[go: up one dir, main page]

WO2018120814A1 - Procédé et appareil d'équilibrage de charge aléatoire, support de stockage d'informations - Google Patents

Procédé et appareil d'équilibrage de charge aléatoire, support de stockage d'informations Download PDF

Info

Publication number
WO2018120814A1
WO2018120814A1 PCT/CN2017/094397 CN2017094397W WO2018120814A1 WO 2018120814 A1 WO2018120814 A1 WO 2018120814A1 CN 2017094397 W CN2017094397 W CN 2017094397W WO 2018120814 A1 WO2018120814 A1 WO 2018120814A1
Authority
WO
WIPO (PCT)
Prior art keywords
output
link
selectable
load balancing
planes
Prior art date
Application number
PCT/CN2017/094397
Other languages
English (en)
Chinese (zh)
Inventor
常艳蕊
Original Assignee
深圳市中兴微电子技术有限公司
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 深圳市中兴微电子技术有限公司 filed Critical 深圳市中兴微电子技术有限公司
Publication of WO2018120814A1 publication Critical patent/WO2018120814A1/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

Definitions

  • the present invention relates to the field of large-capacity data exchange, and in particular, to a method and apparatus for random load balancing, and a computer storage medium.
  • the switching network implements data exchange between switching access devices.
  • the networking of the switching chip becomes complicated, the capacity becomes larger, and the number of Serdes links also increases, and large-capacity switching chips such as 128 ⁇ 128, 168 ⁇ 168, 192 ⁇ 192, 256 ⁇ 256 appear.
  • the switch chip is made in a multi-plane form.
  • FIG. 1 is a schematic diagram of a conventional multi-plane load balancing method, as shown in FIG. It is assumed that the number of links of the switch chip is 96 ⁇ 96, and plane 0, plane 1, plane 2, and plane 3 can all reach the access device through links 0 to 23, and the four planes simultaneously perform load balancing.
  • the output link is 0.
  • embodiments of the present invention are directed to a method and apparatus for random load balancing, and a computer storage medium, so that all cells of multiple routing planes can be uniformly allocated on reachable links to ensure the entire network.
  • the purpose of traffic balancing is to improve bandwidth utilization and system performance.
  • An embodiment of the present invention provides a method for random load balancing, where the method includes:
  • the M-selectable output links of the current sub-load balancing are processed by using a two-choice selector and a unique pseudo-random number in all planes to obtain a current sub-load balancing.
  • Effective output link; the n, M are positive integers;
  • the link number of the valid output link of the current current load balancing is masked in the M selectable link number mask tables of the current secondary load balancing, and the M-1 of the next load balancing is determined.
  • the masked link number in the selectable link number mask table is restored for the next round of load balancing.
  • the M selectable output links of the current sub-load balancing are processed by using a second-selector and a unique pseudo-random number in all planes to obtain an effective output of the current sub-load balancing.
  • Links including:
  • the k-level output chain is performed on the M selectable output links by using a second-selector and a unique pseudo-random number in all planes in the current time and a plane.
  • the grouping of the roads obtains the effective output link of the current current load balancing, including:
  • the m output links of the kth stage are grouped into two groups of adjacent output links, and the selectors of the second selection and the unique ones in all planes are utilized.
  • the m-1 output links of the kth stage are grouped into two groups of adjacent output links, and the selectors are selected and the planes are used in all planes.
  • the M-selectable output link of the current sub-load balancing is processed by using a second-selector and a unique pseudo-random number in all planes, including:
  • the only one of the all planes is a binary number, and the bit width of the only one of the all planes is the selection The number of signals;
  • the signal bit corresponding to the selection signal is selected for the second selection
  • the two selectable output links corresponding to the two inputs of the selector are selected to obtain a selectable output link corresponding to one output, including:
  • the determining, according to a preset rule of the signal bit corresponding to the selection signal, a selectable output link corresponding to the output end including:
  • the selectable output link corresponding to the first input end of the second selection selector is determined as a selectable output chain corresponding to the output end. Or; when the signal bit corresponding to the selection signal is 1, the selectable output link corresponding to the second input end of the second selection selector is determined as one corresponding to the output end; Selected output link.
  • the method further includes:
  • a unique one of all planes in one round of load balancing is generated according to a preset random function.
  • An embodiment of the present invention provides a device for random load balancing, where the device includes:
  • the processing module is configured to process the M selectable output links of the current sub-load balancing by using a second-selector and a unique pseudo-random number in all planes in one round of load balancing.
  • a determining module configured to block the link number of the valid output link of the current current load balancing in the M selectable link number mask tables of the current secondary load balancing, and determine M-1 selectable output links that are load balanced at a time; until all link numbers in the selectable link number mask table are masked;
  • the restoration module is configured to restore the blocked link number in the selectable link number mask table for the next round of load balancing.
  • the processing module is specifically configured to use the two selectors and a unique pseudo random number in all planes in the current time and a plane to the M selectable output chains.
  • the path performs a packet of the k-level output link to obtain the effective output link of the current current load balancing; the k is a positive integer.
  • the processing module is further configured to: in the current time and in a plane, use the two selectors and a unique pseudo random number in all planes to select the M outputs.
  • the link performs two-two packets of k-level adjacent output links;
  • the m output links of the kth stage are grouped into two groups of adjacent output links, and the selectors of the second selection and the unique ones in all planes are utilized.
  • the m-1 output links of the kth stage are grouped into two groups of adjacent output links, and the selectors are selected and the planes are used in all planes.
  • the processing module is further configured to generate a selection signal according to the unique one of the pseudo-random numbers in all planes; the only one of the all planes is a binary number, in all planes The bit width of the unique pseudo random number is the number of the selection signals;
