1232 results about "Host (network)" patented technology

The invention belongs to the technical field of artificial intelligence and high-performance computing, and particularly relates to a neural network large model efficient reasoning method based on multiple GPGPUs. The method aims to solve the problems of high communication overhead, non-uniform load, low resource utilization rate, high data transmission delay and the like among multiple processors. Dividing a calculation task into a plurality of sub-graphs through static analysis and mixed granularity partitioning of a model calculation graph; distributing the sub-graphs to the optimal GPGPU based on a weighted cost function in combination with heterogeneous resource perception and a dynamic mapping strategy; a global pipeline scheduling plan is constructed by using communication topology perception, and calculation and communication overlap are maximized; data are loaded in advance through a host side hierarchical caching and asynchronous prefetching mechanism, and transmission delay is hidden; multi-stream concurrent execution and event-based lightweight synchronization are adopted on each GPGPU, so that waiting overhead is reduced. According to the method, the reasoning delay can be remarkably reduced, the throughput and the hardware utilization rate are improved, and the method has good adaptivity and expandability.

The technology disclosed relates to a method, system, and non-transitory computer-readable media that detects malicious communication between a command and control (C2) cloud resource on a cloud application and malware on an infected host, using a network security system. The network security system reroutes the cloud traffic to the network security system. The incoming requests of the cloud traffic are directed to a cloud application in the plurality of cloud applications, and wherein the cloud application has a plurality of resources. The network security system analyzes the incoming requests, determines that the incoming requests are targeted at one or more malicious resources in the plurality of resources. Also, the network security system prevents transmission of the incoming requests to the malicious resources, by making the malicious resources unavailable for receiving future incoming requests, while keeping other resources in the plurality of resources available for receiving the future incoming requests.

A network interface module for coupling a host device to a switched network as a network node is described. The network interface module comprises a single half-duplex port for communicatively coupling to a shared bus of the switched network, at least one frame queue sized to store one multicast read frame received via the shared bus, and logic circuitry. The logic circuitry is configured to decode a read command for the interface module included in a payload of the multicast read frame that includes multiple read commands for other network nodes of the switched network, and transmit a response frame including read data on the shared bus when detecting the shared bus is available for transmitting.

Methods, apparatus, systems, architectures and interfaces for a Multi-access Edge Computing (MEC) host providing a MEC platform, including a reliable and available wireless (RAW) control entity, are provided. A method may include any of: generating MEC application information associated with a MEC-to-RAW request according to information indicating a MEC application; transmitting, to another network entity, the MEC-to-RAW request including the MEC application information; and receiving, from the other network entity, a MEC-to-RAW response including information indicating any of: (1) a wireless communication network path, and (2) RAW control information associated with the wireless communication network path.

A container-based parallel computing system for executing high-performance computing (HPC) applications. The system leverages container technology to package the applications executed at the nodes in a cluster. To load and execute a job in the parallel computing system, containers are deployed in a cluster that include all the application resources and configuration information that the particular HPC application needs to execute. An event-driven batch scheduler may be used to dynamically allocate resources for executing multi-node jobs in the container-based parallel computing system, handling the coordination of resource allocation for the customer. The scheduler insures that jobs begin executing as fast as possible, and handles failure conditions such as partial scaling. Virtual network interfaces are attached to the containers that allow the containers to connect to and communicate with other containers in the cluster directly through the network interfaces of host machines using IP addresses provided by the virtual network interfaces.

The present application relates to a data replay method, a data processing unit, a network interface card, a device, and a storage medium. The method comprises: when a storage stack is restarted, acquiring a replay-pending request queue from a persistent memory, the replay-pending request queue being a queue obtained by mapping a transmission buffer area into the persistent memory, and the transmission buffer area being used for receiving a request from a host side; on the basis of the replay-pending request queue, acquiring a replay-pending request from the persistent memory; and resending the replay-pending request to a target acceleration component directed to a storage device, so as to instruct the storage device to store target data which resides in the persistent memory and corresponds to the replay-pending request. In the method, a request queue, a request, and data corresponding to the request are stored in the persistent memory in advance, and if the storage stack is restarted, the replay-pending request is recovered from the persistent memory, eliminating the need for DMA and encryption / decryption techniques, and shortening an interruption duration of a storage data path, thereby improving data storage efficiency on the host side.

