Hyper split ab tree progressive switching implementation method

Abstract

The application relates to a HyperSplit AB tree progressive switching implementation method and belongs to the technical field of communication.The application implements a HyperSplit AB tree progressive switching implementation method through the following steps: designing a cutting and identifying method of a subspace range in the HyperSplit AB tree progressive switching implementation mode, designing a judgment algorithm of subspace intersection in the HyperSplit AB tree progressive switching implementation mode, and designing an algorithm of progressively selecting and loading a subspace sequence in the HyperSplit AB tree progressive switching implementation mode.Compared with an existing HyperSplit AB tree direct switching implementation mode, the HyperSplit AB tree progressive switching implementation method can greatly save the storage space occupied in the HyperSplit AB tree switching process.Under the condition that the storage resource is limited, the HyperSplit AB tree switching can still be normally completed.Moreover, the implementation mode can enable more performance optimization methods to be implemented, thereby improving the overall performance of service processing.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, specifically relating to a method for implementing progressive switching of HyperSplit AB trees. Background Technology

[0002] The existing technical solutions for directly switching implementation methods using HyperSplit AB trees are as follows: Figure 1 As shown, this implementation method involves loading the constructed HyperSplit A-tree and HyperSplit B-tree data into the storage space all at once, directly completing the switch between the HyperSplit A-tree and HyperSplit B-tree. The processing method for the A-tree is as follows: First, based on the rules contained in the A-tree, the complete range of the rule space is divided into various A-tree subspaces (Space(Tree A)). Then, the HyperSplit A-tree is constructed based on the divided A-tree subspaces (Space(Tree A)). Finally, all A-tree subspaces (Space(Tree A)) are stored in the storage space (Mem Block). The processing method for the B-tree is as follows: First, based on the rules contained in the B-tree, the complete range of the rule space is divided into various B-tree subspaces (Space(Tree B)). Then, the HyperSplit B-tree is constructed based on the divided B-tree subspaces (Space(Tree B)). Finally, all B-tree subspaces (Space(Tree B)) are stored in the storage space (Mem Block).

[0003] As can be seen, the current implementation of direct switching between HyperSplit AB trees involves loading the constructed HyperSplit A-tree and HyperSplit B-tree data into storage all at once, directly completing the switch between the two. Because this implementation loads all sub-spaces of both the HyperSplit A-tree and HyperSplit B-tree simultaneously, the required storage space can double or even more. In applications with limited storage resources, this implementation is unusable and may prevent the implementation of performance optimization methods due to excessive storage space consumption. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] The technical problem to be solved by this invention is: how to design an implementation method that can significantly reduce the storage space occupied during the HyperSplit AB tree switching process.

[0006] (II) Technical Solution

[0007] To address the aforementioned technical problems, this invention provides a method for implementing progressive switching of a HyperSplit AB tree, comprising the following steps:

[0008] The first step is to divide the complete range of the rules into various subspaces (Tree A) based on the rules contained in the HyperSplit A-tree, set the status of all subspaces (Tree A) to valid, construct the HyperSplit A-tree, and finally store all subspaces (Tree A) of the A-tree in the storage space.

[0009] The second step is to divide the complete range of the rules into various B-tree subspaces (Tree B) according to the rules contained in the HyperSplit B-tree, set the state of all B-tree subspaces (Tree B) to unloaded, and then construct the HyperSplit B-tree.

[0010] The third step is to obtain all the split subspaces that are valid in the HyperSplit A tree;

[0011] Step 4: Determine if there is a valid subspace in the HyperSplit A tree. If it exists, proceed to step 5 and continue execution. If it does not exist, end the current switching process.

[0012] Step 5: Compare all the partitioned and valid subspaces in tree A with all the partitioned and unloaded subspaces in tree B; based on the comparison results, select the subspace Space'(Tree A) from tree A that has the smallest number of intersections with subspace Space(Tree B) in tree B.

[0013] Step 6: Find all B-tree subspaces (Tree B) that intersect with the selected A-tree subspace Space' (Tree A) in the previous step; then load the data of the found B-tree subspaces (Tree B) into the storage space and set their status to loaded; finally, set the status of the A-tree subspace Space' (Tree A) to invalid; then jump to step 3 to continue execution.