  • the processing module is further configured to: if only one of the two selectable output links corresponding to the two inputs of the two-selector is valid, the effective output is The link serves as a selectable output link corresponding to the output;
  • the processing module is further configured to: when the signal bit corresponding to the selection signal is 1, the selectable output link corresponding to the first input end of the second selection selector is determined as a selectable output link corresponding to the output; or, when the signal bit corresponding to the selection signal is 1, the selectable output corresponding to the second input of the second selector The link is determined to be a selectable output link corresponding to the output.
  • the device further includes:
  • a generating module is configured to generate a unique one of all planes in one round of load balancing according to a preset random function.
  • the processing module, the determining module, the restoring module, and the generating module may use a central processing unit (CPU), a digital signal processor (DSP, Digital Singnal Processor), or Field-Programmable Gate Array (FPGA) implementation.
  • CPU central processing unit
  • DSP digital signal processor
  • FPGA Field-Programmable Gate Array
  • the embodiment of the present invention further provides a computer storage medium storing computer executable instructions configured to perform the method of random load balancing according to any one of the foregoing aspects.
  • Method and device for random load balancing provided by embodiment of the present invention, and computer storage
  • the medium through a round of n load balancing, using a two-selection selector and a unique pseudo-random number in all planes, processing the M selectable output links of the current sub-load balancing to obtain a current An effective output link of the secondary load balancing;
  • the n, M are positive integers;
  • the link number of the effective output link of the current current load balancing is M selectable links of the current secondary load balancing Masked in the mask table and determine the M-1 selectable output links for the next load balancing; until all link numbers in the selectable link number mask table are masked;
  • the masked link number in the selectable link number mask table is restored, and the next round of load balancing is performed; all cells of multiple routing planes can be uniformly allocated on the reachable link, and guaranteed.
  • the purpose of traffic balancing across the entire network improves bandwidth utilization and system performance.
  • FIG. 1 is a schematic diagram of a conventional multi-plane load balancing method
  • Embodiment 1 is a flowchart of Embodiment 1 of a method for random load balancing according to the present invention
  • Embodiment 3 is a flowchart of Embodiment 2 of a method for random load balancing according to the present invention
  • FIG. 4 is a schematic structural diagram of a second-selector of a method for random load balancing according to an embodiment of the present invention
  • FIG. 5 is a schematic diagram of a random load balancing structure of a routing plane of a 96 ⁇ 96 switching device according to an embodiment of a method for random load balancing according to the present invention
  • FIG. 6 is a schematic diagram of a load balancing cell flow direction in Embodiment 2 of a method for random load balancing according to the present invention
  • FIG. 7 is a schematic structural diagram of an apparatus for random load balancing according to the present invention.
  • Embodiment 1 of a method for random load balancing according to the present invention
  • the method for random load balancing provided by the embodiment of the present invention may include the following steps:
  • Step 201 In a round of load balancing, using the two selectors and a unique pseudo random number in all planes, the M selectable output links of the current secondary load balancing are processed to obtain a current The effective output link of the secondary load balancing; the n and M are positive integers.
  • the M selectable output links of the current secondary load balancing are performed in the current time and in a plane by using a second selector and a unique pseudo random number in all planes.
  • the packets of the level output link get a valid output link of the current secondary load balancing; where k is a positive integer.
  • a unique one of all planes in one round of load balancing is first generated according to a preset random function, and a selection signal is generated according to the pseudo random number;
  • the pseudo random number is a binary number whose bit width is the number of the selection signals.
  • the number of routing planes in load balancing is a
  • the number of selection signals required for load balancing in each routing plane is b.
  • first A unique pseudo-random number Y in the period is generated according to the preset random function f(x), and a*b selection signals are generated according to Y, wherein the bit width of the pseudo-random number Y is a*b.
  • the K selectable output links of the current sub-load balancing are subjected to k-level phase by using a two-select selector and a selection signal generated based on a unique pseudo-random number in all planes.
  • the two-two packets of the adjacent output link get a valid output link of the current secondary load balancing.
  • the k-th output link is grouped according to the parity of the k-th output link number m, specifically :
  • a k-level phase is performed on the M selectable output links of the current sub-load balancing by using a two-select selector and a selection signal generated according to a unique pseudo-random number in all planes.
  • two selectable output links corresponding to the two input ends of the two-selection selector are selected according to the signal bits corresponding to the selection signal, and the selection process thereof for:
  • the predetermined rule is determined according to the signal bit corresponding to the selection signal.
  • a selectable output link corresponding to the output end specifically: when the signal bit corresponding to the selection signal is 1, the selectable output link corresponding to the first input end of the second selection selector is determined as an output a selectable output link corresponding to the end; or, when the signal bit corresponding to the selection signal is 1, the selectable output link corresponding to the second input end of the second selector is determined as the output end Corresponding to an optional output link.
  • Step 202 Shield the link number of the valid output link of the current current load balancing in the M selectable link number mask tables of the current secondary load balancing, and determine the next load. Equalized M-1 selectable output links; until all link numbers in the selectable link number mask table are masked.
  • the link number of the valid output link is masked in the M selectable link number mask table of the current secondary load balancing, unmasked
  • the link will act as the M-1 actually selectable output links for the next load balancing; until all link numbers in the selectable link number mask table are masked, a round of load balancing is completed.
  • Step 203 Restore the blocked link number in the selectable link number mask table, and perform the next round of load balancing.
