Data center network topology construction and routing design method based on cube structure

By adopting a cube-structured network topology and VEPL routing algorithm in the data center, and using 60GHz wireless communication technology, the deployment difficulties and routing path length problems of traditional wired data centers are solved, efficient data transmission and network connectivity are achieved, and data is ensured. The center is operating normally in the event of a failure.

CN116112370BActive Publication Date: 2024-11-12HENAN UNIVERSITY
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202310133093.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-02-17
Publication Date
2024-11-12
Estimated Expiration
2043-02-17

AI Technical Summary

Technical Problem

Traditional wired data centers have problems such as deployment difficulties, complex cabling and long routing paths, resulting in poor network connectivity and inability to flexibly cope with various data center deployment environments.

Method used

Using the data center network topology structure and routing design method based on cube structure, using 60GHz wireless communication technology, the VEPL routing algorithm of CaF features is designed, including vertex routing algorithm and in-plane Routing algorithm to build a three-dimensional network structure to realize wireless connection.

Benefits of technology

Improves the performance of the data center, ensures high data transmission rates, shortens the routing path length, enhances network connectivity, and ensures the normal operation of the network in the event of rack failure.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116112370B_ABST
    Figure CN116112370B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of data center network topology structure, and specifically to a data center network topology construction and routing design method based on a cube structure, comprising: first determining the network scale of each network unit of the data center network, then deploying the wireless transceiver above each rack in the network unit as a cube unit, then obtaining the relative coordinates of each rack in each network unit and the relative height of each wireless transceiver, and dividing each wireless transceiver in each network unit into vertex nodes, in-plane nodes and in-body nodes, designing a corresponding VEPL routing algorithm according to CaF characteristics, and completing the actual deployment of the CaF three-dimensional hierarchical structure in the data center. The present invention overcomes the defects of poor connectivity, poor flexibility in coping with various data center deployment environments, and large routing paths that still exist in wireless data center networks, and can be applied to the field of constructing wireless data center network topology structures.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data center network topology structure, and in particular to a data center network topology construction and routing design method based on a cube structure. Background Art

[0002] With the increasing development of cloud computing and big data technology, more and more applications need to be supported by data centers. Therefore, in order to meet the scalability, throughput, bandwidth and other characteristics of data centers, the network topology of traditional wired data centers is designed and analyzed. However, traditional wired data centers have problems such as difficult deployment, complex wiring and high cost, which are not conducive to the performance of data centers. In order to overcome the defects of traditional wired data centers, wireless data center networks have been introduced, but the existing wireless data center networks still have problems such as poor connectivity, inability to flexibly cope with the deployment environment of various data centers, and long routing paths. Summary of the invention

[0003] In order to solve the technical problems of poor connectivity, inability to flexibly cope with the deployment environments of various data centers, and long routing paths in the above-mentioned existing data center networks, the purpose of the present invention is to provide a data center network topology construction and routing design method based on a cube structure.

[0004] The present invention provides a data center network topology construction and routing design method based on a cube structure, comprising the following steps:

[0005] Determine the number of row racks in each network unit according to the row spacing of the data center network and the preset network unit size, determine the number of column racks in each network unit according to the column spacing of the data center network and the preset network unit size, determine the position of the central rack based on the number of row racks and the number of column racks in each network unit, and delete the central rack in each network unit based on the position of the central rack;

[0006] Determine the relative coordinates of each rack in each network unit according to the coordinates of each rack in each network unit, the row spacing, column spacing, number of row racks, and number of column racks of the data center network;

[0007] According to the preset maximum transmission distance of the millimeter wave and the preset network unit size corresponding to the data center network, the maximum height at which the antenna horn is deployed is determined; according to the layer number of the wireless transceiver of each rack in each network unit from the top of the rack, the total number of network layers, and the maximum height at which the antenna horn is deployed, the relative height of the wireless transceiver from each layer on the top of the rack is determined;

[0008] According to each wireless transceiver in each network unit, the number of row racks, the number of column racks, and the hierarchical deployment of the wireless transceivers above all racks in each network unit, the hierarchical deployment is a cube unit, constructing the position of each wireless transceiver in each network unit in the cube, and dividing each wireless transceiver in each network unit into vertex nodes, in-plane nodes, and in-body nodes based on the position in the cube;

[0009] A corresponding VEPL routing algorithm is designed according to the CaF characteristics, and the VEPL routing algorithm includes a vertex routing algorithm and an intra-plane routing algorithm.

