WO2016197706A1 - Data migration method and device - Google Patents

Abstract

A data migration method and device. The method comprises: acquiring a start rate of a plurality of nodes in a communication network and a storage space utilization rate of the plurality of nodes; classifying the priority of the plurality of nodes according to a preset rule, the start rate and the storage space utilization rate, so as to obtain node sets classified according to the priority; and migrating data between the node sets according to the priority. By means of the solution, the problem in the relevant art that only one aspect of a space factor such as a space utilization rate or a time factor such as a service quality operation rate is taken into consideration is solved, thereby achieving the effect of improving the system efficiency.

Description

Data migration method and device Technical field

The present application relates to, but is not limited to, the field of communications, and in particular, to a data migration method and apparatus.

Background technique

With the rapid development of communication technology, the traffic of each node (communication device) in the communication network layout is getting larger and larger. Although large-capacity and high-performance communication devices are constantly appearing, in order to rationally utilize resources, each communication is improved. The service processing capability of the node ensures better network quality. It is necessary to migrate the corresponding configuration data to nodes with fast startup rate and small space utilization, so that the overall communication node service processing is more balanced.

There are a plurality of data migration methods in the related art. For example, in a related art solution, a data migration task to be scheduled in each source node is determined according to a quality of service QOS of a source node to a destination node in the network. And sending a scheduling command to the target source node corresponding to the to-be-scheduled data migration task, where the scheduling command is used to schedule the data migration task, which improves data migration efficiency and avoids network congestion, but the method only The "time" factor of quality of service between nodes is used as the criterion for judging the priority of data migration. It does not take into account the "space" factor of storage between nodes, which may lead to excessive storage pressure of some nodes with large space utilization. Reduce the life of your hard drive.

In another related technical solution, a method for calculating an average value of the space utilization ratio of the same type of hard disk is adopted, and space utilization of each hard disk in any one of the storage pools is obtained, and the space of the same type is used. The data in the hard disk having a utilization greater than the average value is migrated to the hard disk in the same type where the space utilization is less than the average value. It reduces the pressure on the hard disk with large data storage capacity and prolongs the service life of the hard disk with large data storage capacity. However, this method only takes the storage space utilization of the node as the priority judgment criterion for data migration, and does not take care of the storage. The "time" factor of the media object is judged.

It can be seen that the related technologies often only consider the spatial factors such as space utilization or the service efficiency of service quality. On the one hand, the two factors of time and space are not comprehensively considered; for the above problems in related technologies, Propose an effective solution.

Summary of the invention

The following is an overview of the topics detailed in this document. This Summary is not intended to limit the scope of the claims.

The main purpose of the present application is to provide a method and an apparatus for migrating data, so as to at least solve the problem of one of the time factors such as space utilization and service quality, such as space utilization and other operational factors during data migration in the related art. .

According to an aspect of the present application, a data migration method is provided, including: acquiring a startup rate of a plurality of nodes in a communication network and a storage space utilization ratio of the plurality of nodes; according to a preset rule, the startup rate, and The storage space utilization classifies the priorities of the multiple nodes to obtain a set of nodes classified according to priorities; and migrates data between the set of nodes according to priorities.

Optionally, the step of acquiring a startup rate of the multiple nodes in the communication network and the storage space utilization of the multiple nodes includes: acquiring a time difference between a power-on time and a normal running time of the multiple nodes, where And using the time difference as a startup rate of the multiple nodes; acquiring a proportion of a space size of all files in the storage space of the multiple nodes to a total space size, and using the specific gravity as the storage space rate.

Optionally, the step of classifying the priorities of the multiple nodes according to the preset rule, the startup rate, and the storage space utilization, and obtaining the node set classified according to the priority includes: Generating the plurality of node feature vectors by using the startup rate and the storage space utilization; classifying the priorities of the plurality of feature vectors by cluster analysis, and dividing the plurality of nodes according to the result of the cluster analysis A first node set and a second node set, wherein the nodes in the first node set have a higher priority than the nodes in the second node set.

Optionally, the step of classifying the priorities of the multiple feature vectors by cluster analysis, and dividing the multiple nodes into the first node set and the second node set according to the result of the cluster analysis includes Presetting a feature vector of the first initial cluster center and the second initial cluster center, wherein the first initial cluster center belongs to the first node set, and the second initial cluster center belongs to the first a two-node set; obtaining a distance value between a feature vector of the plurality of nodes and a feature vector of the first initial cluster center and the second initial cluster center; respectively, according to the distance value and the closest distance a matching principle, the plurality of nodes are allocated to the first node set or the second node set that are closer to each other; and all nodes in the first node set are acquired relative to the first initial cluster a distance average of the center, and obtaining a distance average of all nodes in the second node set relative to the second initial cluster center, and adjusting the first initial cluster center or the initial according to the distance mean respectively The second initial cluster center.