  • the random load balancing method selects M of the current sub-load balancing by using a two-selection selector and a unique pseudo-random number in all planes in one round of load balancing.
  • the output link is processed to obtain a valid output link of the current secondary load balancing; the n and M are positive integers; the link number of the effective output link of the current current load balancing is in the current time Shielding of load-balanced M selectable link number mask tables and determining M-1 selectable output links for next load balancing; until all of the selectable link number mask tables
  • the link number is masked; the shielded link number in the selectable link number mask table is restored, and the next round of load balancing is performed; using a two-choice selector and a load balancing mask
  • All the cells of multiple routing planes can be evenly distributed on the reachable links to ensure the traffic balance of the entire network.
  • the bandwidth utilization and system performance can be improved and can be used in various modes. Net knot
  • Embodiment 3 is a flowchart of Embodiment 2 of a method for random load balancing according to the present invention; this embodiment is described by taking a processing manner of a routing plane of a 96 ⁇ 96 switching device as an example, and the switching device includes four Routing plane: plane 0, plane 1, plane 2 and plane 3; the selection signal (select signal) of the two selectors used by the four routing planes is generated by the same random function generator, and the selectable output link is 96.
  • the method for random load balancing provided by the embodiment of the present invention may include the following steps:
  • Step 301 The 96 selectable output links are grouped into two groups and divided into 48 groups. After the first level selection, 48 output links are generated.
  • the 96 selectable output links are grouped into two groups of adjacent links, which can be divided into 48 groups in total; after the 48 sets of selectable output links are selected by the first choice of 48 two-selectors, Get 48 output links.
  • the selector of two options includes two input terminals, one selection signal terminal, and one output terminal;
  • the inputs of the selected selector are link a and link b, respectively, and one of link a and link b is selected as the output link according to link a, link b and selection signal; specifically, when the link When a is valid and link b is invalid, link a is selected as the output link; when link a is invalid and link b is valid, link b is selected as the output link; when link a and link b are both When valid, the corresponding bit of the selection signal (select signal) is used to select link a or link b as an output link; for example, it may be preset to select a link when the corresponding bit of the selection signal (select signal) is 1.
  • a is used as an output link; or, it may be preset that the link b is selected as an
  • Step 302 The generated 48 output links are grouped into two groups and divided into 24 groups. After the second level is selected, 24 output links are generated.
  • the two output links of the adjacent output links are continued for the 48 output links, and the total can be divided into 24 groups, and then The 24 sets of output links use 24 two-select selectors for second-level selection, resulting in 24 output links.
  • Step 303 The generated 24 output links are grouped into two groups, and are divided into 12 groups. After the third level is selected, 12 output links are generated.
  • the 24 output links After two stages of selection of 96 selectable output links, after generating 24 output links, the 24 output links continue to be grouped into two groups of adjacent output links, divided into 12 groups, and then 12 groups The output link uses 12 two-select selectors for third-level selection, resulting in 12 output links.
  • Step 304 The generated 12 output links are grouped into two groups, and are divided into six groups. After the fourth level is selected, six output links are generated.
  • Step 305 The generated 6 output links are grouped into two groups, and are divided into three groups. After the fifth level is selected, three output links are generated.
  • the two output links After four selections of 96 selectable output links are generated, after six output links are generated, the two output links continue to be grouped into two groups of adjacent output links, and are divided into three groups, and then 3 The group output link uses three alternative selectors for the fifth level selection, resulting in three output links.
  • Step 306 Divide the generated three output links into two groups and one group, and perform a sixth level selection on the output chains of the two groups to generate one output link and one output remaining. link.
  • the first and second output links or the second and third output links of the three output links are divided. For the first group, the remaining one output link is the second group; then, the two output links of the first group are selected by the second selection selector for the sixth level, and one output link is generated. At the same time, there is one output link remaining.
  • Step 307 Performing a seventh-level selection of one output link generated by the sixth stage and one remaining output link of the fifth stage to obtain an effective output link.
  • the output link is the final payload balanced link.
  • Step 308 Shield the obtained link number of a valid output link in the link number mask table of the 96 selectable output links to generate a selectable output link for the next load balancing.
  • the link number of the load balancing link is masked in the link number mask table of the 96 selectable output links. , produces a selectable output link for the next load balancing.
  • the location corresponding to the payload equalization link may be 1 and fed back to the load balancing input end, and the link number is in 96 selectable link number masks. Shielded in the table to obtain the next output loadable selectable output link, so that the next blocked load will not be selected.
  • FIG. 5 is a schematic diagram of a random load balancing structure of a routing plane of a 96 ⁇ 96 switching device according to an embodiment of a method for random load balancing according to the present invention; as shown in FIG. 5, 96 selectable output links are performed by using a selective selector. Two-two packets and selection of seven adjacent output links, and finally an output link is generated as a payload equalization link; and a corresponding mask is fed back to the input of the load balancing, and the payload is balanced. The link number is masked in the link number mask table of the 96 selectable output links, resulting in a selectable output link for the next load balancing.
  • each routing plane in the switching device is identical; that is, steps 301 to 308 are performed on plane 0, plane 1, plane 2, and plane 3, and each of the active load balancing links is synchronously generated. , the output link to the switching device.
  • a Y is generated, and the bit width is the number of selection signals (select signals), that is, 4 ⁇ 95.