[0014] Preferably, in the first step, five HyperSplit A-tree subspaces are segmented according to the rules contained in the HyperSplit A-tree, namely HyperSplit A-tree subspace 1 Tree A Space 1, HyperSplit A-tree subspace 2 Tree A Space 2, HyperSplit A-tree subspace 3 Tree A Space 3, HyperSplit A-tree subspace 4 Tree A Space 4, and HyperSplit A-tree subspace 5 Tree A Space 5.

[0015] Preferably, in the second step, five HyperSplit B-tree subspaces are created based on the rules contained in the HyperSplit B-tree, namely HyperSplit B-tree subspace 1 (Tree B Space 1), HyperSplit B-tree subspace 2 (Tree B Space 2), HyperSplit B-tree subspace 3 (Tree B Space 3), HyperSplit B-tree subspace 4 (Tree B Space 4), and HyperSplit B-tree subspace 5 (Tree B Space 5).

[0016] Preferably, the HyperSplit A tree contains 4 internal nodes and 5 leaf nodes, where the 5 leaf nodes correspond to the 5 subspaces of the HyperSplit A tree; the HyperSplit B tree contains 4 internal nodes and 5 leaf nodes, where the 5 leaf nodes correspond to the 5 subspaces of the HyperSplit B tree.

[0017] Preferably, in the first step, the five subspaces contained in the HyperSplit A tree, Tree A Space 1, Tree A Space 2, Tree A Space 3, Tree A Space 4, and Tree A Space 5, have been stored sequentially into storage unit blocks 1 to 5, namely Mem Block 1, Mem Block 2, Mem Block 3, Mem Block 4, and Mem Block 5;

[0018] During the first loop, in step five, the subspace with the fewest intersections with the HyperSplit B tree subspaces is selected from the five subspaces contained in the HyperSplit A tree. Let's say that the selected subspace is HyperSplit A tree subspace 1Tree A Space 1, which has 1 intersection with the HyperSplit B tree subspace. During the first loop, in step six, HyperSplit B tree subspace 1Tree B Space 1, which intersects with HyperSplit A tree subspace 1Tree A Space 1, is stored in storage unit block 6 (Mem Block 6). Finally, storage unit block 1 containing HyperSplit A tree subspace 1Tree A Space 1 is set to invalid. This completes the switching operation of HyperSplit A tree subspace 1Tree A Space 1.

[0019] Preferably, during the fifth step of the second loop, the subspace with the fewest intersections with the remaining four subspaces of the HyperSplit A tree is selected from the remaining four subspaces of the HyperSplit B tree. In this case, the selected subspace is HyperSplit A tree subspace 2Tree A Space 2, which has 1 intersection with the HyperSplit B tree subspace. During the sixth step of the second loop, HyperSplit B tree subspace 2Tree B Space, which intersects with the HyperSplit A tree subspace from the remaining four subspaces of the HyperSplit B tree, is stored in storage unit block Mem Block 1. Finally, storage unit block 2, where HyperSplit A tree subspace Tree A Space 2 is located, is set to invalid. This completes the switching operation of HyperSplit A tree subspace 2Tree A Space 2.

[0020] Preferably, when performing step 5 in the third loop, firstly, select the subspace with the fewest intersections with the remaining 3 subspaces of the HyperSplit A tree from the remaining 3 subspaces contained in the HyperSplit A tree. In this case, the selected subspace is HyperSplit A tree subspace 3Tree A Space 3, which has 1 intersection with the HyperSplit B tree subspace. When performing step 6 in the third loop, store HyperSplit B tree subspace 3Tree B Space 3, which intersects with HyperSplit A tree subspace 3Tree A Space 3, into storage unit block 2MemBlock 2. Finally, set storage unit block 3 where HyperSplit A tree subspace 3Tree A Space 3 is located to invalid. This completes the switching operation of HyperSplit A tree subspace 3Tree A Space 3.