[0010] Furthermore, the calculation formula for the number of row racks in each network unit is:

[0011]

[0012] Among them, m0 is the number of row racks in each network unit, p is the row spacing of the data center network, and a is the preset network unit size. For Round down;

[0013] The calculation formula for the number of column racks in each network unit is:

[0014]

[0015] Among them, n0 is the number of column racks in each network unit, q is the column spacing of the data center network, and a is the preset network unit size. For Round down.

[0016] Further, based on the number of row racks and the number of column racks in each network unit, determining the location of the central rack includes:

[0017] The ratio of the number of row racks to the value 2 is calculated, and the ratio is used as the abscissa of the central rack. The ratio of the number of column racks to the value 2 is calculated, and the ratio is used as the column coordinate of the central rack. The position of the central rack is determined based on the abscissa and column coordinates of the central rack.

[0018] Further, the relative coordinates of each rack in each network unit are determined according to the coordinates of each rack in each network unit, the row spacing, column spacing, number of row racks, and number of column racks of the data center network, including:

[0019] According to the horizontal coordinates of each rack in each network unit, the row spacing of the data center network, and the number of row racks, the relative horizontal coordinates of each rack in each network unit are determined. The calculation formula of the relative horizontal coordinates of each rack is:

[0020] x=p(i-1)%(m0-1)

[0021] Wherein, x is the relative horizontal coordinate of each rack in each network unit, p is the row spacing of the data center network, i is the horizontal coordinate of each rack in each network unit, m0 is the number of row racks in each network unit, % is the remainder, p(i-1)%(m0-1) is the remainder of the ratio of p(i-1) to (m0-1), and p(i-1) and (m0-1) are both integers;

[0022] According to the ordinates of each rack in each network unit, the column spacing of the data center network, and the number of column racks, the relative ordinates of each rack in each network unit are determined. The calculation formula of the relative ordinates of each rack is:

[0023] y=q(j-1)%(n0-1)

[0024] Among them, y is the relative vertical coordinate of each rack in each network unit, q is the column spacing of the data center network, j is the vertical coordinate of each rack in each network unit, n0 is the number of column racks in each network unit, % is the remainder, q(j-1)%(n0-1) is the remainder of the ratio of q(j-1) to (n0-1), and q(j-1) and (n0-1) are both integers.

[0025] Further, according to the preset maximum transmission distance of the millimeter wave and the preset network unit size corresponding to the data center network, the maximum height at which the antenna horn is deployed is determined, including:

[0026] Calculate the square of the maximum transmission distance of the preset millimeter wave and the square of the preset network unit size, and then calculate the product of the square of the preset network unit size and the value 2, so that the square of the maximum transmission distance of the preset millimeter wave is subtracted from the product of the square of the preset network unit size and the value 2, determine the difference, take the square root of the difference, and use the square root value as the maximum height at which the antenna horn is deployed.

[0027] Further, according to the layer number sequence of the wireless transceiver of each rack in each network unit from the top of the rack, the total number of network layers, and the maximum height at which the antenna horn is deployed, the relative height of the wireless transceiver from each layer of the top of the rack is determined, including:

[0028] The difference between any layer number of the wireless transceiver from the top of the rack and the value 1 is used as the numerator of the ratio, and the difference between the total number of network layers and the value 1 is used as the denominator of the ratio. The product of the ratio and the maximum height at which the antenna horn is deployed is calculated, and the product is used as the relative height of the wireless transceiver from the top layer of the rack corresponding to the layer number.

[0029] Furthermore, a corresponding VEPL routing algorithm is designed according to the CaF characteristics, including:

[0030] First, the vertex routing algorithm is used, and the source node transmits the wireless signal to the vertex node closest to the destination. If the vertex node is in an occupied or faulty state, the intra-plane routing algorithm is used to forward it to the intra-plane node closest to the destination, and then the next hop is selected through the vertex routing algorithm. When the last forwarded node is one hop away from the destination, it is directly forwarded. The last forwarded node is a vertex node or an intra-plane node.

