WO2018036173A1 - Procédé, dispositif et système d'équilibrage de charge de réseau - Google Patents

Procédé, dispositif et système d'équilibrage de charge de réseau Download PDF

Info

Publication number
WO2018036173A1
WO2018036173A1 PCT/CN2017/080586 CN2017080586W WO2018036173A1 WO 2018036173 A1 WO2018036173 A1 WO 2018036173A1 CN 2017080586 W CN2017080586 W CN 2017080586W WO 2018036173 A1 WO2018036173 A1 WO 2018036173A1
Authority
WO
WIPO (PCT)
Prior art keywords
network
stream
sub
load balancing
terminal device
Prior art date
Application number
PCT/CN2017/080586
Other languages
English (en)
Chinese (zh)
Inventor
龚志波
高静
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2018036173A1 publication Critical patent/WO2018036173A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
    • H04L47/2441—Traffic characterised by specific attributes, e.g. priority or QoS relying on flow classification, e.g. using integrated services [IntServ]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
    • H04L47/2483—Traffic characterised by specific attributes, e.g. priority or QoS involving identification of individual flows

Definitions

  • the present invention relates to the field of network communication technologies, and in particular, to a network load balancing method, device, and system.
  • ECMP Equal Cost Multi-Path
  • a concept of flowcell is proposed. The meaning is that when the data flow is sent by the sender of the Transmission Control Protocol (TCP), the data stream is divided into multiple points according to a certain amount of data. Segment, each segment has a unique number, recorded as the flow segment ID Flowcell ID.
  • TCP Transmission Control Protocol
  • the advantage of the stream segmentation method is that there is a clear upper limit on the amount of data for each segment, so load balancing based on stream segmentation can achieve good balance. For example, if the stream segment is divided in units of 64 Kbytes, the load balancing accuracy error of a single data stream does not exceed 64 Kbytes.
  • the probability of generating significant cumulative errors eg, reaching 1% of the gigabit link bandwidth, ie 10 Mbps
  • ECMP is the most common method of load balancing in units of flow.
  • the principle of this method is to use a quintuple (source IP address, destination IP address, source port number, destination port number, protocol number) for hash calculation, mapping each stream to one of multiple available links.
  • quintuple source IP address, destination IP address, source port number, destination port number, protocol number
  • flow-based load balancing has the problem of uneven load sharing. Some paths in the network may not be effectively utilized.
  • the invention is based on the concept of Flowcell, and proposes a network load balancing method, device and system.
  • a network load balancing method includes: the terminal device provides P logical channels, the P is a network maximum load balancing path number P, and P is an integer greater than or equal to 2;
  • the transmitted data stream is divided into units of stream segments to generate a plurality of sub-data streams;
  • the terminal device maps the plurality of sub-data streams to the P logical channels and sends the data to the network device.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • stream segmentation is that when a data stream is sent by a transmission control protocol (TCP) sender, the data stream is divided into a plurality of segments according to a certain amount of data, and each segment has The unique number is recorded as the flow segment ID Flowcell ID.
  • TCP transmission control protocol
  • the method further includes when the length of the Xth sub data stream and the next data stream is accumulated exceeding a maximum value of the stream segment, and the next data stream is taken as the X+1th
  • X is a non-negative integer.
  • the sub-data stream further includes: a stream segment identifier; and, according to the stream segment identifier, retrieve the P, and acquire a logical channel corresponding to the stream segment identifier;
  • the sub-data streams identified by the segments are respectively mapped to the corresponding logical channels.
  • the sub-data stream is mapped to the plurality of logical channels according to the flow segment identifier, and the load balancing is more balanced in the network.
  • the maximum load balancing path number P is determined by reading a memory of the terminal device.
  • the maximum load balancing path number P of the terminal device is set through a console or a command line.
  • the embodiment of the present invention further provides a network load balancing method, including: receiving a plurality of sub-data streams from an upstream node, and fetching the P according to the quintuple of the sub-data stream and the stream segment identifier
  • the hash operation is performed, and the mapping is performed to multiple physical links and sent to the next hop node, where P is the maximum load balancing path number of the network, and P is an integer greater than or equal to 2.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • the P value is obtained by reading the memory.
  • the P value is obtained through a command line or a network management device.
  • the embodiment of the present invention further provides a terminal device, including a processor, configured to provide P logical channels, where P is a network maximum load balancing path number P, and P is an integer greater than or equal to 2;
  • P is a network maximum load balancing path number P, and P is an integer greater than or equal to 2;