[0021] Preferably, when performing the fifth step in the fourth loop, firstly, select the subspace with the fewest intersections with the remaining two subspaces of the HyperSplit A tree from the remaining two subspaces contained in the HyperSplit A tree. In this case, the selected subspace is HyperSplit A tree subspace 4Tree A Space 4, which has 1 intersection with the HyperSplit B tree subspace. When performing the sixth step in the fourth loop, store HyperSplit B tree subspace 4Tree B Space 4, which intersects with HyperSplit A tree subspace 4Tree A Space 4, into storage unit block 3 (Mem Block 3). Finally, set storage unit block 4, where HyperSplit A tree subspace 4Tree A Space 4 is located, to invalidate. This completes the switching operation of HyperSplit A tree subspace 4Tree A Space 4.

[0022] Preferably, during the sixth step of the fifth loop, the last subspace of the HyperSplit B tree, 5Tree B Space 5, which intersects with the last subspace of the HyperSplit A tree, is stored in storage unit block 4Mem Block 4. Finally, storage unit block 5 containing the HyperSplit A tree subspace 5Tree A Space 5 is set to invalid, thus completing the gradual switch from HyperSplit A tree to HyperSplit B tree.

[0023] The present invention also provides a method for saving storage space during HyperSplit AB tree switching using the method described above.

[0024] (III) Beneficial Effects

[0025] This invention implements a progressive switching method for HyperSplit AB trees by designing methods for cutting and identifying subspace ranges, determining subspace intersections, and progressively selecting the loading order of subspaces. Compared to existing direct switching methods, this invention significantly reduces storage space usage during HyperSplit AB tree switching. It can complete HyperSplit AB tree switching even with limited storage resources. Furthermore, this implementation allows for the implementation of more performance optimization methods, thereby improving the overall performance of business processing. Attached Figure Description

[0026] Figure 1 Overall architecture diagram for HyperSplit AB tree direct switching implementation method;

[0027] Figure 2 This is a flowchart illustrating the implementation method of the HyperSplit AB tree progressive switching of the present invention.

[0028] Figure 3 This is an example diagram illustrating the processing procedure of the HyperSplit AB tree progressive switching implementation method of the present invention. Detailed Implementation

[0029] To make the objectives, contents, and advantages of the present invention clearer, the specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples.

[0030] This invention provides a method for progressively switching HyperSplit AB trees, with the primary objective of designing a method that significantly reduces storage space usage during HyperSplit AB tree switching. This method can successfully complete the HyperSplit AB tree switching even under conditions of limited storage resources. Furthermore, this implementation method allows for the implementation of more performance optimization techniques, thereby improving the overall performance of business processing.

[0031] refer to Figure 2 The overall solution for the HyperSplit AB tree progressive switching implementation method provided by this invention is as follows:

[0032] Step 1: First, based on the rules contained in the A-tree, divide the complete range of the rules into various A-tree subspaces Space (Tree A). Then, construct the HyperSplit A-tree based on the divided A-tree subspaces Space (Tree A). Finally, store the data of all A-tree subspaces Space (Tree A) into the storage space Mem Block.

[0033] Step 2: Process the B-tree as follows: First, based on the rules contained in the B-tree, divide the complete range of the rule space into various B-tree subspaces (Tree B). Then, construct the HyperSplit B-tree based on the divided B-tree subspaces (Tree B). After constructing the B-tree, do not directly store all the data of the B-tree subspaces (Tree B) into the storage space (Mem Block). Instead, compare all the subspaces (Tree A) contained in the A-tree with all the subspaces (Tree B) contained in the B-tree. Based on the comparison result, select the subspace (Tree A) from the A-tree with the smallest number of intersections with the B-tree subspace (Tree B). Find all the B-tree subspaces (Tree B) that intersect with the selected A-tree subspace (Tree A) from the previous step. Finally, load the found B-tree subspace (Tree B) data into the storage space (Mem Block). Repeat the above steps until all the B-tree subspace (Tree B) data is loaded into the storage space (Mem Block).