[0031] The present invention has the following beneficial effects:

[0032] The present invention proposes a data center network topology construction and routing design method based on a cube structure. The method utilizes 60GHz wireless communication technology to overcome the disadvantages of difficult deployment, complex wiring and easy generation of hot nodes in traditional wired data centers based on cable or optical connections, which is conducive to ensuring high data transmission rate and improving the performance of the data center. Placing a 60GHz millimeter wave wireless transceiver on the top of a rack can avoid signal blocking, and directly establishing wireless connections between nodes can enable the network to have a larger average point degree, improve network connectivity, and effectively shorten the routing path length between the sending node and the receiving node. The unit and the whole of the CaF three-dimensional network structure are constructed so that adjacent units have common vertices and common edges, and a corresponding routing algorithm is designed so that the normal operation of the entire data center network can be guaranteed even when a forwarding individual rack is in a fault state. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 This is a flow chart of a data center network topology construction and routing design method based on a cube structure of the present invention;

[0035] Figure 2 A top view of a data center network composed of four network units and with racks arranged in an array in an embodiment of the present invention;

[0036] Figure 3 A schematic diagram of a three-dimensional coordinate system corresponding to a basic unit structure of a cube layered network in an embodiment of the present invention;

[0037] Figure 4 It is a schematic diagram of each wireless transceiver device in actual hierarchical deployment in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the technical solutions proposed by the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0039] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0040] This embodiment uses 60GHz wireless communication technology to solve the problems of difficult deployment, complex wiring, and easy generation of hot nodes in traditional wired data centers based on cable connections or optical fiber connections. It can ensure high data transmission rates and greatly improve the performance of data centers. Specifically, a data center network topology construction and routing design method based on a cube structure is proposed, such as Figure 1 As shown, the following steps are included:

[0041] S1, determine the number of row racks in each network unit according to the row spacing of the data center network and the preset network unit size, determine the number of column racks in each network unit according to the column spacing of the data center network and the preset network unit size, determine the position of the central rack based on the number of row racks and the number of column racks in each network unit, and delete the central rack in each network unit based on the position of the central rack.

[0042] It should be noted that when constructing a data center network topology, the network scale of each network unit in the data center network is first determined. The network scale includes the number of row racks and column racks in the network unit, the arrangement of the racks, and the deployment range of the racks in the network unit. The specific steps include:

[0043] The first step is to determine the number of row racks in each network unit according to the row spacing of the data center network and the preset network unit size, and determine the number of column racks in each network unit according to the column spacing of the data center network and the preset network unit size.

[0044] In this embodiment, the data center network topology is constructed from the network unit to the entire data center, and the arrangement of the racks in each network unit can be in the form of an array. The row spacing of the data center network here is the spacing distance value between adjacent racks in each row of racks in the network unit, and the column spacing is the spacing distance value between adjacent racks in each column of racks. The row spacing and column spacing are both known values ​​and can be set according to the deployment of the data center. The network unit size here refers to the length of the actual occupied area of ​​each network unit. The network unit size can be recorded as a, and the actual occupied area of ​​each network unit can be a×a, that is, the network unit can be deployed within the range of a×a. The preset network unit size a can be set by the implementer according to the actual occupied area size of the data center network, without specific limitation.

[0045] The calculation formula for the number of row racks in each network unit is:

[0046]

[0047] Among them, m0 is the number of row racks in each network unit, p is the row spacing of the data center network, and a is the preset network unit size. For Round down.

[0048] The formula for calculating the number of column racks in each network unit is:

[0049]

[0050] Among them, n0 is the number of column racks in each network unit, q is the column spacing of the data center network, and a is the preset network unit size. For Round down.

[0051] It should be noted that determining the network scale of the network unit through the actual data status of the data center network is conducive to improving the performance superiority of the data center network topology structure and facilitating the subsequent determination of the location of the central rack of each network unit. According to the number of row racks m0 and the number of column racks n0 in each network unit, the total number of racks in each network unit is calculated, which is m0×n0-1, that is, each network unit has m0×n0-1 racks.

[0052] In the second step, based on the number of row racks and the number of column racks in each network unit, the position of the central rack is determined, and the central rack in each network unit is deleted based on the position of the central rack.

[0053] In this embodiment, in order to avoid interference and improve the data transmission rate, it is necessary to delete the central rack in each network unit. First, the position of the central rack in each network unit is determined. Specifically, the ratio of the number of row racks to the value 2 is calculated, and the ratio is used as the horizontal coordinate of the central rack. The ratio of the number of column racks to the value 2 is calculated, and the ratio is used as the column coordinate of the central rack. The position of the central rack is determined based on the horizontal coordinate and column coordinate of the central rack. Since the number of row racks is m0 and the number of column racks is n0, the coordinates of the central rack can be like and If it is not an integer, it can be rounded up and the result of the rounding up is used as the position coordinate of the central rack. For example, in a data center network with racks arranged in 7 rows and 5 columns, the coordinate position of the central rack is (4,3). Then, based on the position of the central rack Delete the central rack in each network unit.