Optionally, the step of migrating data between the set of nodes according to the priority includes: migrating data to be migrated of the migrating node in the second set of nodes to the migrating node of the first set of nodes.

Optionally, after the step of migrating data between the node sets according to the priority, the method further includes: when the system of the migrating node needs to access the migrated data, The migrated data in the migrated node is restored to the migrated node.

The application further provides a computer readable storage medium storing computer executable instructions that are implemented when the computer executable instructions are executed.

According to another aspect of the present application, a data migration apparatus is provided, including: an acquisition module configured to acquire a startup rate of a plurality of nodes in a communication network and a storage space utilization ratio of the plurality of nodes; a classification module, setting Sorting the priorities of the multiple nodes according to a preset rule, the startup rate, and the storage space utilization, and obtaining a node set classified according to the priority; the migration module is set to be based on the priority, The data is migrated between the set of nodes.

Optionally, the acquiring module includes: a first acquiring unit, configured to acquire a time difference between a power-on time and a normal running time of the multiple nodes, and use the time difference as a starting rate of the multiple nodes The second obtaining unit is configured to obtain a proportion of a space size of all files in the storage space of the plurality of nodes to a total space size, and use the specific gravity as the storage space utilization rate.

Optionally, the classification module includes: a generating unit, configured to generate the multiple node feature vectors according to the startup rate and the storage space utilization rate; and a classification unit configured to perform the multiple The priority of the feature vector is classified, and the plurality of nodes are divided into a first node set and a second node set according to a result of the cluster analysis, wherein the nodes in the first node set have higher priority than the first node set The priority of the nodes in the second set of nodes.

Optionally, the classification unit includes: a preset subunit, configured to preset in the first initial cluster a feature vector of the heart and the second initial cluster center, wherein the first initial cluster center belongs to the first node set, the second initial cluster center belongs to the second node set; and the subunit is acquired And configured to obtain a distance value between the feature vector of the plurality of nodes and the feature vector of the first initial cluster center and the second initial cluster center respectively; the allocation subunit is configured to be allocated according to the distance value and the closest distance In principle, the plurality of nodes are allocated to the first node set or the second node set that are closer to themselves; the adjusting subunit is configured to acquire all nodes in the first node set relative to the a distance average of the first initial cluster center, and obtaining a distance average of all nodes in the second node set relative to the second initial cluster center, and respectively adjusting the first initial cluster according to the distance mean Center or the initial second initial cluster center.

In the present application, the priority of multiple nodes is classified according to a preset rule, a startup rate, and a storage space utilization, and a node set classified according to priority is obtained, and is migrated between the node sets according to the priority. The data, that is, the startup rate and the storage space utilization ratio of the nodes in the communication network are combined in the present application, and the data between the nodes is migrated in a priority manner, and the related art only considers the data migration when performing data migration. The problem of one of the time factors such as space utilization or service efficiency of service quality, and thus the effect of improving system efficiency.

Other aspects will be apparent upon reading and understanding the drawings and detailed description.

BRIEF abstract

The drawings described herein are intended to provide a further understanding of the present application, and are intended to be a part of this application. In the drawing:

1 is a flowchart of a method of migrating data according to an embodiment of the present invention;

2 is a structural block diagram of a data migration apparatus according to an embodiment of the present invention;

3 is a block diagram showing an optional structure of an obtaining module 22 of a data migration apparatus according to an embodiment of the present invention;

4 is a block diagram 1 of an optional structure of a classification module 24 of a data migration apparatus according to an embodiment of the present invention;

FIG. 5 is an optional structural block of the classification module 24 of the data migration apparatus according to an embodiment of the present invention. Figure II;

6 is a block diagram 1 of an optional structure of a data migration apparatus according to an embodiment of the present invention;

7 is a block diagram showing the structure of a multipoint priority data migration apparatus according to an alternative embodiment of the present invention;

8 is a flowchart of a multi-point priority data migration method in accordance with an alternative embodiment of the present invention;

9 is a flow chart of clustering analysis of a communication node in accordance with an alternative embodiment of the present invention;

10 is a schematic diagram of migration data backup and recovery in accordance with an alternate embodiment of the present invention.