  • preset random function can be set according to actual needs, which is not limited herein.
  • FIG. 6 is a schematic diagram of a load balancing cell flow direction in Embodiment 2 of a method for random load balancing according to the present invention; as shown in FIG. 6, after a load balancing is performed on a total of 96 selectable output links from 0 to 95, a plane is used.
  • the output link generated by 0 is the link 0
  • the output link generated by plane 1 is the link 1
  • the output link generated by plane 2 is the link 2
  • the output link generated by plane 3 is the link 3. That is, four planes can be load balanced to different links respectively, avoiding multi-plane resonance.
  • Step 309 Determine whether all link numbers in the selectable link number mask table are blocked.
  • Step 310 is performed. If all the link numbers in the selectable link number mask table are blocked, step 311 is performed.
  • Step 310 Perform the next load balancing.
  • Step 311 Restore the blocked link number in the selectable link number mask table, and perform the next round of load balancing.
  • the next load balancing is generated according to the mask table of the load balancing link and the selectable output link.
  • the output link can be selected, that is, the link number of the generated payload equalization link is shielded in the currently selectable link number mask table, which can avoid the same round of load balancing twice randomly to the same link;
  • the random selection signal select signal
  • the random load balancing method performs two-two grouping on 96 selectable output links, and after the first-stage selection, generates 48 output links; 48 output links to be generated are performed. Two or two groups, after the second level selection, generate 24 output links; the generated 24 output links are grouped into two groups, and after the third level is selected, 12 output links are generated; 12 will be generated.
  • the output link performs two-two grouping, and after the fourth-stage selection, six output links are generated; the six output links that are generated are grouped into two groups, and after the fifth-level selection, three output links are generated;
  • the generated three output links are divided into two groups and one group, and the second level of the output links of the two groups is selected to generate one output link and one output link remaining;
  • the first output link generated by the sixth stage and the remaining one output link of the fifth stage are subjected to the seventh level selection to obtain an effective output link; the link number of the obtained effective output link is 96.
  • FIG. 7 is a schematic structural diagram of an apparatus for random load balancing according to the present invention.
  • the random load balancing apparatus 07 provided by the embodiment of the present invention includes: a processing module 71, a determining module 72, and a restoring module 73;
  • the processing module 71 is configured to perform the current sub-load balanced M selectable output links by using a second-selector and a unique pseudo-random number in all planes in one round of load balancing. Processing, obtaining a valid output link of the current secondary load balancing; the n and M are positive integers;
  • the determining module 72 is configured to block the link number of the valid output link of the current current load balancing in the M selectable link number mask tables of the current secondary load balancing, and determine M-1 selectable output links that are load balanced at a time; until all link numbers in the selectable link number mask table are masked;
  • the restoration module 73 is configured to restore the blocked link number in the selectable link number mask table for the next round of load balancing.
  • the processing module 71 is configured to: in the current time and in a plane, use a second selector and a unique pseudo random number in all planes, and the M is used.
  • the selectable output links are grouped into k-level output links to obtain the effective output link of the current current load balancing; the k is a positive integer.
  • the processing module 71 is further configured to: in the current time and in a plane, use a second selector and a unique pseudo random number in all planes, M selectable output links for two-two packets of k-level adjacent output links;
  • the m output links of the kth stage are grouped into two groups of adjacent output links, and the selectors of the second selection and the unique ones in all planes are utilized.
  • the m-1 output links of the kth stage are grouped into two groups of adjacent output links, and the selectors are selected and the planes are used in all planes.
  • the processing module 71 is further configured to generate a selection signal according to the unique one of the pseudo-random numbers in all planes; the only one of the all planes is a binary number The bit width of the only one of the pseudo-random numbers in all the planes is the number of the selection signals;
  • the processing module 71 is further configured to: if only one of the two selectable output links corresponding to the two input ends of the two-selection selector is valid, Using the valid output link as a selectable output link corresponding to the output;
  • the processing module 71 is further configured to: when the signal bit corresponding to the selection signal is 1, select a corresponding one of the first input ends of the second selection selector The output link is determined as a selectable output link corresponding to the output end; or, when the signal bit corresponding to the selection signal is 1, the second input end of the second selector is selected Corresponding selectable output links are determined to be an alternative to the output Output link.
  • the device 07 further includes: a generating module 74;
  • the generating module 74 is configured to generate a unique one of all planes in one round of load balancing according to a preset random function.
  • the device of the random load balancing in this embodiment may be used to perform the technical solution of the foregoing method embodiment, and the implementation principle and the technical effect are similar, and details are not described herein again.
  • the processing module 71, the determining module 72, the restoring module 73, and the generating module 74 of the stochastic load balancing device 07 may each be a central processing unit (CPU) located in the random load balancing device 07. ), a microprocessor (Micro Processor Unit, MPU), a digital signal processor (DSP), or a Field Programmable Gate Array (FPGA).
  • CPU central processing unit
  • MPU Micro Processor Unit
  • DSP digital signal processor
  • FPGA Field Programmable Gate Array
  • embodiments of the present invention can be provided as a method, system, or computer program product. Accordingly, the present invention can take the form of a hardware embodiment, a software embodiment, or a combination of software and hardware. Moreover, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • the M selectable output links of the current sub-load balancing are processed by using a two-selection selector and a unique pseudo-random number in all planes in one round of load balancing.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