[0034] The specific implementation process of the HyperSplit AB tree progressive switching is as follows: Figure 3As shown, this mainly includes an example diagram of a HyperSplit A-tree subspace partitioned according to the HyperSplit A-tree rule set (Space decomposition by HyperSplit Tree A). In this example diagram, the X-axis represents the range of source IP addresses (SIP) in the rules, with a maximum value of 2^32, and the Y-axis represents the range of destination IP addresses (DIP) in the rules, with a maximum value of 2^32. It also includes an example diagram of a HyperSplit B-tree subspace partitioned according to the HyperSplit B-tree rule set (Space decomposition by HyperSplit Tree B). In this example diagram, the X-axis represents the range of source IP addresses (SIP) in the rules, with a maximum value of 2^32, and the Y-axis represents the range of destination IP addresses (DIP) in the rules, with a maximum value of 2^32. Additionally, it includes the HyperSplit A-tree (HyperSplit Tree A) constructed from the partitioned HyperSplit A-tree subspaces, the HyperSplit B-tree (HyperSplit Tree B) constructed from the partitioned HyperSplit B-tree subspaces, and the storage of HyperSplit trees. The storage unit blocks (Mem Blocks) for data information in the A tree subspace and the HyperSplit B tree subspace.

[0035] Based on the rules contained in the HyperSplit A-tree, five HyperSplit A-tree subspaces were created: HyperSplit A-tree subspace 1, HyperSplit A-tree subspace 2, HyperSplit A-tree subspace 3, HyperSplit A-tree subspace 4, and HyperSplit A-tree subspace 5.

[0036] HyperSplit A-tree subspace 1 contains source IP address ranges of [0-2] and destination IP address ranges of [0-1]; HyperSplit A-tree subspace 2 contains source IP address ranges of [0-2] and destination IP address ranges of [1-2^32]; HyperSplit A-tree subspace 3 contains source IP address ranges of [2-3] and destination IP address ranges of [0-2^32]; HyperSplit A-tree subspace 4 contains source IP address ranges of [3-2^32] and destination IP address ranges of [3-2^32]; HyperSplit A-tree subspace 5 contains source IP address ranges of [3-2^32] and destination IP address ranges of [0-3].

[0037] Based on the rules contained in the HyperSplit B-tree, five HyperSplit B-tree subspaces were created: HyperSplit B-tree subspace 1, HyperSplit B-tree subspace 2, HyperSplit B-tree subspace 3, HyperSplit B-tree subspace 4, and HyperSplit B-tree subspace 5.

[0038] HyperSplit B-tree subspace 1 contains source IP address ranges of [0-2] and destination IP address ranges of [0-2]; HyperSplit B-tree subspace 2 contains source IP address ranges of [0-2] and destination IP address ranges of [2-2^32]; HyperSplit B-tree subspace 3 contains source IP address ranges of [2-3] and destination IP address ranges of [0-2^32]; HyperSplit B-tree subspace 4 contains source IP address ranges of [3-2^32] and destination IP address ranges of [3-2^32]; HyperSplit B-tree subspace 5 contains source IP address ranges of [3-2^32] and destination IP address ranges of [0-3].

[0039] The HyperSplit A-tree contains 4 internal nodes and 5 leaf nodes, with the 5 leaf nodes corresponding to the 5 subspaces of the HyperSplit A-tree.

[0040] The HyperSplit B-tree contains 4 internal nodes and 5 leaf nodes, with the 5 leaf nodes corresponding to the 5 subspaces of the HyperSplit B-tree.

[0041] In Step 1 of the HyperSplit AB tree switching (Step 1 above in Step 1), the five subspaces (Tree A Space 1, Tree A Space 2, Tree A Space 3, Tree A Space 4, Tree A Space 5) contained in the HyperSplit A tree have been stored sequentially into storage unit blocks 1 to 5 (Mem Block 1, Mem Block 2, Mem Block 3, Mem Block 4, Mem Block 5). (In step 2 above in Step 1,) select the subspace with the fewest intersections with the HyperSplit B tree subspace from the 5 subspaces contained in the HyperSplit A tree. Let's say the selected subspace is HyperSplit A tree subspace 1 (Tree A Space 1), which has 1 intersection with the HyperSplit B tree subspace. Then, store the HyperSplit B tree subspace 1 (Tree B Space 1) that intersects with HyperSplit A tree subspace 1 (Tree A Space 1) in storage block 6 (Mem Block 6). Finally, set storage block 1 containing HyperSplit A tree subspace 1 (Tree A Space 1) to invalid. This completes the switching operation of HyperSplit A tree subspace 1 (Tree A Space 1). This switching only requires one additional storage block 6 (Mem Block 6).