A data processing method, a network interface card, an electronic device, and a storage medium. The method includes: determining a target dispatch queue to be dispatched from a host memory in response to a doorbell signal sent by a target host, where the doorbell signal indicates that there is a target message to be sent; determining a current target dispatch state of the target dispatch queue, where the target dispatch state is obtained based on an activity and a credit value corresponding to the target work queue element; determining whether a dispatch mechanism corresponding to the target dispatch queue is valid based on the target dispatch state, where the dispatch mechanism indicates whether the target dispatch queue is allowed to be dispatched; and performing a dispatch operation on the target work queue element and the target message in response to the dispatch mechanism being valid.

Systems, methods, and apparatus may be provided for identifying a device or user behind a resential gateway. For example, a wireless transmit / receive unit (WTRU) may perform a method that may comprise a number of operations. A Protocol Data Unit (PDU) session to a network node may be established. A first message may be received from a device. The first message may indicate a Dynamic Host Configuration protocol (DHCP) request. A device identifier associated with the DHCP request may be determined based on configuration information and a characteristic of the DHCP request. A second message may be generated. The second message may indicates the DHCP request and a remote identification (ID) option value. The remote ID option value may be set to the device identifier. The second message may be sent to the first network node using the first PDU session.

The invention discloses a software and hardware cooperation system and method for satellite network routing simulation, and mainly solves the problem that an existing simulation scheme is difficult to consider high fidelity of protocol behaviors and high performance of data forwarding in a high dynamic satellite network. The system comprises a software control plane unit, a hardware data plane unit, a satellite network simulation upper computer and a switch, and the switch is bidirectionally connected with the software control plane unit, the hardware data plane unit, the satellite network simulation upper computer and the switch and serves as a data exchange platform of the system. The satellite network simulation upper computer is used for system configuration management, operation state monitoring and result visualization; the software control plane is used for executing simulation calculation and routing decision of a satellite network routing protocol and generating a standardized forwarding table item; and the hardware data plane unit performs high-speed and deterministic table lookup forwarding according to the standardized forwarding table item, and simulates a satellite resource constraint environment. According to the method, the integrity of protocol simulation details is guaranteed, meanwhile, high-efficiency forwarding performance and non-interruption updating in a high-dynamic environment are achieved, and the method can be used for routing protocol verification, network performance testing and satellite-borne algorithm evaluation of large-scale low-orbit satellite constellations.

The present application relates to the technical field of computers. Disclosed are a computing system, a model training method and apparatus, and a device, a medium and a product. The artificial intelligence computing system comprises a plurality of computing nodes. Each computing node comprises a host, a computing device and a network interface card, wherein the computing device comprises a first hardware interface, a first device coherence engine and a first computing core, and is connected to the host by means of the first hardware interface; the network interface card comprises a second hardware interface and a second device coherence engine, and is connected to the host by means of the second hardware interface; and by means of a cache flush operation, the computing device transmits gradient data with a network interface card memory of the network interface card, and the network interface card is used for synchronizing the gradient data with other computing nodes. The present application improves the efficiency of transmitting gradient data between a computing device and a network interface card, thus improving the training efficiency for an artificial intelligence model.

Distributed fiber optic sensing (DFOS) / distributed acoustic sensing (DAS) that illustrate inventive techniques—applicable to all embedded systems—that deliver high-bandwidth traffic from a DFOS system to a cloud or other processing resources in real-time, employ firmware that operates at a register-transfer level and writes data repeatedly into the embedded host memory directly—without software intervention after initial configuration. This technique advantageously enables the use of the relatively large host memory to compensate for any software jitter. Preferably configured, the firmware employs only a small buffer to compensate for a transaction and host-memory arbitration latency. A second dedicated thread sends out data from the user space buffer to a cloud, or a connected server. Remote procedures that are executed on a remote system (a computer, network of computers, or cloud), receive data transmitted from the embedded system, and perform further processing as necessary.

