional network to obtain corresponding task feature vectors; acquiring resource status information of a heterogeneous resource pool; establishing a weighted multi-objective optimization function with maximizing performance indicators and minimizing power consumption indicators as optimization objectives; determining an iterative greedy strategy for approximating the optimization objectives; in each iteration, selecting the optimal candidate resource from the heterogeneous resource pool to add to a virtual computing cluster; and calculating whether the computing resources, storage capacity, and transmission bandwidth provided by the updated virtual computing cluster meet the corresponding requirements, until the target virtual computing cluster is obtained. This disclosure achieves precise matching between task characteristics and hardware resources, dynamically generates optimal resource configuration schemes, and improves system-level energy efficiency while ensuring smooth task execution.

An apparatus comprises a processing device configured to receive, from one of two or more storage nodes of a software-defined storage system comprising a write cache and persistent storage, a request to read a given portion of data. One or more portions of data in the persistent storage are also stored as read-only data in local storage of virtual computing instances implementing the storage nodes. The processing device is configured to read the given portion of data from the local storage of the virtual computing instances responsive to determining that the given portion of data is part of the read-only data in the local storage of the virtual computing instances, and to read the given portion of data from the write cache or persistent storage responsive to determining that the given portion of data is not part of the read-only data in the local storage of the virtual computing instances.

The application is specifically an AI computing power efficiency increasing method and system based on heterogeneous resource pooling and dynamic scheduling, relates to the technical field of distributed AI computing power scheduling, and comprises the following steps: collecting bandwidth, time delay and packet loss rate data of a wide area network link between each regional node in real time, calculating a transmission cost value between any two nodes, and generating a cost matrix.In the application, a standard virtual computing power unit index is introduced, floating point operation capability, video memory bandwidth and operator compatibility coefficient are uniformly included in calculation, different architecture heterogeneous chips obtain unified dimension logical resource representation, thereby providing a quantitative basis for cross-brand task allocation, and the problem that the existing scheme cannot perform equivalent comparison on heterogeneous chips is solved.

A seat testing system is disclosed and includes: at least one of an interface and a control module configured to receive inputs; a dummy positioning module configured i) to position an anthropomorphic test device (ATD) in a seat system to be tested in a virtual computer aided engineering environment, ii) to access target parameters including the inputs, and iii) based on the target parameters, to determine parameters indicative of the positioning of the ATD in the seat system; and a simulation evaluation module configured i) based on the parameters, to run the ATD and the seat system through a simulation of a collision in the virtual computer aided engineering environment, and ii) based on results of the simulation, to generate a rating of the seat system.

An apparatus comprises at least one processing device configured to receive a request, from an application running on a virtual computing instance, to consume data from an external device coupled to a physical hardware port of a host device, the external device having a first interface with a physical driver running on a host operating system of the host device, the first interface having a first interface type. The processing device is also configured to instantiate a second interface on the host operating system of the host device, the second interface having a second interface type different than the first interface type. The processing device is further configured to establish a service mesh for the application by linking the instantiated second interface with (i) a virtual driver running on the virtual computing instance and (ii) the physical driver running on the host operating system of the host device.

The embodiment of the invention discloses an interface interaction control method, a computer system, equipment and a computer storage medium. According to an interface type of a target virtual interface indicated by the creation request, creating a target virtual computing power device corresponding to the interface type in a target virtual environment, responding to a deep learning request of the target virtual environment through the target virtual computing power device, and forwarding the request to a host machine, and scheduling a physical computing power device corresponding to the interface type to respond to the request, and returning a processing result of the request to the target virtual environment through a target virtual computing power device. According to the method, virtualization is carried out from a software interaction level instead of virtualization of computing power hardware, the scheme is completely realized based on software, and a physical GPU is not required to have any special virtualization hardware function, so that almost all mainstream and even old models of GPUs on the market and accelerators such as NPUs with different architectures can be managed. And the user does not need to pay for expensive virtualization dedicated hardware and software license, so that the virtualization deployment cost is reduced.

The present invention discloses a cloud honeypot for the whole ship computing environment, an attack event perception and behavior analysis method, and belongs to the field of data security. The cloud honeypot includes: a link gateway, a TCP / UDP protocol parsing and authentication engine, a data analysis engine, a honeypot simulation computing service node, a honeypot simulation storage node and a proxy gateway, wherein the honeypot simulation computing service node and the honeypot simulation storage node jointly construct a virtual computing storage environment. The present invention constructs a honeypot that has the advantages of both a real system honeypot and a pseudo-system honeypot in the whole ship computing environment, thereby inducing the attacker to attack them, thereby capturing and analyzing the attack behavior, understanding the tools and methods used by the attacker, inferring the attack intentions and motives, and enabling the defender to clearly understand the security threats they face. It solves the problem of capturing and analyzing unknown new attack methods when it is impossible to obtain Internet cloud empowerment in real time, thereby enhancing the security protection capabilities of the actual system.