  • the data stream is divided into units of stream segments to generate a plurality of sub-data streams; and the transceiver maps the plurality of sub-data streams to the P logical channels for transmission to the network device.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • the sub-data stream further includes: a stream segment identifier; the processor is further configured to: take a P share according to the stream segment identifier, and obtain a corresponding to the stream segment identifier a logical channel; respectively mapping the sub-data streams identified by the stream segment to the corresponding logical channel.
  • the sub-data stream is mapped to the plurality of logical channels according to the flow segment identifier, and the load balancing is more balanced in the network.
  • the processor is configured to read a storage unit of the terminal device to determine the maximum load balancing path number P.
  • the transceiver is further configured to send a first message to the network device, to request the network device to feed back a P value.
  • the embodiment of the present invention further provides a network device, including a processor, that receives multiple sub-data streams from an upstream node, where each sub-data stream carries a stream segment identifier, and a transceiver is configured to use the sub- The quintuple of the data stream and the stream segment identifier hash the P remainder and map to multiple physical links Send to the next hop node, where P is the number of network maximum load balancing paths, and P is an integer greater than or equal to 2.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • the P value is obtained by reading the memory.
  • the P value is obtained through a command line or a network management device.
  • a fifth aspect a communication system comprising the terminal device according to the third aspect and the network device according to the fourth aspect.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • FIG. 1A is a schematic structural diagram of a communication network according to an embodiment of the present invention.
  • FIG. 1B is a schematic structural diagram of another communication network according to an embodiment of the present invention.
  • 2A is a schematic diagram of data of different granularities according to an embodiment of the present invention.
  • FIG. 2B is a schematic structural diagram of a novel TCP packet according to an embodiment of the present invention.
  • FIG. 3 is a schematic flowchart of a method for network load balancing according to an embodiment of the present invention
  • FIG. 4 is a schematic diagram of mapping a sub-data stream to a logical channel according to an embodiment of the present invention
  • FIG. 5 is a schematic flowchart of a method for network load balancing according to an embodiment of the present disclosure
  • FIG. 6 is a schematic diagram of mapping multiple sub-data streams to multiple physical channels according to an embodiment of the present invention.
  • FIG. 7 is a schematic diagram of mapping multiple sub-data streams to multiple physical channels according to an embodiment of the present invention.
  • FIG. 8 is a schematic structural diagram of a network device according to an embodiment of the present disclosure.
  • FIG. 9 is a schematic structural diagram of a network device according to an embodiment of the present invention.
  • FIG. 10 is a schematic structural diagram of a network device according to an embodiment of the present invention.
  • the terminal device of the present invention can communicate with one or more network devices via a network, and the user device can refer to a user terminal, a computer, a server, a telephone, a computer, a handheld device, a printer, a palmtop computer, a fax machine, Multi-function devices, projectors, plotters and other equipment.
  • the network device of the present invention may be a network side device for communicating with the terminal device, for example, may be a router, a switch, a bridge, a hub, a modem, or other device having a function of forwarding a message.
  • FIG. 1A is a schematic structural diagram of a communication network 100 according to an embodiment of the present invention.
  • the terminal device 110 is connected to the terminal device 140 through the network devices 120, 130.
  • FIG. 1B is a schematic structural diagram of another communication network 100 according to an embodiment of the present invention.
  • the terminal device 110 is connected to the terminal device 140 through the network devices 120, 130, 132.
  • FIG. 1A and FIG. 1B are merely examples, and the actual networking architecture may be much more complicated. There may be multiple network nodes from the network device 120 to the terminal device 140, and FIG. 1A and FIG. 130 and/or 132 are examples.
  • the source node device 110 wants to send a data stream to the destination node 140, and the network device 120 and 130, 132, wherein there are three physical paths in the network device 120 to the destination node 140, and an embodiment of the present invention provides a network load balancing method.
  • the data stream can be equalized and transmitted on the above three physical paths.
  • FIG. 2A shows a schematic diagram of data of different granularities.
  • Figure 2A shows an example of the flow segmentation from the TCP data stream to the stream segmentation and then to the packet.
  • the TCP data stream is divided into multiple stream segments according to a fixed upper length limit. If the length of a TCP data stream and the previous data stream exceeds the upper limit, the data stream is divided into the next stream segment.
  • the stream segmentation identifier is assigned to the stream segment while segmenting the stream segment.
  • the stream segment is further divided into packets, and the stream segment identifier is encapsulated into a packet header of each packet. In this way, the plurality of TCP data streams that are divided into the same stream segment carry the same stream segment identifier in the packets transmitted on the network link.