A memory management communication interface for a network-on-chip and a DDR controller. The memory management communication interface is connected to a host by means of the network-on-chip, and is connected to a DDR memory chip by means of the DDR controller. The memory management communication interface comprises: a request processing module, configured to receive a request packet from the host by means of the network-on-chip, and verify the integrity of the data format of the request packet; a response processing module, configured to send an ID unlocking request to the request processing module; and an output control module, configured to receive a read acknowledgment packet from the request processing module and a response packet from the response processing module, and further configured to handle a situation where the read acknowledgment packet and the response packet are received simultaneously. The memory management communication interface can ensure the security of data during the transmission of requests of multiple hosts over the network-on-chip, thereby improving the performance of a multi-chiplet Internet system.

The present disclosure relates to a 5G or 6G communication system for supporting higher data transmission rates. A method of a network entity included in a home network in a wireless communication system, according to one embodiment of the present disclosure, comprises the steps of receiving first steering of roaming (SoR)-related terminal information from a serving network to which a terminal is connected; determining, on the basis of the first SoR-related terminal information, whether to steer the terminal to a hosting network that provides a localized service; and transmitting first SoR information to the terminal through the serving network, if it is determined that the terminal will be steered to the hosting network that provides the localized service, wherein the first SoR information includes an indication that the terminal is to be switched into a localized service access mode.

Apparatuses, systems, and techniques for detecting that a host device is subject to a distributed denial of service (DDOS) attack using a machine learning (ML) detection system are described. A computing system includes a switch with port interfaces, a central processing unit (CPU) that implements a machine learning (ML) detection system, and network monitoring logic. The network monitoring logic can extract features from network data and send the extracted features to the ML detection system. The ML detection system determines whether the host device is subject to a DDOS attack using the extracted features. The ML detection system can send an alert to the host device responsive to a determination that the host device is subject to the DDOS attack.

The invention provides a packaging method and system for process communication between hosts of a heterogeneous operating system, and relates to the technical field of heterogeneous host communication.The method comprises the steps that a source host obtains a communication request, containing a target process identifier and communication data, of a source process, and the isomerism of the source host and a target host operating system is recognized; determining a corresponding encapsulation protocol containing a data conversion specification and a process identifier mapping rule; serializing communication data according to specifications, coding a target process identifier into a general process identifier according to rules, combining the target process identifier and the general process identifier into a transmission packet, and sending the transmission packet to a target host through a TCP network; and finally, analyzing the head of the transmission packet at the target host, extracting the general process identifier and the system type information, decoding to obtain a local process identifier and deserialized load area data, and delivering the data to the target process, so that standardized packaging and cross-platform data interaction of process communication between hosts of the heterogeneous operating system can be realized.

The disclosed systems and methods track and continuously monitor data about a player or a multiple players and create a non-playing character (NPC) and / or modify an existing NPC that replicates the player(s) play style. The disclosed systems implement an artificial intelligence engine that monitors how a real player responds to one or more events in a game and correlates game outcomes with real player actions, with the actions or reactions of third players, and / or with an amount or extent of engagement. The engine may be used to generate, host, or otherwise provide data representative of one or more NPCs to multiple different games, being hosted by one or more servers, concurrently.

The present application provides a system and a method for detecting an Apache HTTP Server path traversal vulnerability. The method comprises: inputting a target IP address list into a scanning module, and probing the network reachability of each IP address in the target IP address list; if there is a network-reachable IP address, probing all ports of the IP address, and recording open port numbers; identifying services running on the open ports and performing service classification; if a service is classified as an Apache service, initiating a Curl request command to an icons directory path of the Apache service; if the Curl request succeeds and an / etc / passwd file is read, parsing the / etc / passwd file and extracting users of a host under test; performing detection on the users of the host to obtain an Apache HTTP Server path traversal vulnerability; and generating a detection report for the host on the basis of the Apache HTTP Server path traversal vulnerability.