Les modes de réalisation de la présente invention concernent un procédé d'équilibrage de charge aléatoire, comprenant les étapes suivantes : pendant un cycle de n équilibres de charge, utilisation d'un sélecteur deux en un et d'un nombre pseudo-aléatoire unique sur tous les plans pour traiter M liaisons de sortie sélectionnables de l'équilibre de charge courant afin d'obtenir une liaison de sortie effective de l'équilibre de charge courant; n et M sont des nombres entiers positifs; masquage du numéro de liaison de la liaison de sortie effective de l'équilibrage de charge courant dans une table de codes de masque des M numéros de liaison sélectionnables de l'équilibre de charge courant, et détermination de M -1 liaisons de sortie sélectionnables de l'équilibre de charge suivant; poursuite jusqu'à ce que tous les numéros de liaison dans la table de codes de masque de liaison sélectionnable soient masqués; rétablissement des numéros de liaison masqués dans la table de codes de masque de liaison sélectionnable, et réalisation du prochain cycle d'équilibrage de charge. Un appareil d'équilibrage de charge aléatoire et un support de stockage d'informations sont également décrits dans les modes de réalisation de la présente invention.
PCT/CN2017/094397 2016-12-26 2017-07-25 Procédé et appareil d'équilibrage de charge aléatoire, support de stockage d'informations WO2018120814A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201611215878.2A CN108243113B (zh) 2016-12-26 2016-12-26 随机负载均衡的方法及装置
CN201611215878.2 2016-12-26