[0054] S2, determining the relative coordinates of each rack in each network unit according to the coordinates of each rack in each network unit, the row spacing, column spacing, the number of row racks, and the number of column racks of the data center network.

[0055] It should be noted that the position of a rack in each network unit in the coordinate system is (i, j), that is, the rack is located in the i-th row and j-th column, and the relative coordinates (x, y) of the rack are the positions located in the i-th row and j-th column after adding the row and column spacing. The relative coordinate position is relative to the data center to which it is deployed. The CaF network structure is constructed in three dimensions from the unit to the entire data center, so it is necessary to determine the relative position (x, y) of the rack located in the i-th row and j-th column in each network unit. At the same time, i is the horizontal coordinate of each rack in each network unit, and j is the vertical coordinate of each rack in each network unit. The steps include:

[0056] The first step is to determine the relative horizontal coordinates of each rack in each network unit according to the horizontal coordinates of each rack in each network unit, the row spacing of the data center network, and the number of row racks. The calculation formula of the relative horizontal coordinates of each rack is:

[0057] x=p(i-1)%(m0-1)

[0058] Among them, x is the relative horizontal coordinate of each rack in each network unit, p is the row spacing of the data center network, i is the horizontal coordinate of each rack in each network unit, m0 is the number of row racks in each network unit, % is the remainder, p(i-1)%(m0-1) is the remainder of the ratio of p(i-1) to (m0-1), and p(i-1) and (m0-1) are both integers.

[0059] For p(i-1)%(m0-1) in the calculation formula, the numbers before and after % are integers, and the positive or negative sign of the remainder is determined by the integer to the left of %. If the number to the left of % is positive, then the remainder is also positive. If the number to the left of % is negative, then the remainder is also negative. For example, a rack is located in the third row, that is, i is equal to 3. From the calculation formula of the relative horizontal coordinate, it can be known that the relative horizontal coordinate x of the rack is x=2p%(m0-1). The row spacing p and the number of row racks m0 are both known values. The number to the left of % should be a positive number, and the relative horizontal coordinate x is also a positive number.

[0060] The second step is to determine the relative ordinate of each rack in each network unit according to the ordinate of each rack in each network unit, the column spacing of the data center network, and the number of column racks. The calculation formula of the relative ordinate of each rack is:

[0061] y=q(j-1)%(n0-1)

[0062] Among them, y is the relative vertical coordinate of each rack in each network unit, q is the column spacing of the data center network, j is the vertical coordinate of each rack in each network unit, n0 is the number of column racks in each network unit, % is the remainder, q(j-1)%(n0-1) is the remainder of the ratio of q(j-1) to (n0-1), and q(j-1) and (n0-1) are both integers.

[0063] At this point, based on the relative horizontal and vertical coordinates of each rack in each network unit, the relative coordinates of each rack in each network unit can be obtained. It is worth noting that adjacent network units in the data center can communicate through common vertices and common edges, which helps to achieve the interconnection of the entire network and improve the development of wireless data centers. The top view of the data center network composed of 4 network units and the racks are arranged in an array is shown in the figure below. Figure 2 shown.

[0064] S3, according to the preset maximum transmission distance of the millimeter wave corresponding to the data center network and the preset network unit size, determine the maximum height at which the antenna horn is deployed, and according to the layer number of the wireless transceiver of each rack in each network unit from the top of the rack, the total number of network layers and the maximum height at which the antenna horn is deployed, determine the relative height of the wireless transceiver from each layer on the top of the rack, wherein the steps include:

[0065] The first step is to determine the maximum height at which the antenna horn is deployed based on the preset maximum transmission distance of the millimeter wave and the preset network unit size corresponding to the data center network.

[0066] To facilitate subsequent calculations of the relative height of the wireless transceiver device from the top layer of each rack, it is necessary to determine the maximum height at which the antenna horn is deployed. Specifically, calculate the square of the maximum transmission distance of the preset millimeter wave and the square of the preset network unit size, and then calculate the product of the square of the preset network unit size and the value 2. Subtract the product of the square of the preset network unit size and the value 2 from the square of the maximum transmission distance of the preset millimeter wave to determine the difference, take the square root of the difference, and use the value after taking the square root as the maximum height at which the antenna horn is deployed.