Preferred embodiment of the invention

It should be noted that the embodiments in the present application and the features in the embodiments may be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings.

This embodiment provides a method for migrating data. FIG. 1 is a flowchart of a method for migrating data according to an embodiment of the present invention. As shown in FIG. 1, the steps of the method include:

Step S102: Acquire a startup rate of multiple nodes in the communication network and a storage space utilization ratio of the multiple nodes.

Step S104: classify the priorities of the multiple nodes according to a preset rule, a startup rate, and a storage space utilization, and obtain a node set classified according to the priority;

Step S106: migrating data between the node sets according to the priority.

In the foregoing steps S102 to S106 of the embodiment, the priority of the multiple nodes is classified according to the startup rate and the storage space utilization by using a preset rule, and the node set classified according to the priority is obtained, and the priority is determined according to the priority. The data is migrated between the node sets, that is, in this embodiment, the startup rate and the storage space utilization rate of the nodes in the communication network are combined, and the data is migrated between the nodes according to the priority, and the migration data in the related art is solved. At the time, only one aspect of the time factor such as space utilization or service quality operation efficiency is considered, and the effect of improving system efficiency is achieved.

The method of obtaining the startup rate of the multiple nodes in the communication network and the storage space utilization ratio of the multiple nodes in the step S102 of the embodiment may be implemented by the following steps in the optional implementation manner of the embodiment:

Step S11: Obtain a time difference between a power-on time and a normal running time of the multiple nodes, and use the time difference as a starting rate of the multiple nodes.

Step S12: Acquire the proportion of the space size of all files in the storage space of the plurality of nodes to the total space size, and use the specific gravity as the storage space utilization rate.

In an application scenario of this embodiment, the startup rate and the storage space utilization involved in the foregoing step S11 and the step S12 may be:

If the node is powered on at 10:00, and the node returns to normal operation after 5 minutes, the startup rate of the node is 5 minutes. The step may be a program running by the node. If the node power-on time is t _i1 , when the node is put into operation normally, and the corresponding working time t _i2 is counted again, the startup rate of the node x _i may be expressed as follows: v _i =t _i2 -t _i1

If the total storage capacity of the second node is 4G, and the total storage file size of the node is 2G, the node space utilization is 50%. The step may be a program running in the node. If the file size of the memory card in the node is s ₁ and the total memory capacity of the node is s ₂ , the storage space utilization of the node may be expressed as follows: s _i = s ₁ /s ₂

For the step S104 in this embodiment, the priority of the multiple nodes is classified according to the startup rate and the storage space utilization by using a preset rule, and the node set classified according to the priority is obtained. In the way, it is implemented as follows:

Step S22: Generate multiple node feature vectors according to the startup rate and the storage space utilization rate;

Step S23: classifying the priorities of the plurality of feature vectors by cluster analysis, and dividing the plurality of nodes into a first node set and a second node set according to a result of the cluster analysis, wherein the first node The nodes in the set have a higher priority than the nodes in the second set of nodes.

For the step S23, the priority of the plurality of feature vectors is classified by cluster analysis, and the plurality of nodes are divided into a first node set and a second node set according to the result of the cluster analysis. include:

Step S31: Presetting the feature vectors of the first initial cluster center and the second initial cluster center, wherein the first initial cluster center belongs to the first node set, and the second initial cluster center belongs to the second node set;

Step S32: acquiring feature vectors of the plurality of nodes and the first initial cluster center and the first The distance value between the feature vectors of the two initial cluster centers;

Step S33: assigning a plurality of nodes to the first node set or the second node set that are closer to each other according to the distance value and the nearest distance allocation principle;

Step S34: Acquire a distance average value of all nodes in the first node set relative to the first initial cluster center, and obtain a distance average value of all nodes in the second node set relative to the second initial cluster center, and adjust according to the distance average value respectively. The first initial cluster center or the initial second initial cluster center.

