TWI684342B - Path routing method for hexagonal optical switching network - Google Patents

Abstract

This invention relates to a path routing method for hexagonal optical switching network, which is a high-speed optical switching network used in the Internet data center and uses a hexagonal optical switch as the switching core component. By conversion of the polar coordinates of each node configured in the hexagonal structure of the optical switch group, the cost function is used to calculate the minimum value of the path from starting point of the path, through the three adjacent nodes, to the end of the path in order to find the shortest optical path route between the starting point of the path and the end point of the path of the optical switch, and achieve the purpose of fast searching and obtaining the optical path transmission route.

Description

Path routing method for hexagonal optical switching network

The invention relates to a configuration method of a hexagonal optical switching network, in particular to a path routing method for a hexagonal optical switching network.

Existing optical switch routing technology, such as Republic of China Invention Patent No. I552536, is an optical data center network system and optical switch. The optical switch used is a commercial optical length switch (Wavelength Selective) Switch, WSS), and divide the network system of the optical data center into a three-layer architecture, which includes multiple first-layer optical switches, multiple second-layer optical switches, and multiple third-layer optical switches In which, a plurality of first-layer optical switches are connected to each other through a ribbon fiber to form a group; a plurality of second-layer optical switches are connected to each other through a ribbon fiber to form a giant group Group (macro pod), and each second-layer optical switch is connected to all the first-layer optical switches in the group, and multiple third-layer optical switches are connected to each other through a ribbon fiber, each third layer The optical switch is connected to all the second layer optical switches in the giant group to which it belongs.

In addition, SHORTEST PATH ROUTING SYSTEMS AND METHODS FOR NETWORKS WITH NON-FULLY of US Patent Publication No. 2013/0070617A1 MESHED VERTICES proposed that in a non-fully meshed network, the shortest path routing of two nodes, and the use of modified Dijkstra's algorithm to achieve, is a Greedy Method (Greedy Method) technology.

Existing optical switch routing configuration methods often calculate the optimal route at a speed that is not as expected, and cannot quickly search for the shortest path, or it is necessary to traverse all existing paths to find the best or shortest path, and the optical transmission route cannot be reached. The purpose of quick search and acquisition.

In view of the shortcomings of the aforementioned systems and methods, we are eager to innovate and invent. After a long period of painstaking and painstaking research, we finally successfully developed the shortest path routing method of the flexible fault-tolerant hexagonal switching network of the invention.

The invention proposes the shortest path routing method applied to the hexagonal optical switching switching network, which is used for the high-speed optical switching network of the Internet data center, uses the hexagonal optical switch as the switching core component, and uses the shortest path routing of the present invention Method to find the most suitable optical path between two nodes of the optical switch to achieve the purpose of rapid search and acquisition of optical path transmission routes.

Accordingly, the present invention provides a path routing method for a hexagonal optical switching network. The hexagonal optical switching network includes a plurality of optical switch groups with a hexagonal configuration in a honeycomb configuration, wherein the method The system includes: let a group of optical switch groups in the plurality of optical switch groups be the starting unit, and make the starting unit the first layer; let the plurality of optical switch groups centering around the starting unit be the first Two to N layers to calculate the maximum radius of the first to N layers; let the optical switch on the first direction extension line of the hexagonal structure of the starting unit be the starting point and along the first to N layers The optical switches in the second direction are numbered sequentially; the corresponding numbers and corresponding maximum radii of the path start point and the path end point are respectively obtained to calculate the polar coordinates of the path start point and the path end point; and the path start point and the path are input The polar coordinates of the end point are in the cost function to calculate starting from the path Point to the shortest path of the end of the path. The cost function calculates the minimum cost value from the starting point of the path through the adjacent three nodes and from the adjacent three nodes to the end of the path, respectively.

As in the aforementioned path routing method, the cost function is calculated based on the sufficient optical path bandwidth resources, the remaining part of the optical path bandwidth resources, and the exhaustion of all optical path bandwidth resources between the path start point and the three adjacent nodes Value.

As in the aforementioned path routing method, the cost function calculates the Euclidean distance or the Hamilton distance between the adjacent three nodes and the end of the path, respectively.

As in the aforementioned path routing method, it further includes checking the termination condition to end the cost function. The step of checking the termination condition is to check whether the available nodes in the three adjacent nodes are the end of the path; if so, the cost function is ended, otherwise, the cost function continues to be calculated.

As in the aforementioned path routing method, it further includes checking the termination condition to end the cost function. The step of checking the termination condition is to check whether the available nodes are blocked in the process of the adjacent three nodes and backtrace to the previous path whether the path is the starting point; if so, end the cost function, otherwise Then continue to calculate the cost function.