[0042] In Step 2 of the HyperSplit AB tree switching, the remaining four subspaces (TreeA Space 2, Tree A Space 3, Tree A Space 4, Tree A Space 5) contained in the HyperSplit A tree are still stored sequentially in storage cell blocks 2 to 5 (Mem Block 2, Mem Block 3, Mem Block 4, Mem Block 5). First, select the subspace with the fewest intersections with the remaining four subspaces of the HyperSplit A tree from the remaining four subspaces of the HyperSplit B tree. In this case, the selected subspace is HyperSplit A tree subspace 2 (Tree ASpace 2), which has one intersection with the HyperSplit B tree subspace. Then, store HyperSplit B tree subspace 2 (Tree B Space 2), which intersects with HyperSplit A tree subspace 2 (Tree A Space 2), into storage unit block 1 (Mem Block 1). Finally, set storage unit block 2, containing HyperSplit A tree subspace 2 (Tree A Space 2), to invalidate. This completes the switching operation for HyperSplit A tree subspace 2 (Tree A Space 2). This switching does not require an additional storage unit block (Mem Block).

[0043] In step 3 of the HyperSplit AB tree switching, the remaining three subspaces (Tree A Space 3, Tree A Space 4, and Tree A Space 5) of the HyperSplit A tree are still stored sequentially in storage blocks 3 to 5 (Mem Block 3, Mem Block 4, and Mem Block 5). First, the subspace with the fewest intersections with the remaining three subspaces of the HyperSplit B tree is selected from the remaining three subspaces of the HyperSplit A tree. In this case, the selected subspace is HyperSplit A tree subspace 3, which has 1 intersection with the HyperSplit B tree subspace. Then, the HyperSplit B-tree subspace 3 that intersects with HyperSplit A-tree subspace 3 is stored in storage block 2. Finally, storage block 3 containing HyperSplit A-tree subspace 3 is set to invalid, thus completing the switching operation for HyperSplit A-tree subspace 3. This switching does not require an additional storage block.

[0044] In step 4 of the HyperSplit AB tree switching, the remaining two subspaces (Tree A Space 4 and Tree A Space 5) of the HyperSplit A tree are still stored sequentially in storage blocks 4 and 5 (Mem Block 4 and Mem Block 5). First, the subspace with the fewest intersections with the remaining two subspaces of the HyperSplit B tree is selected from the remaining two subspaces of the HyperSplit A tree. In this case, the selected subspace is HyperSplit A tree subspace 4, which has 1 intersection with the HyperSplit B tree subspace. Then, the HyperSplit B-tree subspace 4 (Tree B Space 4), which intersects with HyperSplit A-tree subspace 4, is stored in storage block 3 (Mem Block 3). Finally, storage block 4, containing HyperSplit A-tree subspace 4, is set to invalid, thus completing the switching operation for HyperSplit A-tree subspace 4 (Tree A Space 4). This switching does not require an additional storage block (Mem Block).

[0045] In step 5 of the HyperSplit AB tree switching process, the last subspace (Tree A Space 5) of the HyperSplit A tree is still stored in storage block 5 (Mem Block 5). The last subspace (Tree B Space 5) of the HyperSplit B tree that intersects with the last subspace (Tree A Space 5) of the HyperSplit A tree is stored in storage block 4 (Mem Block 4). Finally, storage block 5 containing the HyperSplit A tree subspace (Tree A Space 5) is set to invalid. This completes the gradual switch from HyperSplit A tree to HyperSplit B tree. Ultimately, the entire gradual switch from HyperSplit A tree to HyperSplit B tree requires only one additional storage block (Mem Block).

[0046] The specific processing flow of HyperSplit's progressive AB tree switching implementation is as follows: Figure 2As shown, the first step involves dividing the complete rule range space into various A-tree subspaces (Space(Tree A)) based on the rules contained in the A-tree, setting the state of all A-tree subspaces (Space(Tree A)) to valid, and then constructing the HyperSplit A-tree. Finally, all A-tree subspaces (Space(Tree A)) are stored in the storage space (Mem Block).