In combination with the above-mentioned starting rate and the value of the storage space utilization, in the manner of the foregoing step S22, an application scenario in this embodiment may be:

The obtained pattern feature vector is used as the input of the unit for cluster analysis. First, the distance between each feature vector is counted. The distance between the node x _i and the node x _j is expressed as follows:

According to the criterion of clustering with priority weights, K initial cluster centers are selected;

Wherein, in this alternative example, K=2, by heuristically selecting the initial values of the K initial cluster centers, assuming that the second node is the initial cluster center, then z ₂ = (v ₂ , s ₂ ), the initial value of the cluster center can be expressed as follows: Z = {z _j | j = 1, ... k}

The sample x _i in the sample set is assigned to the nearest neighbor cluster z _j according to the principle of minimum distance, which can be expressed as follows:

d _ij =min(||x _i -z _j ||), x _i ∈X,z _j ∈Z

It can be seen that, in the application scenario, two node sets are obtained according to the two initial cluster centers, and the start rate and the storage space utilization are divided in the two sets, and two node sets with different priorities are obtained.

For the step S106 of the embodiment, the step of migrating data between the node sets includes: migrating the data to be migrated of the migrating node in the second node set to the migrating node of the first node set.

In addition, in another optional implementation manner of this embodiment, after the step of migrating data between the node sets according to the priority, the method in this embodiment further includes: the system in the migrating node needs to access When the data is migrated, the data moved in the moved node is restored to the evicted node.

Embodiments of the present invention further provide a computer readable storage medium storing a computer executable The instructions, when the computer executable instructions are executed, implement the above method.

In the embodiment, a data migration device is further provided, and the device is used to implement the foregoing embodiments and preferred embodiments, and details are not described herein. As used below, the term "module" may implement a combination of software and/or hardware of a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware, or a combination of software and hardware, is also possible and contemplated.

2 is a structural block diagram of a data migration apparatus according to an embodiment of the present invention. As shown in FIG. 2, the apparatus includes: an acquisition module 22 configured to acquire a startup rate of a plurality of nodes in a communication network and the plurality of nodes. The storage space utilization; the classification module 24 is coupled to the acquisition module 22, and is configured to classify the priorities of the multiple nodes according to the preset rules, the startup rate, and the storage space utilization, and obtain the node sets classified according to the priority. The migration module 26 is coupled to the classification module 24 and configured to migrate data between sets of nodes in accordance with priority.

FIG. 3 is a block diagram showing an optional structure of the acquiring module 22 of the data migration apparatus according to the embodiment of the present invention. As shown in FIG. 3, the obtaining module 22 includes: a first obtaining unit 32 configured to obtain power-on of multiple nodes. The time difference between the time and the normal running time, and the time difference is used as the starting rate of the plurality of nodes; the second obtaining unit 34 is configured to obtain the space size of all the files in the storage space of the plurality of nodes. The proportion of the size of the space, and the specific gravity as the storage space utilization.

4 is a block diagram of an optional structure of a classification module 24 of a data migration apparatus according to an embodiment of the present invention. As shown in FIG. 4, the classification module 24 includes: a generation unit 42 configured to be based on a startup rate and a storage space utilization rate. Generating a plurality of node feature vectors; the classifying unit 44 is coupled to the generating unit 42 and configured to classify the plurality of feature vectors by cluster analysis, and divide the plurality of nodes according to the result of the cluster analysis. The first node set and the second node set, wherein the nodes in the first node set have higher priority than the node priority in the second node set.

FIG. 5 is a block diagram 2 of an optional structure of a classification module 24 of a data migration apparatus according to an embodiment of the present invention. As shown in FIG. 5, the classification unit 44 includes: a preset subunit 52 configured to preset a first initial aggregation. a feature vector of the class center and the second initial cluster center, wherein the first initial cluster center belongs to the first node set, the second initial cluster center belongs to the second node set; the obtaining sub-unit 54, and the preset sub-unit 52 a coupling connection, configured to acquire feature vectors of the plurality of nodes and first a distance value between the feature vector of the initial cluster center and the second initial cluster center; the assigning subunit 56 is coupled with the obtaining subunit 54 and configured to allocate the plurality of nodes according to the distance value and the nearest distance allocation principle. a closer first node set or a second node set; the adjusting subunit is coupled to the assigning subunit 56, configured to obtain a distance average of all nodes in the first node set relative to the first initial cluster center, and obtain The distance between all nodes in the second node set relative to the second initial cluster center is averaged, and the first initial cluster center or the initial second initial cluster center is respectively adjusted according to the distance mean.

Optionally, the migration module 26 is further configured to migrate the data to be migrated of the migrated node in the second node set to the migrated node of the first node set.

FIG. 6 is a block diagram of an optional structure of a data migration apparatus according to an embodiment of the present invention. As shown in FIG. 6, the apparatus further includes: a recovery module 62 coupled to the migration module 26 and configured to be required by the system of the migration node. When accessing the migrated data, the migrated data in the migrated node is restored to the migrated node.