  • a new type of TCP packet including:
  • Source port and destination port each occupying 2 bytes, which is the service interface between the transport layer and the application layer.
  • Serial number 4 bytes. Each byte in the data stream transmitted by the TCP connection is numbered. The value of the sequence number field in the header refers to the sequence number of the first byte of the data sent in this segment.
  • Confirmation number 4 bytes, which is the sequence number of the first byte of data expected to receive the next segment of the other party.
  • Data Offset 4 bits, which indicates how far the data at the beginning of the segment is from the beginning of the TCP segment. In fact, it is the length of the header of the TCP segment.
  • the Flowcell ID in the embodiment of the present invention may be extended in the reserved field, or the flow segment identifier may be carried by defining a field named Flowcell ID.
  • Acknowledge bit ACK When the ACK value is 1, the acknowledgment number field is valid. When the ACK value is 0, the acknowledgment number field is invalid.
  • Push Bit PUSH When a receiver receives a segment with a PUSH value of 1, it will deliver it to the receiving application process as soon as possible, and will not wait until the entire receive buffer is full before delivering it.
  • Reset bit RST When the RST value is 1, it indicates that there is a serious error in the TCP connection and the connection must be released.
  • the reset bit is also used to reject an illegal segment or refuse to open a connection.
  • Synchronization bit SYN used to synchronize the sequence number when the connection is established.
  • the value of SYN is 1, the value of ACK is When 0, it indicates that this is a connection request segment. If the other party agrees to establish a connection, the SYN shall be 1 and the ACK shall be 1 in the response segment. Therefore, a value of 1 for SYN indicates that this is a connection request or a connection reception message.
  • Termination bit FIN When the FIN value is 1, it indicates that the data of the sender of this segment has been sent and requires the release of the transport connection.
  • Window 2 bytes, used to control the amount of data sent by the other party.
  • the unit is byte, indicating the upper limit of the sending window of the other party.
  • Checksum 2 bytes. The range of the check includes the header and the data. When calculating the checksum, you need to add a 12-byte pseudo header to the segment.
  • Urgent pointer 2 bytes, indicating the sequence number of the last byte in the urgent data in this segment. Only valid when the emergency bit URG has a value of 1.
  • TCP only specifies one option, the Maximum Segment Size (MSS).
  • MSS Maximum Segment Size
  • inventive embodiments are applicable to data communication networks having multiple physical links, such as, but not limited to, wide area networks, enterprise networks, data center networks, and the like. Please refer to the following examples for specific solutions:
  • the first embodiment of the present invention provides a method 300 for network load balancing, including:
  • Step 310 The terminal device provides P logical channels, where P is the maximum load balancing path number P of the network, and P is an integer greater than or equal to 2;
  • Step 320 The terminal device divides the data stream to be sent into units of stream segments, and generates multiple sub-data streams.
  • Step 330 The terminal device maps the multiple sub-data streams to the P logical channels and sends the data to the network device.
  • the network maximum load balancing path number P refers to the maximum number of paths from one node to another in the entire network.
  • the terminal device obtains the maximum load balancing path number P value of the network, and has multiple implementation modes:
  • the first implementation manner is: setting the network maximum load balancing path number P as a part of the network protocol, and the terminal device and the network device in the network all follow the P value specified by the network protocol, that is, the device has already been P before leaving the factory.
  • the value is written to the device's memory and is not changed subsequently.
  • the protocol specifies that P is 6, and the P value of the terminal device and the network device defaults to 6.
  • step 310 can be understood as determining the P value by reading the memory of the terminal device or the network device.
  • the second implementation manner is: the network administrator sets the P value of the network device through the console or the command line, and then the terminal device acquires the P value from the network device by using signaling or a message.
  • the specific implementation steps are:
  • Step A The terminal device sends a first message to the network device, where the first message is used to request the network device to feed back the maximum load balancing path number P of the network;
  • Step B The network device sends a second message to the terminal device, where the second message carries the P value.
  • the P value of the network device is set by the network administrator through a console or a command line. This P value can be changed later.
  • the format of the first message and the second message may refer to the prior art TCP.
  • the frame format of the protocol is not described here.
  • the P value can be carried by extending the reserved field of the existing TCP protocol.
  • the method also includes each of the sub-data streams carrying a stream segment identifier.
  • the meaning of the flow segment identifier is that when the data flow is sent by the sender of the Transmission Control Protocol (TCP), the data stream is divided into multiple segments according to a certain amount of data, and each segment has a unique number. , recorded as the stream segment ID Flowcell ID.