As in the aforementioned path routing method, it further includes checking the termination condition to end the cost function. The step of checking the termination condition is to check whether the previous node is the starting point of the path, or the current node is the starting point of the path; if so, output a corresponding message and temporarily buffer the end of the path to the waiting queue of the beginning of the path (waiting queue), and end the cost function, otherwise, backtrace (Backtrace) to the previous node, and execute the cost function again.

，其中，該第N層的該最大半徑設為R_N，第N層設為N，且N為大於或等於1的整數。 As in the aforementioned path routing method, in the step of calculating the maximum radius of the first to N layers, the equation for calculating the maximum radius is:

, Where the maximum radius of the Nth layer is set to R _N and the Nth layer is set to N, and N is an integer greater than or equal to 1.

As in the aforementioned path routing method, the first direction is the nine o'clock direction of the hexagonal structure of the initiating unit, and the second direction is the clockwise outer edge of the combination of the hexagonal structure of the Nth layer direction.

，

，其中，該第N層的該最大半徑設為R_N，該第N層設為N，由該起始機組中該第一方向之延伸線上的該起始點為第零個點起算且沿著該第N層的該六角形架構之組合外緣之第二方向上的該光交換機依序編號的第M+1個點設為M。 As in the aforementioned path routing method, in the step of calculating the polar coordinate values of the start point and the end point of the path, the equation used to calculate the polar coordinate values is:

,

, Where the maximum radius of the Nth layer is set to R _N and the Nth layer is set to N, starting from the zeroth point and starting from the zeroth point on the extension line of the first direction in the starting unit The M+1th point of the sequentially numbered optical switches in the second direction of the combined outer edge of the hexagonal structure of the Nth layer is set as M.

The present invention is in the field of all-optical switching technology research, through the hexagonal optical switching network to achieve the fastest and simplest shortest path routing, and the present invention by finding the most suitable optical path between two optical switch nodes on the network , To achieve the purpose of rapid search and acquisition of optical transmission routes.

The shortest path routing method of the present invention is used to find the most suitable optical path between two nodes of the optical switch. It belongs to the Heuristic Algorithm of AI, which has the ability to quickly search, and the best case time complexity is O( log n ), the worst case time complexity is O(n×log n ), and the general case time complexity is O(log n )~O(n×log n ), making the average time complexity O(n) .

The design of the hexagonal optical switch network used in the present invention is different from the traditional three-layer switching architecture of the data center or the Spine-Leaf two-layer switching architecture. The present invention uses the current optical switch as the switching core and uses the hexagonal The two-dimensional deployment method is used to expand the entire all-optical switching network. The shortest path routing method is proposed by the shortest path routing method described above to find the most suitable optical path routing between two nodes of the optical switch.

10‧‧‧ Optical switch group

11‧‧‧ Starting unit

12‧‧‧ Starting point

13‧‧‧End of route

L1‧‧‧First floor

L2‧‧‧Second floor

L3‧‧‧The third floor

L4‧‧‧Fourth floor

S01~S07‧‧‧Step

FIG. 1 is a flow chart of the steps of the path routing method for a hexagonal optical switching network of the present invention.

Figure 2 is a schematic diagram of the configuration of the hexagonal structure of the optical switchboard of the present invention.

FIG. 3 is a schematic diagram of implementing the path routing method of the present invention in an embodiment of a seven-layer optical switch group.

The present invention relates to a path routing method for a hexagonal optical switching network, which is a hexagonal high-speed optical switching network for an Internet data center. As shown in Figures 2 and 3, the present invention uses an optical switch To exchange core components, six optical switches are configured as hexagonal optical switch groups 10, and each optical switch group 10 is arranged in a honeycomb shape.

In the following, with reference to Figures 2 and 3, a flowchart of the steps of the path routing method for the hexagonal optical switching network in Figure 1 will be described.

In step S01, a group of optical switch groups in the plurality of optical switch groups 10 is selected as the starting unit 11, and the starting unit 11 is the first layer L1, that is, the six of the starting unit 11 The optical switch on each end of the angular architecture is located on the first layer L1.

In step S02, with the initial unit 11 as the center, the plurality of optical switch groups 10 surrounding the initial unit 11 are sequentially from the second layer L2 to the Nth layer; that is, directly connected to the first layer L1 The plurality of optical switch groups 10 (a total of six groups) of the starting unit 11 is the second layer L2, and the plurality of optical switch groups 10 (a total of twelve groups) directly connected to the second layer L2 is the third Layer L3, and so on to the Nth layer (the seventh layer L7 in the example in Figure 2, but not limited to this).