The present application is described in detail below in conjunction with an optional embodiment of the present invention;

The present invention provides a multi-point priority data migration method and apparatus. According to the optional embodiment, the configuration data with low priority weight can be migrated based on the comprehensive consideration of the time factor and the spatial factor of the communication node. The node with low priority weight effectively integrates system resources and improves the overall service processing capability of the communication node.

FIG. 7 is a structural block diagram of a multi-point priority data migration apparatus according to an alternative embodiment of the present invention. As shown in FIG. 7, the apparatus includes: a data pre-processing module (the acquisition module 22 and the classification module 24 in the foregoing embodiment). a data backup module (corresponding to the migration module 26 in the above embodiment) and a data recovery module (corresponding to the recovery module 62 in the above embodiment), wherein the data preprocessing module includes: a statistical startup rate unit ( Corresponding to the first acquisition unit 32), the statistical space utilization unit (corresponding to the second acquisition unit 34 in the above embodiment), the mode feature processing unit (corresponding to the generation unit 42 of the above embodiment), and the right in the above embodiment. A value priority processing unit (corresponding to the classification unit 44 in the present embodiment described above).

FIG. 8 is a flowchart of a multi-point priority data migration method according to an alternative embodiment of the present invention. Based on the module of the apparatus in FIG. 7, the steps of the method shown in FIG. 8 include:

Step S802: number each node in the communication network layout as a subsequent pattern clustering Collection of pattern samples;

Step S804: Counting the startup rate of each node in the communication network layout;

Step S806: Counting the space utilization rate of the memory card of each node in the communication network layout;

Step S808: Taking the startup rate and the space utilization rate of each node as the classification feature vector of the node;

Step S810: Perform cluster analysis on feature vectors of each node, and classify the sample set of the communication node into two categories: high and low data migration priority weights;

Step S812: The configuration data of the corresponding node with a lower priority weight is migrated to the classification node with a higher priority weight.

Based on the foregoing data pre-processing module, optionally, step S804 includes: calculating, by using a statistical start rate unit, a difference between a power-on time and a normal running time of each node in the communication layout as a startup rate of the node.

Based on the foregoing data pre-processing module, optionally, step S806 includes: obtaining a space utilization ratio of the node by using a ratio of a space size of all files of the memory card of each node calculated by the space utilization unit to a total space size.

Based on the data pre-processing module, optionally, step S808 includes: using a value obtained by each node statistical start rate unit and a statistical space utilization unit as a priority processing unit by the mode feature processing unit to obtain a corresponding mode feature vector. .

Based on the foregoing data pre-processing module, optionally, step S810 includes: performing cluster analysis on the pattern feature vectors of the respective nodes by using the weight priority processing unit to obtain two sets of high priority weights and low priority weights. .

For step S804 to step S810, optionally,

Step S804 includes: counting the power-on time of each node in the communication network layout and the time after the node is working normally by using the statistical start rate unit, and using the power-on time and the normal working time as the input of the statistical start rate unit, and outputting the node as the corresponding Start rate.

Step S806 includes: counting the size of all the files of the memory cards of each node in the communication network layout and the total capacity of the memory card by using the statistical space utilization unit, and using all the file sizes and the total capacity of the memory card as the statistical space utilization unit. Input, output is the corresponding empty space of the node Utilization rate.

Step S808 includes: passing, by the mode feature processing unit, an output of the statistical start rate unit and an output of the statistical space utilization unit as an input of the mode feature processing unit, and the output of the mode feature processing unit is a mode feature vector of the corresponding node.

Step S810 includes: inputting, by the weight priority processing unit, an output of the pattern feature processing unit as a weight priority processing unit, and processing, by the unit, the output of the weight priority processing unit is a priority weight and priority Two sets of low-level nodes.

Optionally, the processing method of the weight priority processing unit includes:

Dividing two classification sets; one of them is a classification with high priority weight, which is characterized by small space utilization of nodes and high startup rate of nodes, and one type is classification with low priority weights, which is characterized by space of nodes. The utilization rate is large and the startup rate of the node is low. An initial cluster center is determined for the two classification sets respectively, the sample nodes are allocated to the nearest cluster set according to the nearest distance allocation principle, and the sample mean of each cluster set is used as the new cluster center. Repeat the above steps so that the cluster centers of the two cluster sets no longer change. If the last sample points are the same distance from the two cluster centers, then the sample points are removed, indicating that the node does not need data migration. . Finally, two cluster sets with corresponding priority weights and low priority weights are obtained.