  • the logical channel in step 320 can represent a sub-connection, each logical channel is similar to a TCP connection, with independent transmission windows and control parameters.
  • the terminal device sends the packets of each logical channel in parallel, and each logical channel uses a TCP transmission window to control the transmission rate.
  • the terminal device can perform accurate path congestion state measurement for each logical channel and perform corresponding flow control. for example:
  • Step a logical channel #1 sends the message that the current sending window allows to send
  • Step b waiting for the receiving end (network device) to feed back the acknowledgement message of the logical channel #1;
  • the receiving end finds that the data sequence number of the packet received in the same flow segment is not continuous, it indicates that packet loss occurs.
  • Step c The receiving end feeds back the acknowledgement packet, and requests to retransmit the lost packet.
  • Step d When the transmitting end receives the request to retransmit the lost packet, in addition to retransmitting the lost packet, the logical channel is considered to be congested;
  • Step e The sending end halved the size of the sending window, that is, reduces the rate at which the message is sent.
  • each logical channel Since each logical channel has its own independent window, the logical channel #1 is congested. It only needs to reduce the packet sending rate of the logical channel #1. The other logical channels are not affected, and the packets are sent normally.
  • step 330 “the segmentation of each segment is mapped to the P logical channels”, which specifically includes:
  • the stream segment identification value is used to allocate P to obtain which logical channel each stream segment corresponds to.
  • P takes a value of 3, and the data stream to be transmitted is divided into 5 Flowcells in units of 64 Kbytes (ie, the size of each stream segment is 64 kbytes). Recorded as Flowcell ID#0 ⁇ Flowcell ID#4, respectively.
  • Flowcell ID#0 corresponds to logical channel
  • Flowcell ID#1 corresponds to logical channel 1
  • Flowcell ID#4 corresponds to logical channel 1.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • a second embodiment of the present invention provides a method 500 for network load balancing, including:
  • Step 510 The network device receives multiple sub-data streams from the upstream node, where each sub-data stream carries a stream segment identifier.
  • Step 520 Perform a hash operation on the P remainder according to the quintuple of the sub-data stream and the stream segment identifier, and map to multiple physical links to send to the next hop node, where P is the maximum load of the network.
  • P is the maximum load of the network.
  • the number of equalized paths, P is an integer greater than or equal to 2.
  • the next hop node may be an intermediate node (network device) or a destination node, and is a terminal device that the plurality of sub-data streams finally arrive at.
  • the network device identifies that the packet belonging to one logical channel passes the six-tuple information, and the so-called six-tuple includes the traditional five-tuple and the stream segment identifier, and the six-member group includes: the source IP, Source port, destination IP, destination port, protocol number, and stream segment identifier %P, where % represents the remainder operation.
  • the same message of the six-tuple comes from a logical channel.
  • the network device When receiving a packet, the network device firstly searches the routing table according to the destination IP address of the packet, and the obtained route includes M next hops, that is, M links of adjacent downstream physical links. M is greater than or equal to 1.
  • the value of the logical channel P is 3, that is, the packets received on the three logical channels are sent through three physical links, where the packets on the first logical channel are sent through the first physical link, and the second logical channel is sent.
  • the packet is sent through the second physical link; the packet on the third logical channel is sent through the third physical link. That is, in this implementation manner, the number of the physical link corresponds to the number of the logical channel.
  • the logical channel P takes a value of 3
  • the physical path physical link takes a value of 2, that is, the received packet on the three logical channels passes through two physical entities.
  • the path physical link is sent, where the packet on the first logical path is sent through the physical link of the first physical path; the packet on the second logical channel and the third logical channel is sent through the physical link of the second physical path. If the next hop node of the physical link of the second physical path has two physical path physical links, the packet sent by the physical link of the second physical path can be divided by the secondary.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • another embodiment of the present invention provides a terminal device 800, where the terminal device 800 includes a processor 810 and a transceiver 820, where:
  • the processor 810 is configured to provide P logical channels, where P is the maximum load balancing path number of the network, P is an integer greater than or equal to 2, and the data stream to be sent is segmented by using a stream segment as a unit to generate multiple sub-data. flow;
  • the transceiver 820 is configured to map the multiple sub-data streams to the P logical channels and send the data to the network device.
  • Each sub-data stream carries a stream segment identifier.
