10results about How to "Reduce forwarding delay" patented technology

The invention discloses a multi-substream parallel forwarding scheduling method for a novel intelligent computing center network. The whole communication forwarding process is divided into three stages of data stream segmentation and distribution, parallel transmission and receiving end sorting and recombination. The cooperation of a data stream segmentation distribution strategy of a sending end and an out-of-order data packet sorting and recombination operation of a receiving end is comprehensively considered, an optimal distribution ratio and a corresponding optimal forwarding path are determined, and the optimal distribution ratio and the corresponding optimal forwarding path are mapped to a physical network to complete parallel transmission of multiple substreams. According to the method, the problem of failure of a traditional network scheduling method caused by new characteristics of small flow quantity and large single flow bandwidth of distributed training service data flow of an AI large model in a novel intelligent calculation center is solved.

The embodiment of the invention provides a switch chip, a switch, a computing system and a message transmission method. The scheme is suitable for the switch chip comprising a CISC instruction set, an RISC reduced instruction set (especially an RISC-V instruction set) or a VLIM instruction set architecture. The switch chip comprises an on-chip network, a plurality of port clusters and a plurality of physical layer ports, the port cluster comprises a plurality of switch ports, each switch port is connected with one physical layer port, and the switch ports are connected with the on-chip network; the physical layer port is used for sending the received message to the switch port connected with the physical layer port; and the switch port is used for caching the received message to an inlet end data caching unit included in the switch port, and sending the cached message to the output port through the network-on-chip after receiving permission information sent by the output port through the network-on-chip. According to the scheme, the message transmission delay of the switch chip can be reduced.

The application discloses an IP bridging method and system based on star flash SLE, belongs to the technical field of short-distance wireless communication and Internet of Things, and aims at the problems of complex multi-connection management, high deployment cost of heterogeneous network intercommunication of the existing low-power wireless IP scheme.The center device is configured as an access point role, the mapping relationship between terminal MAC addresses and connection identifiers is maintained through a connection management list, a unique SLE virtual network interface is created in an IP protocol stack, bidirectional data forwarding of a single interface to multiple SLE connections is realized, a data link layer bridging channel of SLE and WiFi can be further established, and two-layer transparent forwarding is realized.The application can simplify multi-device access management logic, reduce networking complexity, and improve the efficiency and compatibility of low-power terminal IP access.

The invention discloses a message processing method and system, and relates to the technical field of network communication. The specific embodiment of the method comprises the following steps: receiving a first message through a field programmable gate array unit, and querying a flow table according to message information carried in the first message; forwarding the first message to a processor unit in response to the situation that the message information is not queried in the flow table, so that the processor unit encapsulates the first message to obtain a second message; receiving a second message sent by the processor unit, analyzing the second message to obtain packaging information, and configuring the message information and the packaging information into a flow table; and sending the second message to a message outlet. According to the embodiment of the invention, data interaction for many times when the processor is utilized to configure the flow table is avoided, and the time required for configuring the flow table is shortened, so that a new message can hit the flow table in time, the message is forwarded in time, the message forwarding time delay is reduced, and the user experience of establishing network connection is improved.

The invention relates to the technical field of communication, and discloses a load balancing method, device and equipment for a data packet, which are used for solving the problem of low reordering efficiency of a downlink line card caused by high reordering complexity of the downlink line card in a load balancing method in the related technology. In the embodiment of the invention, the plurality of data packets are divided into different logic frames according to the lengths of the data packets, and the serial numbers of the logic frames and the logic frame end marks EOP are marked for the data packets meeting the conditions, so that the data length difference between the logic frames is reduced, the clear upper bound of the length of the logic frames is ensured, and the data transmission efficiency is improved. The load balancing effect and the applicability in different scenes and hardware are improved, the downlink line card can timely determine whether forwarding of the data packet in the logic frame is ended or not through the logic frame end mark EOP, the data packet forwarding efficiency is improved, the end-to-end data packet forwarding time delay in the device is reduced, the rearrangement complexity of the downlink line card is reduced, and the user experience is improved. And the reordering efficiency is improved.