，其中，R_N為該第N層的該最大半徑，N為第N層，且N為大於或等於1的整數。 In step S03, the maximum radius of the first to N layers is calculated, where the equation for calculating the maximum radius is:

, Where R _N is the maximum radius of the Nth layer, N is the Nth layer, and N is an integer greater than or equal to 1.

Taking the first layer L1 as an example, the maximum radius R ₁ of the first layer L1 is 1.

。 Taking the second layer L2 as an example, the maximum radius R ₂ of the second layer L2 is

.

。 Taking the third layer L3 as an example, the maximum radius R ₃ of the third layer L3 is

.

。 Taking the fourth layer L4 as an example, the maximum radius R ₄ of the fourth layer L4 is

.

In step S04, the optical switch on the extension line in the first direction of the hexagonal structure of the starting unit 11 is used as the starting point (ie, node (1,0)) and along the second direction of the Nth layer The optical switches are numbered sequentially. The first direction is the nine o'clock direction of the hexagonal structure of the starting unit 11, and the second direction is the clockwise direction of the combined outer edge of the hexagonal structure of the Nth layer. The starting point of each layer (a total of N layers) on the extension line of the hexagonal structure of the starting unit 11 is set to the zeroth point (N, 0), and the six along the Nth layer The outer edge of the angular structure is numbered sequentially in the clockwise direction, points 1, 2, 3, 4....M points (that is, numbered (N, 1), (N, 2), (N, 3), (N ,4)...(N,M)).

Please refer to Figures 2 and 3 at the same time. The first node number of the starting unit 11 (first layer L1) is (1,0); the node (4, 2) is from the starting unit 11 The optical switch of the fourth layer L4 on the extension line of the nine o'clock direction is regarded as the zero point, and the second point in the clockwise direction along the outer edge of the fourth layer L4 (that is, the third point including the zero point ); and so on, the node (3,14) is the The optical switch of the third layer L3 on the extension line at the nine o'clock direction serves as the zeroth point of the starting point, and the fourteenth point in the clockwise direction along the outer edge of the third layer L3 (that is, including the zeroth point Fifteen points).

In step S05, the corresponding numbers and corresponding maximum radii of the path start point 12 (Source Node) and the path end point 13 (Destination Node) are respectively obtained to calculate the polar coordinates of the path start point 12 and the path end point 13.

，其中，R為半徑，N為該第N層，M為編號。 The representation of polar coordinates is (X,Y), X = R cos( θ ), Y = R sin( θ ), θ =

, Where R is the radius, N is the Nth layer, and M is the number.

，

，其中，R_N為該第N層的該最大半徑，N為該第N層，M為編號。 Therefore, the polar coordinate value can be combined and expressed as:

,

, Where R _N is the maximum radius of the Nth layer, N is the Nth layer, and M is the number.

If the aforementioned node (1,0) is used, its polar coordinate value (X _S , Y _S ) is (0, 0).

,

,

)。 If the node (4, 2) is taken as the starting point of the path 12, the polar coordinate value (X _S , Y _S ) of the starting point 12 of the path is (

,

,

).

,

,

)。 If the node (3, 14) is taken as the end point 13 of the path, the polar coordinates (X _d , Y _d ) of the end point 13 of the path are (

,

,

).

Next, proceed to step S06, input the polar coordinates of the path start point 12 and the path end point 13 in a cost function to calculate and record the shortest path from the path start point 12 to the path end point 13; wherein, the The cost function calculates the minimum cost value from the starting point 12 of the path through the adjacent three nodes and the adjacent three nodes to the end point 13 of the path.

An example of the cost function is as follows. Select the path start point 12 (Source Node) in the hexagonal optical switching network, take the path start point (N, M) as an example, and select the path end point 13 (Destination Node) in the hexagonal optical switching network. Taking the end point of the path (T, U) as an example, starting from the start point of the path (N, M), the next point N _i of the best routing path is selected, where N _i is the adjacent 1 to the starting point 12 (N, M) of the path Three optical switch nodes, which will be calculated as: cost (start of path (N, M), next node N ₁ ) + cost (N ₁ , end of path (T, U)); cost (start of path (N, M), the next node N ₂ )+cost(N ₂ , path end (T,U)); and cost(path start (N,M), next point N ₃ )+cost(N ₃ , path end (T ,U)).