Publications (1)

Publication Number Publication Date
WO2018120814A1 true WO2018120814A1 (fr) 2018-07-05

Family

ID=62705016

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/094397 WO2018120814A1 (fr) 2016-12-26 2017-07-25 Procédé et appareil d'équilibrage de charge aléatoire, support de stockage d'informations

Country Status (2)

Country Link
CN (1) CN108243113B (fr)
WO (1) WO2018120814A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113767597A (zh) * 2020-04-03 2021-12-07 华为技术有限公司 用于基于周期的负载均衡的网络设备、系统和方法
CN116170449A (zh) * 2023-02-24 2023-05-26 浪潮电子信息产业股份有限公司 一种流量均衡方法、系统、装置及计算机可读存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111935029B (zh) * 2020-09-18 2021-06-08 腾讯科技（深圳）有限公司 网关负载均衡方法和装置、存储介质及电子设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1426211A (zh) * 2001-12-06 2003-06-25 富士通株式会社 服务器负载分担系统
CN101217476A (zh) * 2008-01-07 2008-07-09 华为技术有限公司 一种多平面包交换方法、系统及设备
US7480246B2 (en) * 2002-03-06 2009-01-20 Agere Systems Inc. Characterizing transmission of data segments within a switch fabric using multiple counters for each destination node
US7596135B1 (en) * 2003-05-23 2009-09-29 Cisco Technology, Inc. Method and apparatus for mixed-cast routing through a Clos-like network
CN102238072A (zh) * 2010-05-06 2011-11-09 中兴通讯股份有限公司 一种动态选择路由的方法及clos交换网系统
CN105337883A (zh) * 2015-08-20 2016-02-17 电子科技大学 一种支持多业务的网络交换装置及其实现方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101227394B (zh) * 2008-02-18 2011-07-13 中兴通讯股份有限公司 一种大容量无阻塞路由矩阵
CN102065014B (zh) * 2010-12-29 2014-12-31 中兴通讯股份有限公司 数据信元处理方法和装置
CN103152281B (zh) * 2013-03-05 2014-09-17 中国人民解放军国防科学技术大学 基于两级交换的负载均衡调度方法
CN105376168B (zh) * 2014-08-25 2019-06-11 深圳市中兴微电子技术有限公司 一种负载均衡的方法与装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1426211A (zh) * 2001-12-06 2003-06-25 富士通株式会社 服务器负载分担系统
US7480246B2 (en) * 2002-03-06 2009-01-20 Agere Systems Inc. Characterizing transmission of data segments within a switch fabric using multiple counters for each destination node
US7596135B1 (en) * 2003-05-23 2009-09-29 Cisco Technology, Inc. Method and apparatus for mixed-cast routing through a Clos-like network
CN101217476A (zh) * 2008-01-07 2008-07-09 华为技术有限公司 一种多平面包交换方法、系统及设备
CN102238072A (zh) * 2010-05-06 2011-11-09 中兴通讯股份有限公司 一种动态选择路由的方法及clos交换网系统
CN105337883A (zh) * 2015-08-20 2016-02-17 电子科技大学 一种支持多业务的网络交换装置及其实现方法