[0047] The second step is to divide the complete rule range space into various B-tree subspaces (Space(Tree B)) based on the rules contained in the B-tree, set the state of all B-tree subspaces (Space(Tree B)) to unloaded, and then construct the HyperSplit B-tree.

[0048] The third step is to obtain all the split and valid subspaces (Tree A) contained in the HyperSplit A tree.

[0049] The fourth step is to determine whether there is a valid subspace Space(Tree A) in the HyperSplit A tree. If it exists, proceed to the fifth step to continue execution; otherwise, end the current switching process.

[0050] The fifth step compares all the partitioned and valid subspaces Space(Tree A) in tree A with all the partitioned and unloaded subspaces Space(Tree B) in tree B. Based on the comparison results, the subspace Space'(Tree A) with the smallest number of intersections with the subspace Space(Tree B) in tree A is selected.

[0051] Step 6: Locate all B-tree subspaces (Tree B) that intersect with the selected A-tree subspace (Space'(Tree A)) from the previous step. Then, load the data of the found B-tree subspaces (Space'(Tree B)) into the storage space MemBlock and set its status to loaded. Finally, set the status of the A-tree subspace (Space'(Tree A)) to invalid; then jump to step 3 to continue execution.

[0052] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for implementing progressive switching of a HyperSplit AB tree, characterized in that, Includes the following steps: The first step is to divide the complete rule range space into various Space Tree A subspaces based on the rules contained in the HyperSplit A-tree, and set the status of all Space Tree A subspaces to valid. Then, the HyperSplit A-tree is constructed, and finally, all Space Tree A subspace data is stored in the storage space. The maximum value of the source IP address range in the rules contained in the HyperSplit A-tree is 2^32, and the maximum value of the destination IP address range in the HyperSplit A-tree is 2^32. The second step is to divide the complete rule range space into various B-tree subspaces Space Tree B according to the rules contained in the HyperSplit B-tree, set the status of all B-tree subspaces Space Tree B to unloaded, and then construct the HyperSplit B-tree; the maximum value of the source IP address range in the rules contained in the HyperSplit B-tree is 2^32, and the maximum value of the destination IP address range in the HyperSplit B-tree is 2^32. The third step is to obtain all the split subspaces that are valid in the HyperSplit A tree; Step 4: Determine if there is a valid subspace in the HyperSplit A tree. If it exists, proceed to step 5 and continue execution. If it does not exist, end the current switching process. Step 5: Compare all the partitioned and valid subspaces in tree A with all the partitioned and unloaded subspaces in tree B; based on the comparison results, select the subspace Space'Tree A from tree A that has the smallest number of intersections with subspace SpaceTree B in tree B. Step 6: Find all B-tree subspaces Space'Tree B that intersect with the A-tree subspace Space'Tree A selected in the previous step; then load the data of the found B-tree subspace Space'Tree B into the storage space and set its status to loaded; finally, set the status of the A-tree subspace Space'Tree A to invalid; then jump to step 3 to continue execution.

2. The method as described in claim 1, characterized in that, In the first step, five HyperSplit A-tree subspaces were created based on the rules contained in the HyperSplit A-tree: HyperSplit A-tree subspace 1 (Tree A Space 1), HyperSplit A-tree subspace 2 (Tree A Space 2), HyperSplit A-tree subspace 3 (Tree A Space 3), HyperSplit A-tree subspace 4 (Tree A Space 4), and HyperSplit A-tree subspace 5 (Tree A Space 5).

3. The method as described in claim 2, characterized in that, In the second step, five HyperSplit B-tree subspaces were created based on the rules contained in the HyperSplit B-tree: HyperSplit B-tree subspace 1 (Tree B Space 1), HyperSplit B-tree subspace 2 (Tree B Space 2), HyperSplit B-tree subspace 3 (Tree B Space 3), HyperSplit B-tree subspace 4 (Tree B Space 4), and HyperSplit B-tree subspace 5 (Tree B Space 5).