FIG. 9 is a flowchart of performing cluster analysis on a communication node according to an alternative embodiment of the present invention. As shown in FIG. 9, the steps of the process include:

Step S902: obtaining a feature vector of the mode sample;

Step S904: Heuristically selecting initial values of two cluster centers;

Step S906: each data point is classified into a classification to which the cluster center closest to it belongs;

Step S908: Calculate a new cluster center of each category;

Step S910: determining whether the cluster center changes, and determining that there is no change, executing step S906, determining that there is a change, executing step S912;

Step S912: The node set is divided into two categories according to the weight of the priority.

The method steps involved in the foregoing step S812 in the present embodiment involving performing backup by the data backup module include:

Step S41: determining, according to the output of the data migration pre-processing unit, that is, the classification set of the two nodes, the data migration node and the data migration node;

Step S42: determining basic configuration data to be migrated by the data migration node and a storage path of the data to be migrated;

Step S43: determining the IP address of the node where the data is moved into the node, the slot number of the rack, and the disk storage path of the migrated data;

Step S44: The corresponding basic configuration data in the source node (the node where the data is migrated) is migrated to the corresponding storage path of the destination node (the node into which the data is moved).

For the data recovery method of the data migration in the alternative embodiment, the data recovery method of the data migration may be implemented by using the data recovery module in the optional embodiment, and the method includes:

Step S51: determining the source node IP address of the data migration, the slot number of the rack, and the storage path of the data;

Step S52: determining an IP address of the data migration destination node, a slot number of the rack, and a storage path of the data;

Step S53: When the system of the data migration node needs to access the migrated configuration data, the corresponding configuration data in the node where the data is migrated is restored to the node where the data is moved out.

For the above steps S41 to S44, and the data backup and recovery involved in steps S51 to S53, as shown in FIG. 10, FIG. 10 is a schematic diagram of migration data backup and recovery according to an alternative embodiment of the present invention.

According to the multi-point priority data migration method of the alternative embodiment, after the multi-node is started, the priority weight of the node is determined according to the node startup rate and the storage space utilization, and the node basic configuration data with the lower priority weight is set. It is migrated to a node with a higher priority weight. This method not only improves the service life of the memory of a node with a large storage capacity, but also improves the system efficiency.

The present application is described in conjunction with the above-mentioned FIG. 7 and FIG. 8 and the specific embodiments of the present optional embodiment;

The data migration method provided in this alternative embodiment is applied to the communication network layout. Each node in the communication network layout may be a communication board, and the migrated storage unit is a memory card in the board. The data migration method includes the following steps:

Step S61, numbering each node in the communication network layout;

Wherein, assuming that the total number of nodes in the network layout that need to perform data migration is n, n nodes in the communication network layout can be expressed as follows:

X={x ₁ ,x ₂ ,...x _i ,...x _n }

Step S62, the rate of starting the node is counted by the statistics start rate unit;

For example, if the node is powered on at 10:00, and the node returns to normal operation after 5 minutes, the startup rate of the node is 5 minutes. The statistical start rate unit may be a program running by the node. If the node power-on time is t _i1 , when the node is put into normal operation and the corresponding working time t _i2 is counted again, the startup rate of the node x _i may be expressed as follows. :

v _i =t _i2 -t _i1

Step S63, collecting the space utilization rate of the node memory card by using the statistical space utilization unit;

If the total capacity of the memory card of the second node is 4G, and the total file size of the memory card is 2G, the space utilization rate of the node is 50%. The statistical space utilization unit can be used as a program running in the node. If the file size of the memory card in the node is s ₁ and the total storage capacity of the node is s ₂ , the storage space utilization of the node can be expressed as follows:

s _i =s ₁ /s ₂

Mode wherein the processing unit will start node rate of v _i x _i and s _i with a space utilization mode that performs clustering feature vector, a feature vector pattern node x _i may be expressed as follows:

y _i =(v _i ,s _i )i=1,2,...n

Step S64, the priority weight processing unit performs cluster analysis on the mode feature vector obtained by the mode feature processing unit as an input of the unit;

First, the distance between each feature vector is first counted, and the distance between the node x _i and the node x _j is expressed as follows:

Step S65, selecting K initial cluster centers according to clustering criteria of priority weights;

Wherein, in this alternative example, K=2, by heuristically selecting the initial values of the K initial cluster centers, assuming that the second node is the initial cluster center, then z ₂ = (v ₂ , s ₂ ), the initial value of the cluster center can be expressed as follows:

Z={z _j |j=1,...k}

The sample x _i in the sample set is assigned to the nearest neighbor cluster z _j according to the principle of minimum distance, which can be expressed as follows:

d _ij =min(||x _i -z _j ||), x _i ∈X,z _j ∈Z

In step S66, using the distance average of the samples in each cluster relative to the initial cluster center as the new cluster center, step S64 and step S65 are repeated until the cluster center point of the sample center does not change.