[0067] In this embodiment, the calculation formula for the maximum height at which the antenna horn is deployed can be:

[0068]

[0069] where h is the maximum height at which the antenna horn is deployed, Q is the maximum transmission distance of the preset millimeter wave, and the preset millimeter wave can be 60 GHz, and a is the preset network unit size.

[0070] In the second step, according to the layer number sequence of the wireless transceiver device in each rack from the top of the rack, the total number of network layers, and the maximum height at which the antenna horn is deployed, determine the relative height of the wireless transceiver device from the top layer of each rack.

[0071] Take the difference between any layer number sequence of the wireless transceiver device from the top of the rack and the value 1 as the numerator of the ratio, and take the difference between the total number of network layers and the value 1 as the denominator of the ratio. Calculate the product of the ratio and the maximum height at which the antenna horn is deployed, and use the product as the relative height of the wireless transceiver device from the top layer of the rack corresponding to the layer number sequence.

[0072] In this embodiment, the calculation formula for the relative height of the wireless transceiver device from the top layer of each rack can be:

[0073]

[0074] where h k is the relative height of the wireless transceiver device from the k-th layer at the top of the rack, k is the layer number sequence of the wireless transceiver device in each rack from the top of the rack, the layer number sequence k is greater than 0 and less than the total number of network layers (0 < k < w), w is the total number of network layers, and the total number of network layers is a known value, and h is the maximum height at which the antenna horn is deployed.

[0075] S4, according to the hierarchical deployment of each wireless transceiver in each network unit, the number of row racks, the number of column racks, and the wireless transceivers above all racks in each network unit, the hierarchical deployment is a cube unit, construct the position of each wireless transceiver in each network unit in the cube, and divide each wireless transceiver in each network unit into vertex nodes, surface nodes, and body nodes based on the position in the cube.

[0076] It should be noted that the 60GHz wireless transceivers above each rack in each network unit are layered and deployed as completely identical cubic units, and any two wireless transceivers in the cubic unit can be within the wireless communication range. A maximum of two wireless transceivers are in a straight line in a cubic unit. If there are three wireless transceivers in a straight line, when the first wireless transceiver wants to communicate with the third transceiver, the second wireless transceiver will be blocked, and normal communication will not be possible. Therefore, in order to avoid signal obstruction, any three wireless transceivers in the cubic unit are not in a straight line. Two wireless transceivers on the same rack are located at different layers and can participate in communication, so there will be one or two 60GHz wireless transceivers above each rack, and the network unit and the cubic unit are in a one-to-one relationship.

[0077] In this embodiment, based on the number of row racks and column racks of each network unit, the racks in each network unit are arranged, that is, the wireless transceivers are arranged, and the positions of the wireless transceivers in each network unit in the cube can be determined. The wireless transceivers here are nodes. The principle diagram of the three-dimensional coordinate system corresponding to the basic unit structure of the cube layered network is as follows: Figure 3 As shown. Figure 3 In the figure, there are nodes at the 8 vertices of the 5-layer CaF structure, 4 nodes in each face, and 8 nodes inside. The 8 nodes are located at the center of the 8 directions. Based on the position of each node in the cube, each wireless transceiver in each network unit is divided into vertex nodes, in-face nodes, and body nodes. The nodes at the vertex position of the cube are vertex nodes, the nodes on the surface of the cube are in-face nodes, and the nodes inside the cube are body nodes. Figure 3 shown.

[0078] S5, designing a corresponding VEPL routing algorithm according to the CaF characteristics, wherein the VEPL routing algorithm includes a vertex routing algorithm and an intra-plane routing algorithm.

[0079] First, the vertex routing algorithm is used, and the source node transmits the wireless signal to the vertex node closest to the destination. If the vertex node is in an occupied or faulty state, the intra-plane routing algorithm is used to forward it to the intra-plane node closest to the destination, and then the next hop is selected through the vertex routing algorithm. When the last forwarded node is one hop away from the destination, it is directly forwarded. The last forwarded node is a vertex node or an intra-plane node.