4. The method as described in claim 3, characterized in that, A HyperSplit A-tree contains 4 internal nodes and 5 leaf nodes, with the 5 leaf nodes corresponding to the 5 subspaces of the HyperSplit A-tree; a HyperSplit B-tree contains 4 internal nodes and 5 leaf nodes, with the 5 leaf nodes corresponding to the 5 subspaces of the HyperSplit B-tree.

5. The method as described in claim 4, characterized in that, In the first step, the five subspaces contained in the HyperSplit A tree, Tree A Space 1, Tree A Space 2, Tree A Space 3, Tree A Space 4, and Tree A Space 5, have been stored sequentially into storage unit blocks 1 to 5, namely Mem Block 1, Mem Block 2, Mem Block 3, Mem Block 4, and Mem Block 5. When executing the fifth step in the first loop, select the subspace with the fewest intersections with the subspaces of the HyperSplit B tree from the five subspaces contained in the HyperSplit A tree. Let's say that the selected subspace is the HyperSplit A tree subspace 1, which has 1 intersection with the subspaces of the HyperSplit B tree. During the sixth step of the first loop, the HyperSplit B subspace Tree B Space 1, which intersects with HyperSplit A subspace Tree A Space 1, is stored in storage block 6 (Mem Block 6). Finally, storage block 1 containing HyperSplit A subspace Tree A Space 1 is set to invalid, thus completing the switching operation of HyperSplit A subspace Tree A Space 1.

6. The method as described in claim 5, characterized in that, During the fifth step of the second loop, the subspace with the fewest intersections with the remaining four subspaces of the HyperSplit A tree is selected from the remaining four subspaces of the HyperSplit B tree. In this case, the selected subspace is HyperSplit A tree subspace 2 (Tree A Space 2), which has 1 intersection with the HyperSplit B tree subspace. During the sixth step of the second loop, the HyperSplit B tree subspace 2 (Tree B Space), which intersects with the HyperSplit A tree subspace, is stored in storage block Mem Block 1. Finally, storage block 2, where HyperSplit A tree subspace 2 is located, is set to invalid. This completes the switching of HyperSplit A tree subspace 2 (Tree A Space 2).

7. The method as described in claim 6, characterized in that, During the fifth step of the third loop, the subspace with the fewest intersections with the remaining three subspaces of the HyperSplit A tree is selected from the remaining three subspaces of the HyperSplit B tree. In this case, the selected subspace is HyperSplit A tree subspace 3 (Tree A Space 3), which has 1 intersection with the HyperSplit B tree subspace. During the sixth step of the third loop, HyperSplit B tree subspace 3 (Tree B Space 3), which intersects with HyperSplit A tree subspace 3 (Tree A Space 3), is stored in storage unit block 2 (Mem Block 2). Finally, storage unit block 3 containing HyperSplit A tree subspace 3 (Tree A Space 3) is set to invalid, thus completing the switching operation of HyperSplit A tree subspace 3 (Tree A Space 3).

8. The method as described in claim 7, characterized in that, During the fifth step of the fourth loop, the subspace with the fewest intersections with the remaining two subspaces of the HyperSplit A tree is selected from the remaining two subspaces of the HyperSplit B tree. In this case, the selected subspace is HyperSplit A tree subspace 4 (Tree A Space 4), which has 1 intersection with the HyperSplit B tree subspace. During the sixth step of the fourth loop, the HyperSplit B tree subspace 4 (Tree B Space 4), which intersects with HyperSplit A tree subspace 4 (Tree A Space 4), is stored in storage unit block 3 (Mem Block 3). Finally, storage unit block 4, where HyperSplit A tree subspace 4 (Tree A Space 4) is located, is set to invalid. This completes the switching operation of HyperSplit A tree subspace 4 (Tree A Space 4).

9. The method as described in claim 8, characterized in that, During the fifth iteration, in the sixth step, the last subspace of the HyperSplit B tree, Tree B Space 5, which intersects with the last subspace of the HyperSplit A tree, is stored in storage block 4 (Mem Block 4). Finally, storage block 5 containing the HyperSplit A tree subspace Tree A Space 5 is set to invalid. This completes the gradual switch from HyperSplit A tree to HyperSplit B tree.

Images