  • the number of network maximum load balancing paths P is obtained, and there are multiple implementation manners:
  • the first implementation manner is: setting the network maximum load balancing path number P as a part of the network protocol, and the terminal device and the network device in the network all follow the P value specified by the network protocol, that is, the device has already been P before leaving the factory. The value is written to the device's memory and is not changed subsequently.
  • the processing unit 810 reads the terminal setting The standby storage unit determines the maximum load balancing path number P.
  • the second implementation manner is: the network administrator sets the P value of the network device through the console or the command line, and then the terminal device acquires the P value from the network device by using signaling or a message.
  • the specific implementation is:
  • the transceiver 820 is configured to send a first message to the network device, to request the network device to feed back a maximum network load balancing path number P value;
  • the transceiver 820 is further configured to receive a second message from the network device that carries the P value.
  • the format of the first message and the second message may refer to the frame format of the TCP protocol of the prior art, and details are not described herein again.
  • the P value can be carried by extending the reserved field of the existing TCP protocol.
  • the processor is further configured to select P logical channels according to the flow segment identifier of the plurality of sub-data streams, and the P logical channel is selected.
  • P logical channels according to the flow segment identifier of the plurality of sub-data streams
  • the third embodiment of the present invention is an apparatus embodiment corresponding to the first embodiment of the method, and the description of the first embodiment of the method should also be applied to the third embodiment of the present invention.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • an embodiment of the present invention provides a network device 900, including:
  • the transceiver 920 receives a plurality of sub-data streams from an upstream node, wherein each sub-data stream carries a stream segment identifier.
  • the processor 910 is configured to perform a hash operation on the P-capacity according to the quintuple of the sub-data stream and the stream segment identifier, and map the data to multiple physical links to send to the next hop node, where P is The maximum number of load balancing paths on the network. P is an integer greater than or equal to 2.
  • the fourth embodiment of the present invention is an apparatus embodiment corresponding to the second embodiment of the method, and the description of the second embodiment of the method should also be applied to the fourth embodiment of the present invention.
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize the load balancing effect while avoiding network devices recording the status and path information of a single data stream.
  • FIG. 10 is a schematic block diagram of a network device according to another embodiment of the present invention.
  • the network device 1000 includes a processor 1010, a memory 1020, a bus 1030 and a user interface 1040, and a network interface 1050.
  • processor 1010 controls the operation of network device 1000, which may be a general purpose processor, a digital signal processor, an application specific integrated circuit, a field programmable gate array, or other programmable logic device.
  • a user interface 1060 configured to connect to a lower layer terminal device
  • the network interface 1050 is configured to connect to an upper layer network device.
  • bus 1050 which in addition to the data bus includes a power bus, a control bus, and a status signal bus.
  • various buses are labeled as bus system 1050 in the figure. It should be noted that the foregoing description of the structure of the network element can be applied to the embodiment of the present invention.
  • the memory 1020 may include a Read Only Memory (ROM) and a Random Access Memory (RAM), or other types of dynamic storage devices that may store information and instructions, or may be a disk storage.
  • the memory 1020 can be used to store instructions that implement the related methods provided by embodiments of the present invention. It will be appreciated that at least one of the cache and long term storage is programmed or loaded by the processor 1010 of the network element 1000 by programming or loading.
  • the memory is for storing computer executable program code, wherein when the program code includes an instruction, when the processor executes the instruction, the instruction causes the network Meta performs the following operations:
  • P is the maximum load balancing path number P of the network, and P is an integer greater than or equal to 2;
  • the data stream to be sent is divided into units of stream segments to generate a plurality of sub-data streams
  • the instruction causes the network element to perform the following operations:
  • each sub-data stream carries a stream segment identifier
  • the present invention can implement a fine-grained, load-sharing and more balanced technical solution in the network by dividing the data stream to be sent in units of stream segmentation and mapping to multiple logical channels, thereby effectively improving network utilization. Optimize load balancing.
  • a person skilled in the art may understand that all or part of the steps of implementing the above embodiments may be completed by hardware, or may be instructed by a program to execute related hardware, and the program may be stored in a computer readable storage medium.
  • the storage medium mentioned may be a read only memory, a magnetic disk or an optical disk or the like.