[0080] In this embodiment, based on the vertex nodes, in-plane nodes and body nodes in the data center network, the corresponding VEPL (Vertex and In-plane Routing Algorithm, vertex and in-plane routing algorithm) routing algorithm is designed using the CaF (the name of the cube-based topological structure in the self-defined wireless data center network) feature. The VEPL routing algorithm includes the vertex routing algorithm and the in-plane routing algorithm. When designing the corresponding VEPL routing algorithm, the vertex routing algorithm is used first. If the vertex node is in an occupied state or a fault state, the in-plane routing algorithm is used for routing communication. For example, the relative coordinates of the source node Src are (2,2), and the relative coordinates of the destination node Des are (12,11). First, the vertex routing algorithm is used to find the vertex node closest to the relative coordinates (12,11) of the destination. The relative coordinates of the nearest vertex node are (5,5). Then, the search for the next hop is continued. It is found that the relative coordinates (9,9) of the vertex node closest to the destination is occupied. Then, the intra-plane routing algorithm is used to find the intra-plane node closest to the destination. The relative coordinates of the intra-plane node are (9,8). Then, the vertex routing algorithm is used to find the vertex node closest to the destination. The relative coordinates of the nearest vertex node are (10,10), which are in the same basic unit as the relative coordinates (12,11) of the destination. It can be reached by direct one-hop forwarding. The final path is (2,2)→(5,5)→(9,8)→(10,10)→(12,11). The corresponding VEPL routing algorithm designed according to the CaF characteristics shortens the routing path length between nodes, can effectively deal with path failures, and significantly enhances the performance of the data center network.

[0081] Each cubic basic unit in this embodiment is composed of multiple wireless transceivers of different heights that can perform 360-degree omnidirectional data transmission. Based on the network scale of the network unit, the relative coordinates in the rack array, and the relative heights of the wireless transceivers, the actual deployment of the CaF three-dimensional hierarchical structure in the data center is completed. The schematic diagram of each wireless transceiver in the actual hierarchical deployment is shown in FIG. Figure 4 As shown. Figure 4 It can be seen that the wireless transceiver rod is thin and bendable enough and will not block the wireless signal.

[0082] The present invention provides a data center network topology construction and routing design method based on a cube structure. The method uses 60GHz wireless communication technology to solve the deployment difficulties, complex wiring and easy generation of hot node defects in traditional wired data centers. It can also ensure high data transmission rates, greatly improving the performance of the data center. Placing a 60GHz millimeter wave wireless transceiver on the top of the rack can effectively avoid signal blocking, and directly establish wireless connections between nodes, so that the network has a larger average point degree and improves the connectivity of the network, thereby effectively shortening the routing path length between the sending node and the receiving node. By designing a corresponding routing algorithm, even if individual racks fail during forwarding, the normal operation of the entire network can be guaranteed.

[0083] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A data center network topology construction and routing design method based on cube structure, It is characterized in that The following steps are involved: Determine the number of row racks in each network unit according to the row spacing of the data center network and the preset network unit size, determine the number of column racks in each network unit according to the column spacing of the data center network and the preset network unit size, determine the position of the central rack based on the number of row racks and the number of column racks in each network unit, and delete the central rack in each network unit based on the position of the central rack; Determine the relative coordinates of each rack in each network unit according to the coordinates of each rack in each network unit, the row spacing, column spacing, number of row racks, and number of column racks of the data center network; According to the preset maximum transmission distance of the millimeter wave and the preset network unit size corresponding to the data center network, the maximum height at which the antenna horn is deployed is determined; according to the layer number of the wireless transceiver of each rack in each network unit from the top of the rack, the total number of network layers, and the maximum height at which the antenna horn is deployed, the relative height of the wireless transceiver from each layer on the top of the rack is determined; According to each wireless transceiver in each network unit, the number of row racks, the number of column racks, and the hierarchical deployment of the wireless transceivers above all racks in each network unit, the hierarchical deployment is a cube unit, constructing the position of each wireless transceiver in each network unit in the cube, and dividing each wireless transceiver in each network unit into vertex nodes, in-plane nodes, and in-body nodes based on the position in the cube; Design a corresponding VEPL routing algorithm according to the CaF characteristics, wherein the VEPL routing algorithm includes a vertex routing algorithm and an intra-plane routing algorithm; Design the corresponding VEPL routing algorithm according to the characteristics of CaF, including: CaF is the name of the cube-based topology in the self-defined wireless data center network. VEPL is Vertex and In-plane Routing Algorithm, which means the vertex and in-plane routing algorithm. First, the vertex routing algorithm is used, and the source node transmits the wireless signal to the vertex node closest to the destination. If the vertex node is in an occupied or faulty state, the intra-plane routing algorithm is used to forward it to the intra-plane node closest to the destination, and then the next hop is selected through the vertex routing algorithm. When the last forwarded node is one hop away from the destination, it is directly forwarded. The last forwarded node is a vertex node or an intra-plane node.