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113767597A (zh) * 2020-04-03 2021-12-07 华为技术有限公司 用于基于周期的负载均衡的网络设备、系统和方法
CN116170449A (zh) * 2023-02-24 2023-05-26 浪潮电子信息产业股份有限公司 一种流量均衡方法、系统、装置及计算机可读存储介质
CN116170449B (zh) * 2023-02-24 2025-08-26 浪潮电子信息产业股份有限公司 一种流量均衡方法、系统、装置及计算机可读存储介质

Also Published As

Publication number Publication date
CN108243113A (zh) 2018-07-03
CN108243113B (zh) 2020-06-16

Similar Documents

Publication Publication Date Title
US9246821B1 (en) Systems and methods for implementing weighted cost multi-path using two-level equal cost multi-path tables
Kim et al. Adaptive routing in high-radix clos network
US8819272B2 (en) Multiprocessor communication networks
CN111913749A (zh) 基于流水线的sm3算法fpga实现方法及系统
Hu et al. DMesh: a diagonally-linked mesh network-on-chip architecture
WO2018120814A1 (fr) Procédé et appareil d'équilibrage de charge aléatoire, support de stockage d'informations
CN109391547B (zh) 一种网络拓扑系统及其拓扑建立与路由表建立方法
Xu et al. Reliable service function chain provisioning in software-defined networking
CN107959642B (zh) 用于测量网络路径的方法、装置和系统
Miura et al. An adaptive routing of the 2-D torus network based on turn model
Swaminathan et al. Enhanced Noxim simulator for performance evaluation of network on chip topologies
Bhaskar et al. A study of the effect of virtual channels on the performance of network-on-chip
Hu et al. A symmetric odd-even routing model in network-on-chip
CN115208769B (zh) 适用于Dragonfly拓扑的环形通信方法
Tu et al. Design a simple and high performance switch using a two-stage architecture
Liang et al. Beyond the performance of three-tier fat-tree: equality topology with low diameter
Punhani et al. E-XY: an entropy based XY routing algorithm
Lei et al. Vertical-mesh-conscious-dynamic routing algorithm for 3D NoCs
Miura et al. The Static and Dynamic Performance of an Adaptive Routing Algorithm of 2-D Torus Network Based on Turn Model
Momeni et al. Improved-XY: A High Performance Wormhole-Switched Routing Algorithm for Irregular 2-D Mesh NoC
Chauhan et al. Comparative analysis of traffic patterns on centre connected topologies based on burton normal form
Dai et al. PDA-HyPAR: Path-diversity-aware hybrid planar adaptive routing algorithm for 3D NoCs
Menon et al. Adaptive look ahead algorithm for 2-D mesh NoC
CN104734995B (zh) 一种链路汇聚流量上限的设置方法和网络控制器
Saravanakumar et al. Energy and throughput analysis of multicast routing algorithm for 2D mesh network on chip

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: 17887932

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: 17887932

Country of ref document: EP

Kind code of ref document: A1