The application discloses a kind of financial special network optimization method and system for AI big model, it is related to communication network technical field.The method is by collecting AI business demand and network state, using the SID programming ability of SRv6, generates exclusive path strategy for AI training, inference and sensitive data flow;Real-time monitoring of computing power node load, link computing power state and network scheduling, realize the collaborative optimization of "network-computing power";Introduce the hop-by-hop path authentication mechanism based on dynamic authentication segment identifier Authen-SID, enhance the security of data transmission path;End-to-end visualization operation and maintenance and closed-loop feedback of service flow are realized by using stream detection technology.The application solves the problems of traffic scheduling mismatch, network and computing power collaboration deficiency, path security guarantee deficiency and weak operation and maintenance visualization capability in the prior art, realizes the deep adaptation of network resources and AI business demand, improves the computing power collaborative efficiency, path security and operation and maintenance intelligent level.

The application discloses a kind of communication and control deep fusion's intelligent distribution network edge control equipment applied to electric communication technical field, including mapping database, instruction generation module, identification processing module, photoelectric conversion module, wavelength allocation module, multi-port DWDM optical module and task verification execution module, different emergency degree's task instruction type is directly corresponded to different physical wavelength channel and optical wavelength identification, the physical layer solidification of task priority is realized, the queuing delay problem caused by the same channel shared by multiple tasks is reduced, simultaneously, by the correspondence of optical wavelength identification and task instruction or not, a judging means is provided for the safety of data, whether there is the possibility of malicious modification instruction by criminal, intent to destroy equipment operation is effectively judged, in addition, by the task midway upgrading means, when emergent situation occurs, optical wavelength identification is modified automatically, the timeliness and correctness of downstream equipment task response are further guaranteed.

The present specification provides a packet forwarding method and device, electronic equipment and medium. The method comprises: receiving a first packet; assigning a target slice channel for the first packet from a plurality of slice channels, wherein the plurality of slice channels are obtained by channel slicing of link aggregation channels of a stacking system according to a common multiple of a total number of aggregation interfaces and a total number of available processors of the stacking system, the stacking system comprising a plurality of network devices, each network device corresponding to at least one slice channel; and sending the first packet to a first network device corresponding to the target slice channel through the target slice channel.

The invention relates to the technical field of communication, and discloses a load balancing method, device and equipment, which are used for solving the problem that a load balancing method in the related technology is easy to cause message probabilistic routing polarization. In the embodiment of the invention, polling is carried out on a received message on a forwarding port according to a descrambling queue identifier and a serial number of a logic frame to which the message belongs, so that all the forwarding ports at least forward one logic frame, and a corresponding relationship between the logic frame and the forwarding port is recorded in a state storage table; the condition that one forwarding port is congested and the other forwarding port is idle can be avoided, the probability of message probabilistic routing polarization is reduced, and by recording the corresponding relation between the logic frame subjected to load balancing and the forwarding port in the state storage table, the probability of message probabilistic routing polarization is reduced on the premise that the logic frame is subjected to load balancing. The messages in the same logic frame are ensured to be forwarded to the same path, disorder of the messages in the logic frame is avoided, and high load balance of the router is met.

The invention relates to a distributed matrix type switching system, a message forwarding method and a chip, the system comprises a plurality of input ports and a plurality of output ports, each input port is connected with a switching module, the number of the input ports is consistent with that of the output ports, and the number of the output ports is equal to that of the input ports. Each output port is connected with one buffer module, each switching module is respectively connected with all the buffer modules, and each switching module is used for carrying out message scheduling forwarding on original messages received by the corresponding input port. Therefore, the message after the message scheduling and forwarding processing is output from any one output port in all the output ports. When the system is used for forwarding the message, the'header blocking problem 'is avoided, the message forwarding delay is reduced, and the forwarding efficiency of the whole chip is improved.