2. According to the method for designing a data center network topology and routing based on a cube structure according to claim 1, It is characterized in that The calculation formula for the number of row racks in each network unit is: Among them, m 0 is the number of row racks in each network unit, p is the row spacing of the data center network, a is the preset network unit size, For Round down; The calculation formula for the number of column racks in each network unit is: Among them, n 0 is the number of column racks in each network unit, q is the column spacing of the data center network, a is the preset network unit size, For Round down.

3. According to the method for designing a data center network topology based on a cube structure and routing, It is characterized in that Based on the number of row racks and column racks in each network unit, determine the location of the central rack, including: The ratio of the number of row racks to the value 2 is calculated, and the ratio is used as the abscissa of the central rack. The ratio of the number of column racks to the value 2 is calculated, and the ratio is used as the column coordinate of the central rack. The position of the central rack is determined based on the abscissa and column coordinates of the central rack.

4. According to the method for designing a data center network topology and routing based on a cube structure as described in claim 1, It is characterized in that According to the coordinates of each rack in each network unit, the row spacing, column spacing, number of row racks and number of column racks of the data center network, the relative coordinates of each rack in each network unit are determined, including: According to the horizontal coordinates of each rack in each network unit, the row spacing of the data center network, and the number of row racks, the relative horizontal coordinates of each rack in each network unit are determined. The calculation formula of the relative horizontal coordinates of each rack is: x=p(i-1)%(m 0 -1) Among them, x is the relative horizontal coordinate of each rack in each network unit, p is the row spacing of the data center network, i is the horizontal coordinate of each rack in each network unit, and m is the horizontal coordinate of each rack in each network unit. 0 is the number of row racks in each network unit, % is the remainder, p(i-1)%(m 0 -1) to find p(i-1) and (m 0 -1), p(i-1) and (m 0 -1) are all integers; According to the ordinates of each rack in each network unit, the column spacing of the data center network, and the number of column racks, the relative ordinates of each rack in each network unit are determined. The calculation formula of the relative ordinates of each rack is: y=q(j-1)=(n 0 -1) Among them, y is the relative vertical coordinate of each rack in each network unit, q is the column spacing of the data center network, j is the vertical coordinate of each rack in each network unit, and n 0 is the number of column racks in each network unit, % is the remainder, q(j-1)%(n 0 -1) to find q(j-1) and (n 0 -1), q(j-1) and (n 0 -1) are all integers.

5. According to the method for designing a data center network topology and routing based on a cube structure as described in claim 1, It is characterized in that According to the preset maximum transmission distance of the millimeter wave and the preset network unit size corresponding to the data center network, the maximum height at which the antenna horn is deployed is determined, including: Calculate the square of the maximum transmission distance of the preset millimeter wave and the square of the preset network unit size, and then calculate the product of the square of the preset network unit size and the value 2, so that the square of the maximum transmission distance of the preset millimeter wave is subtracted from the product of the square of the preset network unit size and the value 2, determine the difference, take the square root of the difference, and use the square root value as the maximum height at which the antenna horn is deployed.

6. According to the method for designing a data center network topology and routing based on a cube structure as described in claim 1, It is characterized in that According to the layer number of the wireless transceiver of each rack in each network unit from the top of the rack, the total number of network layers, and the maximum height at which the antenna horn is deployed, determine the relative height of the wireless transceiver from each layer of the top of the rack, including: The difference between any layer number of the wireless transceiver from the top of the rack and the value 1 is used as the numerator of the ratio, and the difference between the total number of network layers and the value 1 is used as the denominator of the ratio. The product of the ratio and the maximum height at which the antenna horn is deployed is calculated, and the product is used as the relative height of the wireless transceiver from the top layer of the rack corresponding to the layer number.

Citation Information

Cited By

  • Data transmission method and system based on multi-superframe structure of wireless cube clustering network

    CN120881688A