Two classifications with high and low data migration priority weights are obtained. The classification with high priority is characterized by high node space utilization and slow node startup rate. The classification feature with low priority weight is low node space utilization and node startup. The rate is fast.

Step S67: The data backup unit migrates the data migrated out of the set and the data migration set by the data migration preprocessing module to perform corresponding backup;

The IP address of the node where the data is migrated is determined. The board in the communication node is the same as the standby board and the standby board. The IP address is the same. You need to continue to determine the slot number of the rack. Based on the IP address and the slot number, a node where data is migrated can be determined, and then the data migration unit of the data migration node is determined, and the storage path of the migration data is counted. For example, if the IP address is 192.100.90.1 and the slot number is 2, the basic data in the storage path /home/sd is migrated out.

Step S68, determining an IP address, a slot number, and a data storage path of the data moving into the node;

For example, the data migration node IP address is 192.100.90.3, the destination node slot number is 1, and the data migration path is /home/sd/bak.

Step S69, the corresponding configuration data of the data migration node is migrated to the storage location corresponding to the corresponding storage path of the data migration node through the network;

For example, if the IP address is 192.100.90.1 and the slot number is 2, the data in the /home/sd directory is migrated to the IP address of 192.100.90.3 and the slot number is 1. The storage location corresponding to /home/sd/bak.

Step S70: When the system of the data migration node needs to access the migrated configuration data, the data recovery process is triggered;

The IP address of the source node and the destination node, the slot number of the rack, and the storage path of the data unit are first restored.

In step S71, the corresponding configuration data in the node where the data is moved is restored to the node where the data is moved out.

For example, if the IP address is 192.100.90.3 and the slot number is 1, the configuration data of the storage path is /home/sd/bak is restored to the IP address of 192.100.90.1 and the storage path of the node with slot number 2 is / The storage location corresponding to home/sd.

In another embodiment, software is also provided for performing the technical solutions described in the above embodiments and alternative embodiments.

In another embodiment, a storage medium is further provided, wherein the software includes the above-mentioned software, including but not limited to: an optical disk, a floppy disk, a hard disk, an erasable memory, and the like.

Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device such that they may be stored in the storage device for execution by the computing device and, in some cases, may be performed in a different order than herein. The steps shown or described are either made separately into individual integrated circuit modules, or a plurality of modules or steps are fabricated as a single integrated circuit module. Thus, the application is not limited to any particular combination of hardware and software.

The above is only an alternative embodiment of the present invention, and is not intended to limit the present application, and various changes and modifications may be made in the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of this application are intended to be included within the scope of the present application.

One of ordinary skill in the art will appreciate that all or a portion of the above steps may be performed by a program to instruct related hardware, such as a processor, which may be stored in a computer readable storage medium, such as a read only memory, disk or optical disk. Wait. Alternatively, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the above embodiment may be implemented in the form of hardware, for example, by implementing an integrated circuit to implement its corresponding function, or may be implemented in the form of a software function module, for example, executing a program stored in the memory by a processor. / instruction to achieve its corresponding function. Embodiments of the invention are not limited to any particular form of hardware and The combination of software.

Industrial applicability

According to the preset rule, the startup rate, and the storage space utilization, the priorities of the multiple nodes are classified, the node sets classified according to the priority are obtained, and the data is migrated between the node sets according to the priority, that is, in the present The application combines the startup rate and storage space utilization of nodes in the communication network, and migrates data between nodes in a priority manner, which solves the space in the related art when only data utilization is considered. The problem of one of the factors such as factors or service quality operation efficiency, and thus the effect of improving system efficiency.