The invention relates to the technical field of chip function testing, and particularly discloses a multi-site testing method for TOF sensor chip function testing, which comprises the following steps: completing hardware self-inspection and software loading; the host receives a'ready 'signal fed back by the sorting machine / probe station through GPIB communication, and confirms that each slave and the corresponding peripheral are in a'to-be-tested' state through TCP communication or Pipe interaction; the host issues an independent test instruction to each slave; after each slave receives the instruction, the corresponding test resource board card is controlled to provide a test condition for the chip to be tested, and the chip output data is collected at the same time; after completing the single-station test, each slave uploads a test result to the host; the host captures the test data from the corresponding network disk and counts the yield; and the host sends the final result of the multi-station test to the sorting machine / probe station through the communication driver and automatically grabs the next batch of chips to be tested. According to the method, the overall test time is greatly shortened through the multi-site parallel test.

The present application relates to a data processing method and apparatus, a server, and a storage medium. The data processing method is applied to a storage server, the storage server includes a network interface card and a solid-state drive, and the solid-state drive includes a controller memory buffer. The method includes: receiving, by the network interface card, a data operation command which is sent by a host, the data operation command carrying input / output data; sending the input / output data to the solid-state drive according to the data operation command by the network interface card; and operating on the controller memory buffer according to the input / output data by the solid-state drive. By adopting this method, data transmission may bypass the CPU of the storage server, thus reducing the consumption of CPU resources.

A network-packet filtering system is provided. The system includes one or more end hosts and a network-packet filtering device communicable with the end hosts and an external network. Each end host obtains its host-performance vector by executing a first neural network model based on its system-performance information. The network-packet filtering device determines whether to drop a packet from the external network or to pass the packet by executing a second neural network model based on the host-performance vectors and the packet.

Systems and methods are described for providing effective hardware acceleration by performing a combination of string matching and range comparison. According to one embodiment, acceleration device of a host device associated with datacenter receives an input stream of information. The received information is matched with contents of a hash-based lookup table to identify one or more units, which satisfy at least one condition for any or a combination of a string match and a range comparison. The identified one or more units are correlated based on a set of conditions, which define at least one rule related to any of a network policy definition, a packet inspection rule, a database operation command or a format of the input stream. Any or a combination of exact string matching and exact range matching is then performed based on the at least one set of correlated units.

Techniques are described for communications in an L2 virtual network. In an example, the L2 virtual network includes a plurality of L2 compute instances hosted on a set of host machines and a plurality of L2 virtual network interfaces and L2 virtual switches hosted on a set of network virtualization devices. An L2 virtual network interface emulates an L2 port of the L2 virtual network. Storm control information applicable to the L2 port is sent to a network virtualization device that hosts the L2 virtual network interface.

Apparatuses, systems, and techniques for detecting distributed denial of service (DDoS) attacks are described. A system includes a plurality of switches in a monitored data center, each switch comprising network monitoring logic to sample network packets and generate flow records representing behavior of network traffic. A dataflow collector receives the flow records from the plurality of switches. A streaming pipeline coupled to the dataflow collector processes the flow records. A data store stores the flow records processed by the streaming pipeline. A trainer accesses the flow records in the data store and trains one or more machine learning (ML) models to detect DDoS attacks based on the flow records. At least one of the one or more ML models is deployable to at least one switch of the plurality of switches to determine whether a host device coupled to the at least one switch is subject to a DDoS attack.

The invention relates to the technical field of industrial automatic measurement and control communication, and particularly discloses multi-protocol fusion intelligent measurement and control host equipment, which comprises a data acquisition module, a health assessment module, a thermal simulation module and a cooperative control module. The equipment adopts a dual-core hardware architecture with a bottom plate and a top plate, integrates seven communication interfaces, realizes multi-protocol parallel processing through a hardware accelerator and a FreeRTOS system, the protocol switching time does not exceed 10ms, and the health assessment module adopts an online incremental learning recurrent neural network to calculate the health state index of the super capacitor module in real time, so that the super capacitor module can be used as the health state index of the super capacitor module. The thermal simulation module predicts the thermal runaway risk level based on an embedded finite volume method, and the cooperative control module dynamically adjusts the rotating speed of a fan and the switching frequency of a power supply by adopting a Nash equilibrium optimization algorithm, so that early performance attenuation recognition, thermal runaway risk prediction and intelligent cooperative control of the super-capacitor module in an extreme environment are realized. And the reliability and the safety of the system are obviously improved.