Abstract

L'invention concerne un procédé d'équilibrage de charge de réseau, comportant les étapes suivantes: un dispositif terminal met en place P canaux logiques, P étant le nombre maximal P de chemins pour l'équilibrage de charge de réseau, et P étant un entier supérieur ou égal à 2; le dispositif terminal divise un flux de données à émettre en prenant un segment de flux d'un seul tenant, et génère une pluralité de sous-flux de données; et le dispositif terminal associe la pluralité de sous-flux de données aux P canaux logiques, et les envoie à un dispositif de réseau. Selon la présente invention, en divisant un flux de données à émettre en prenant un segment de flux d'un seul tenant et en les associant à une pluralité de canaux logiques, la solution technique faisant intervenir une granularité fine et une répartition de charge plus équilibrée peut être réalisée dans un réseau, le taux d'utilisation du réseau est amélioré en pratique, l'effet d'équilibrage de charge est optimisé, et un dispositif de réseau est empêché d'enregistrer les informations d'état et de chemin concernant un seul flux de données à la fois.
PCT/CN2017/080586 2016-08-23 2017-04-14 Procédé, dispositif et système d'équilibrage de charge de réseau WO2018036173A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610710518.3 2016-08-23
CN201610710518.3A CN107770085B (zh) 2016-08-23 2016-08-23 一种网络负载均衡方法、设备及系统

Publications (1)

Publication Number Publication Date
WO2018036173A1 true WO2018036173A1 (fr) 2018-03-01

Family

ID=61246377

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/080586 WO2018036173A1 (fr) 2016-08-23 2017-04-14 Procédé, dispositif et système d'équilibrage de charge de réseau

Country Status (2)

Country Link
CN (1) CN107770085B (fr)
WO (1) WO2018036173A1 (fr)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110995609A (zh) * 2019-12-20 2020-04-10 新华三半导体技术有限公司 报文发送方法、装置、电子设备及存储介质
CN111586099A (zh) * 2020-04-01 2020-08-25 烽火通信科技股份有限公司 一种跨节点的数据备份同步方法及系统
CN113542051A (zh) * 2021-05-27 2021-10-22 贵州电网有限责任公司 一种基于软件定义网络的监控流量负载均衡方法
CN114095433A (zh) * 2021-11-25 2022-02-25 新华三技术有限公司合肥分公司 一种报文发送方法、装置、电子设备及介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112272933B (zh) * 2018-06-05 2022-08-09 华为技术有限公司 队列控制方法、装置及存储介质
CN111817973B (zh) * 2020-06-28 2022-03-25 电子科技大学 一种数据中心网络负载的均衡方法
CN115085757B (zh) * 2022-07-07 2023-08-25 Oppo广东移动通信有限公司 射频网络优化方法及相关装置

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6778495B1 (en) * 2000-05-17 2004-08-17 Cisco Technology, Inc. Combining multilink and IP per-destination load balancing over a multilink bundle
US20060007929A1 (en) * 2001-03-14 2006-01-12 Desai Bhavesh N Transmit and receive system for a cable data service
CN103348728A (zh) * 2010-12-20 2013-10-09 阿尔卡特朗讯 用于多信道分组传送的系统和方法