Claims (10)

1. A method of data migration, including:

   Obtaining a startup rate of a plurality of nodes in the communication network and a storage space utilization ratio of the plurality of nodes;

   Sorting priorities of the multiple nodes according to a preset rule, the startup rate, and the storage space utilization, to obtain a node set classified according to priority;

   Data is migrated between the set of nodes according to priority.
2. The method of claim 1, wherein the step of obtaining a startup rate of a plurality of nodes in the communication network and a storage space utilization of the plurality of nodes comprises:

   Obtaining a time difference between a power-on time and a normal running time of the multiple nodes, and using the time difference as a starting rate of the multiple nodes;

   Obtaining a proportion of a space size of all files in the storage space of the multiple nodes to a total space size, and using the specific gravity as the storage space utilization rate.
3. The method according to claim 1, wherein the classifying priorities of the plurality of nodes according to a preset rule, the startup rate, and the storage space utilization, and obtaining nodes classified according to priorities The steps of the collection include:

   Generating a feature vector of the plurality of nodes according to the startup rate and the storage space utilization;

   Sorting the priorities of the plurality of feature vectors by cluster analysis, and dividing the plurality of nodes into a first node set and a second node set according to a result of the cluster analysis, wherein the first node set The priority of the nodes in the middle is higher than the priority of the nodes in the second set of nodes.
4. The method according to claim 3, wherein said prioritizing said plurality of feature vectors by cluster analysis, and dividing said plurality of nodes into a first node set according to a result of cluster analysis The steps of the second node set include:

   Presetting a feature vector of the first initial cluster center and the second initial cluster center, wherein the first initial cluster center belongs to the first node set, and the second initial cluster center belongs to the second Node set

   Obtaining feature vectors of the plurality of nodes and a first initial cluster center and a second initial cluster respectively The distance between the feature vectors of the center;

   And assigning, according to the distance value and the nearest distance allocation principle, the plurality of nodes to the first node set or the second node set that are closer to themselves;

   Obtaining a distance average of all nodes in the first node set relative to the first initial cluster center, and obtaining an average distance of all nodes in the second node set relative to the second initial cluster center, and The first initial cluster center or the initial second initial cluster center is respectively adjusted according to the distance mean.
5. The method of claim 4 wherein said step of migrating data between said set of nodes in accordance with a priority comprises:

   The data to be migrated of the migrated node in the second node set is migrated to the migrated node of the first node set.
6. The method of claim 5, after the step of migrating data between the set of nodes according to a priority, the method further comprises:

   When the system of the migrating node needs to access the migrated data, the data migrated in the migrating node is restored to the migrating node.
7. A data migration device comprising:

   Obtaining a module, configured to obtain a startup rate of a plurality of nodes in the communication network and a storage space utilization ratio of the plurality of nodes;

   a classifying module, configured to classify priorities of the plurality of nodes according to a preset rule, the startup rate, and the storage space utilization, to obtain a node set classified according to priorities;

   The migration module is configured to migrate data between the set of nodes according to a priority.
8. The apparatus of claim 7, wherein the obtaining module comprises:

   a first acquiring unit, configured to acquire a time difference between a power-on time and a normal running time of the multiple nodes, and use the time difference as a starting rate of the multiple nodes;

   The second obtaining unit is configured to obtain a proportion of a space size of all files in the storage space of the plurality of nodes to a total space size, and use the specific gravity as the storage space utilization rate.
9. The apparatus of claim 7 wherein said classification module comprises:

   Generating unit, configured to generate a feature vector of the plurality of nodes according to the startup rate and the storage space utilization rate;

   a classifying unit configured to classify priorities of the plurality of feature vectors by cluster analysis, and divide the plurality of nodes into a first node set and a second node set according to a result of the cluster analysis, where The nodes in the first set of nodes have a higher priority than the nodes in the second set of nodes.
10. The apparatus of claim 9, wherein the classification unit comprises:

    a preset subunit, configured to preset feature vectors of the first initial cluster center and the second initial cluster center, wherein the first initial cluster center belongs to the first node set, and the second initial gather The class center belongs to the second node set;

    Obtaining a subunit, configured to obtain a distance value between a feature vector of the plurality of nodes and a feature vector of the first initial cluster center and the second initial cluster center;

    Assigning a subunit, configured to allocate the plurality of nodes to the first node set or the second node set that are closer to each other according to the distance value and the nearest distance allocation principle;

    Adjusting a subunit, configured to obtain a distance average of all nodes in the first node set relative to the first initial cluster center, and obtain all nodes in the second node set relative to the second initial cluster The distance of the center is averaged, and the first initial cluster center or the initial second initial cluster center is respectively adjusted according to the distance mean.

Images