Find the minimum cost value by the cost function, which can be written as: cost function cost (the starting point of the path, the path end point) = min {cost (the starting point of the _{path, N i) + cost (Ni} , end of the path)}, i = 1 ~ 3 .

The value of the cost function cost (start point of the path, end point of the path) is the value from the start point 12 of the path through the adjacent 3 nodes N ₁ , N ₂ , and N ₃ to the end point 13 of the path, and the minimum value is taken as the cost function cost (start point of the path) , The end of the path), and record the relevant selection and optical path occupation.

cost (the starting point of the path, N _i) of the values defined as follows: If the light path with sufficient bandwidth resource path between the starting point 12 to the point N _i, the cost (the starting point of the path, Ni) = 1; if the path 12 to the lower starting point The bandwidth of the optical path of the remaining resources between N _i , then cost(start of path, Ni) = 10~100; and if the bandwidth of the optical path has exhausted all resources between the path starting point 12 and the next point N _i , then cost( Starting point of the path, Ni)=∞.

cost(N _i , the end of the path) is equal to the Euclidean distance (N _i , the end of the path), or the Hamilton distance (N _i , the end of the path). The Euclidean distance is a straight line between two points calculated using polar coordinates The distance, and the Hamilton distance is the total number of side lengths required to be calculated along the side length of the node, which can be used in the network topology of cyclic nodes.

，節點(4,2)的X座標為

，Y 座標為

；節點(3,14)的R₃為

，節點(3,14)的X座標為

，Y座標為

，由前述的節點(4,2)及節點(3,14)的X、Y座標即可算出兩者之間的歐幾里得距離。 As shown in Figure 3, the distance from node (4,2) to node (3,14) is 11; as mentioned above, R ₄ of node (4,2) is

, The X coordinate of node (4,2) is

, The Y coordinate is

; R ₃ of node (3,14) is

, The X coordinate of node (3,14) is

, The Y coordinate is

, The Euclidean distance between the two nodes (4, 2) and the nodes (3, 14) can be calculated from the X and Y coordinates.

Then, proceed to step S07, and check the termination condition to exit the cost function.

Check whether the next node N _i of the minimum cost value of the cost function is equal to the end of the path 13, if it is, then output the corresponding message to find the shortest path route, output each node and light path from the path start 12 to the path end 13 Record the status and end the cost function; if not, continue the following inspection steps.

Check whether the path starting point 12 to the adjacent three nodes N _i is blocked (blocking) state, if it is, then output the corresponding message that this node can not pass, record the selected state of this node as selected and blocked Status, and backtrace (Backtrace) to return to the previous node of the path to continue searching from the possible second shortest path or the third path.

Next, check whether the previous node is the path start point 12, or the current node is the path start point 12, if it is, the corresponding message that the best path cannot be found is output, and the path end point 13 is temporarily stored (buffer) to the waiting queue at the starting point of the path 12 to wait for the next free optical path to be released and the optical path state changes before processing, then record the optical path usage status, and end the cost function; if not , Then backtrack to return to the previous node of the path, and execute the cost function again, that is, execute the cost function cost (previous path start point, path end point).

If none of the above conditions, move forward a node, continue to find, and record the shortest path routing state, then the path starting point 12 to node Ni is selected; then execute the cost function cost (Ni, path end point) again, until the end If the condition is true, the cost function is ended.

The shortest path routing method for hexagonal optical switching network of the present invention is used to find the most suitable optical path between two nodes of an optical switch to achieve the purpose of fast search and acquisition of optical path transmission routes. It has the following advantages in other conventional technologies:

The shortest path routing method of the present invention has the ability to quickly search, if in the best case, that is, when the entire optical path is free, the shortest straight line distance (air distance) can be found from three adjacent nodes along the optical path to reach the path End point 13, assuming that n is the number of nodes, in a two-dimensional hexagonal optical switching network, it can be regarded as a perfect circular network, the distance between any two points is less than or equal to the diameter of this circle, so the best case It is to find the end point all the way, the number of nodes walking the best path will be less than or equal to the number of nodes of the diameter of this circle, so the time complexity is O(log n ), which is the best and fastest search speed. If in the worst case, the node that walked to the end encounters the blocking state, and then retreats to find other paths and all the nodes along the way encounter the blocking state until it returns to the beginning of the path 12, so the worst situation It may be that n nodes have been found, so the time complexity is O(n×log n ).

Therefore, the shortest path routing method of the present invention generally has a time complexity of O(log n )~O(n×log n ) under the general condition of “partially advancing, partly retreating and then advancing”, and its average time complexity is O(n) means that in general, the present invention has a faster search capability.