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8711703B2 (en) * 2010-10-29 2014-04-29 Telefonaktiebolaget L M Ericsson (Publ) Load balancing in shortest-path-bridging networks
CN102368742B (zh) * 2011-12-11 2014-08-06 湖南科技大学 一种公平性好效率高的多链路数据传输方法
CN103312607B (zh) * 2012-03-09 2016-12-21 华为技术有限公司 一种传输路径选择方法及装置
CN104954249B (zh) * 2014-03-27 2018-09-21 华为技术有限公司 一种报文转发方法、系统及装置
CN104581892B (zh) * 2014-12-05 2018-08-21 华为技术有限公司 一种终端的数据传输方法和装置
CN105791145B (zh) * 2014-12-18 2019-10-25 南京中兴新软件有限责任公司 基于等价多路径ecmp的报文传输方法及装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6778495B1 (en) * 2000-05-17 2004-08-17 Cisco Technology, Inc. Combining multilink and IP per-destination load balancing over a multilink bundle
US20060007929A1 (en) * 2001-03-14 2006-01-12 Desai Bhavesh N Transmit and receive system for a cable data service
CN103348728A (zh) * 2010-12-20 2013-10-09 阿尔卡特朗讯 用于多信道分组传送的系统和方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
A. FORD: "Architectural Guidelines for Multipath TCP Development", IETF RFC6182, 21 March 2011 (2011-03-21), pages 1 - 28, XP055603069, ISSN: 2070-1721 *
FU FA ET AL.: "Performance Analysis of MPTCP Protocol in Multiple Scenarios", COMPUTER ENGINEERING AND APPLICATIONS, vol. 52, no. 5, 31 March 2016 (2016-03-31) *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110995609A (zh) * 2019-12-20 2020-04-10 新华三半导体技术有限公司 报文发送方法、装置、电子设备及存储介质
CN111586099A (zh) * 2020-04-01 2020-08-25 烽火通信科技股份有限公司 一种跨节点的数据备份同步方法及系统
CN111586099B (zh) * 2020-04-01 2023-03-24 烽火通信科技股份有限公司 一种跨节点的数据备份同步方法及系统
CN113542051A (zh) * 2021-05-27 2021-10-22 贵州电网有限责任公司 一种基于软件定义网络的监控流量负载均衡方法
CN114095433A (zh) * 2021-11-25 2022-02-25 新华三技术有限公司合肥分公司 一种报文发送方法、装置、电子设备及介质
CN114095433B (zh) * 2021-11-25 2024-02-23 新华三技术有限公司合肥分公司 一种报文发送方法、装置、电子设备及介质

Also Published As

Publication number Publication date
CN107770085B (zh) 2021-08-31
CN107770085A (zh) 2018-03-06

Similar Documents

Publication Publication Date Title
WO2018036173A1 (fr) Procédé, dispositif et système d'équilibrage de charge de réseau
CN109412964B (zh) 报文控制方法及网络装置
US10554538B2 (en) Dynamic link state routing protocol
US7729328B2 (en) Real-time sessions for wireless mesh networks
US8942094B2 (en) Credit-based network congestion management
US20120275304A1 (en) Hierarchical profiled scheduling and shaping
WO2016062106A1 (fr) Procédé, dispositif et système de traitement de paquets
US8885502B2 (en) Feedback protocol for end-to-end multiple path network systems
CN105391567A (zh) 流量管理实现方法、装置和网络设备
CN111800351A (zh) 由交换机进行的拥塞通知分组生成
WO2021244450A1 (fr) Procédé et appareil de communication
WO2019001484A1 (fr) Procédé, appareil, et système de réglage d'un débit de côté envoi
CN111555982A (zh) 一种基于IPv6扩展头的报文智能选路的方法和系统
US8838782B2 (en) Network protocol processing system and network protocol processing method
US11245546B2 (en) Methods and systems for transmitting and receiving data packets through a bonded connection
CN109219944B (zh) 用于减小分组网络中mtu大小的系统和方法
JP5254916B2 (ja) リング網における通信装置および通信制御方法
WO2022242775A1 (fr) Procédé et système de traitement de paquets, et dispositif de réseau
JP6268027B2 (ja) 通信システム、送信装置、及び通信方法
JP5216830B2 (ja) データ転送装置及び方法
JP4828555B2 (ja) ノード装置および帯域制御方法
WO2024065481A1 (fr) Procédé et appareil de traitement de données, et dispositif de réseau et support d'enregistrement
JP4797033B2 (ja) Ｔｃｐフローレート制御エッジノードにおけるフローレート制御方法及びエッジノード
US7260114B1 (en) System and method for tracking multilink packet information
JP2021022838A (ja) データ伝送装置およびプログラム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17842598

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17842598

Country of ref document: EP

Kind code of ref document: A1