At a security gateway residing within a trusted computer network, a service request is obtained from an external client and is authenticated. In response, one instance of a plurality of available application instances is identified. The instances are hosted on a plurality of hosts having a plurality of firewalls, and the firewalls and the hosts reside within the network. The firewalls have been initially defaulted to block external access to all instances. A corresponding one of the hosts, within the trusted computer network, and corresponding to the identified one of the plurality of available application instances, is instructed to open a corresponding one of the firewalls to a network address corresponding to the external client. The external client is advised of a destination identifier and port corresponding to the opening of the firewall, to facilitate the external client communicating with the corresponding one of the hosts.

A method and apparatus for secured, peer-to-peer transfer of data rights over a computer network, the method being accomplished by a decentralized computing system including a distributed ledger platform and an off-chain data host platform. A data owner authorizes the minting and distribution of an on-chain authorization token using the distributed ledger. The authorization token includes primary and secondary encryption levels to ensure secure data transfer. The authorization tokens are used to track data access rights, enforce access policies, and control distribution of encryption keys transferred between the data owner, an administrator, a recipient, and / or a delegate.

The invention relates to the technical field of cross of storage controllers and artificial intelligence acceleration, and discloses a large model reasoning system and method based on combination of a flash memory controller and an NPU (Network Processing Unit), and the large model reasoning system comprises the flash memory controller, the NPU and a flash memory array, the flash memory controller integrates a host interface module, a flash memory interface module, an independent AI acceleration interface module and an AI management engine, and the AI management engine autonomously completes NPU initialization, model weight direct loading, KV Cache hierarchical management, RAG knowledge base retrieval and model switching; the flash memory array is divided into a firmware partition, an AI special partition and a user storage partition, and different data storage requirements are met. According to the method, large model reasoning with low delay and low CPU dependence can be realized, and the model loading delay is reduced from 5-30 seconds to lt; after 500 milliseconds, the CPU occupancy rate of the host is reduced from 15-25% to lt; 2%, and concurrent operation of 4-8 models is supported. According to the invention, integration of storage and calculation is realized, and edge end, data center and mobile equipment scenes are adapted.

The invention relates to a large-model heterogeneous reasoning engine method, device and equipment and a storage medium. The method comprises the following steps: acquiring an end side model generated by performing model conversion on a neural network model by a host end; constructing a calculation graph at the equipment end according to the end side model; a model reasoning task request is received, a task type is obtained, the computational graph is analyzed, differentiated task distribution is conducted on all nodes of the analyzed computational graph based on the task type, and all the nodes are bound with heterogeneous computing units executing different tasks; and according to a task allocation result, executing a model reasoning task through the heterogeneous computing unit bound with the allocated node. According to the method, the nodes are bound with the heterogeneous computing units executing different tasks, differentiated task distribution is performed on the nodes in the computing graph according to the task types, the computing power utilization rate of the heterogeneous computing units can be maximized when the end-side large model reasoning task is executed, the high throughput and low time delay requirements during reasoning task execution are ensured, and the reasoning efficiency of the end-side large model reasoning task is improved. The execution efficiency is improved.

A method and system for replicating data change operations in a virtualized environment are provided. A data change filter in a hypervisor of a virtualization host intercepts data change operations from a virtual machine. A network connection is established between the data change filter and a replication processing service running on a separate replication host. The replication processing service receives the data change operations from the data change filter via the network connection and replicates them to a backup location.

The utility model relates to a circuit for avoiding train running data sharing storage duplex, and belongs to the technical field of train running data safety. According to the utility model, the analog switch is arranged to realize the switching between the memory and the equipment master control module and the external access module, and at most only one of the equipment master control module and the external access module can access the memory at the same time, so that the console is not networked through the equipment master control module; the problem of data leakage of the console connected with the external access module when the external access module and the equipment master control access the memory at the same time is avoided, and the data security is improved.