On the cost function of cost (Ni, the end of the path), the present invention uses Euclidean distance or Hammillun distance for implementation. Euclidean distance uses polar coordinates to calculate the straight-line aviation distance between two points; The distance is calculated along the side length of the node to calculate the total number of side lengths traversed. It is more suitable for practical use in the network topology of cyclic nodes.

The above detailed description is a specific description of a feasible embodiment of the present invention, but this embodiment is not intended to limit the patent scope of the present invention, and any equivalent implementation or change without departing from the technical spirit of the present invention should be included in The patent scope of this case.

To sum up, this case is not only innovative in terms of technical ideas, but also possesses the above-mentioned multiple functions that are not achievable by traditional methods, and has fully met the novelty and progressive legal invention patent requirements File, you file an application in accordance with the law, and urge your office to approve this application for a patent for invention to encourage the invention and reach a sense of virtue.

10‧‧‧ Optical switch group

11‧‧‧ Starting unit

12‧‧‧ Starting point

13‧‧‧End of route

L1‧‧‧First floor

L2‧‧‧Second floor

L3‧‧‧The third floor

L4‧‧‧Fourth floor

Claims (9)

A path routing method for a hexagonal optical switching network, wherein the hexagonal optical switching network includes a plurality of optical switch groups with a hexagonal configuration in a honeycomb configuration, and the method includes: making the multiple A group of optical switch groups in the optical switch group is used as the initial unit, and the initial unit is the first layer; the multiple optical switch groups around the initial unit are the second to N layers to calculate The maximum radius of the first to N layers; let the optical switch on a line extending in a first direction in the hexagonal structure of the starting unit be the starting point, along the second direction of the first to N layers The optical switches on the system are numbered sequentially; the corresponding numbers and corresponding maximum radii of the path start point and the path end point are respectively obtained to calculate the polar coordinates of the path start point and the path end point; and the polar coordinates of the path start point and the path end point are input The value is in the cost function to calculate the shortest path from the start of the path to the end of the path; where the cost function calculates the path from the start of the path through the adjacent three nodes and from the adjacent nodes to the end of the path The minimum generation value. The path routing method as described in item 1 of the patent application scope, wherein the cost function is based on the sufficient optical path bandwidth resources, the remaining part of the optical path bandwidth resources, and the consumption between the path start point and the three adjacent nodes Use all optical bandwidth resources to calculate the value. The path routing method as described in item 1 of the patent application scope, wherein the cost function calculates the Euclidean distance or the Hamilton distance between the adjacent three nodes and the end of the path, respectively. The route routing method as described in item 1 of the patent application scope further includes: Check the termination condition to end the cost function; wherein, the step of checking the termination condition is to check whether the available nodes in the adjacent three nodes are the end of the path; if it is, then end the cost function, otherwise, continue to calculate the cost function. The path routing method as described in item 1 of the patent application scope further includes: checking the termination condition to end the cost function; wherein, the step of checking the termination condition is to obtain the available node during the process of checking the adjacent three nodes as In the congested state and back to the previous path is the starting point of the path; if it is, the cost function is ended, otherwise, the cost function continues to be calculated. The path routing method as described in item 1 of the patent application scope further includes: checking the termination condition to end the cost function; wherein, the step of checking the termination condition is to check whether the previous node is the starting point of the path or the current node Is the starting point of the path; if it is, it outputs a corresponding message, temporarily stores the end point of the path in the waiting queue of the starting point of the path, and ends the cost function, otherwise, it goes back to the previous node and executes the cost function again. The path routing method as described in item 1 of the patent application scope, wherein in the step of calculating the maximum radius of the Nth layer, the equation for calculating the maximum radius is:

, Where the maximum radius of the Nth layer is set to R _N and the Nth layer is set to N, and N is an integer greater than or equal to 1. The path routing method as described in item 1 of the patent scope, wherein the first direction is the nine o'clock direction of the hexagonal structure of the starting unit, and the second direction is the hexagonal shape of the Nth layer The clockwise direction of the outer edge of the combined structure. The path routing method as described in item 1 of the patent application scope, wherein, in the step of calculating the polar coordinate values of the start point and the end point of the path, the equation for calculating the polar coordinate values is:

,

, Where the maximum radius of the Nth layer is set to R _N and the Nth layer is set to N, starting from the zeroth point and starting from the zeroth point on the extension line of the first direction in the starting unit The M+1th point of the sequentially numbered optical switches in the second direction of the combined outer edge of the hexagonal structure of the Nth layer is set as M.

Images