A content distribution method, apparatus, electronic device, and storage medium

Abstract

The application discloses a content distribution method and device, electronic equipment and storage medium, and relates to the technical field of communication. The method comprises the following steps: receiving an operation of triggering a content distribution function by a user, and acquiring a historical edge node accessed by the user; constructing an interest propagation model for a content distribution network based on the historical edge node; determining a target time window node, predicting a target edge node to be accessed by the user at the target time window node based on the interest propagation model; determining content to be distributed corresponding to the target edge node, and distributing the content to be distributed to the user based on the target edge node. The technical scheme provided by the application can accurately predict the content that the user is interested in at a certain time window in the future, so that the prediction is more accurate and unitized, and the network congestion can be improved and the user access response speed and content hit rate can be improved. Compared with the current general nearby access method, intelligent scheduling and resource allocation are performed, and resources are best utilized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to a content distribution method and device, electronic equipment and storage medium. BACKGROUND

[0002] Through a content delivery network (CDN), a user can directly access a node to obtain resources of a website without directly accessing a source website. A content distribution strategy of the CDN determines whether a cache server stores resources, and the content distribution strategy affects the time delay experience of the user and the bandwidth flow of a core backbone network, and the bandwidth flow directly affects the cost of consumption, so that the time delay and the bandwidth flow of the core network are important indicators for measuring the content distribution strategy. The purpose of the CDN distribution is to enable resources to be distributed to edge nodes closest to users as much as possible, so as to enable users to access resources on the edge nodes closest to them without accessing a central node far away. Therefore, nearby access is also an important indicator for measuring the content distribution strategy.

[0003] A user usually accesses content related to work and study during working hours, and usually accesses content related to entertainment and games after work. In addition, the user is interested in some sudden hot spots for a period of time, causing a sharp increase in the flow of CDN nodes, and most of the bandwidth is in an idle state at night. When the content to be distributed by the edge node is not of interest to the user, network congestion, slow user access response speed, and low content hit rate are caused, thereby increasing the scheduling pressure of the CDN, and also affecting the playback experience of the content to be distributed. Therefore, how to design a content distribution method to provide services for users based on the above defects has become a problem to be solved. SUMMARY

[0004] The present application provides a content distribution method, device, electronic equipment and storage medium, which can accurately predict the content of interest to the user at a certain time window in the future, so that the prediction is more accurate and unitized, and the network congestion can be improved and the user access response speed and content hit rate can be improved.

[0005] In a first aspect, the present application provides a content distribution method, which comprises:

[0006] receiving an operation of triggering a content distribution function by a user, and obtaining a historical edge node accessed by the user;

[0007] constructing an interest propagation model for a content distribution network (CDN) based on the historical edge node; the interest propagation model is determined based on a center node, a time window node and an edge node, and the center node and the edge node are adjacent nodes of the time window node;

[0008] determine a target time window node, and predict a target edge node to be accessed by the user at the target time window node based on the interest propagation model;

[0009] determine content to be distributed corresponding to the target edge node, and distribute the content to be distributed to the user based on the target edge node.

[0010] In a second aspect, the present application provides a content distribution device, which comprises:

[0011] an information obtaining module, configured to receive an operation of triggering a content distribution function by a user, and obtain a historical edge node accessed by the user;

[0012] a model constructing module, configured to construct an interest propagation model for a content distribution network (CDN) based on the historical edge node; the interest propagation model is determined based on a center node, a time window node and an edge node, and the center node and the edge node are adjacent nodes of the time window node;

[0013] a node predicting module, configured to determine a target time window node, and predict a target edge node to be accessed by the user at the target time window node based on the interest propagation model;

[0014] a content distribution module, configured to determine content to be distributed corresponding to the target edge node, and distribute the content to be distributed to the user based on the target edge node.

[0015] In a third aspect, the present application provides an electronic device, which comprises:

[0016] at least one processor; and

[0017] a memory connected with the at least one processor; wherein,

[0018] the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the content distribution method according to any of the embodiments of the present application.

[0019] In a fourth aspect, the present application provides a computer readable storage medium, which stores computer instructions for enabling a processor to execute the content distribution method according to any of the embodiments of the present application.

[0020] The embodiment of the present application provides a content distribution method, comprising the following steps: receiving an operation of triggering a content distribution function by a user, obtaining a historical edge node accessed by the user; constructing an interest propagation model for a content distribution network (CDN) based on the historical edge node; determining a target time window node, predicting a target edge node to be accessed by the user at the target time window node based on the interest propagation model; determining content to be distributed corresponding to the target edge node, and distributing the content to be distributed to the user based on the target edge node. The present application analyzes interest preferences from the historical edge node accessed by the user, and considers the timeliness of the interest, and then constructs an interest propagation model; the present application predicts the probability of the user accessing other edge nodes at a future time window in combination with transmission cost in the CDN network, pre-distributes the content to the target edge node with the highest interest value, and provides content distribution services to the user through the target edge node. The present application adds a time window node in the interest propagation model, can accurately predict the content interested by the user at a future time window, makes the prediction more accurate and unitized, and can improve network congestion and improve user access response speed and content hit rate. The content distribution method of the present application optimally utilizes resources compared with the current general nearest access method, and intelligently schedules and allocates resources.

[0021] It should be noted that the computer instructions described above can be stored in whole or in part on a computer readable storage medium. The computer readable storage medium can be packaged together with the processor of the content distribution device, or can be packaged separately from the processor of the content distribution device, and the present application does not limit this.

[0022] The description of the second aspect, the third aspect and the fourth aspect in the present application can refer to the detailed description of the first aspect; and the beneficial effects of the description of the second aspect, the third aspect and the fourth aspect can refer to the beneficial effect analysis of the first aspect, which will not be repeated here.

[0023] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor to limit the scope of the present application. Other features of the present application will become apparent through the following description.

[0024] It can be understood that, before using the technical solutions disclosed in the embodiments of the present application, the type, use range and use scene of the personal information involved in the present application should be informed to the user and the authorization of the user should be obtained through appropriate means according to relevant laws and regulations. BRIEF DESCRIPTION OF DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments description. Obviously, the drawings described in the following embodiments are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0026] Figure 1 The first flow diagram of a content distribution method provided by the embodiments of the present application;

[0027] Figure 2 The schematic diagram of an interest propagation model provided by the embodiments of the present application;

[0028] Figure 3A The schematic diagram of a first propagation path provided by the embodiments of the present application;

[0029] Figure 3B The schematic diagram of a second propagation path provided by the embodiments of the present application;

[0030] Figure 3C The schematic diagram of a third propagation path provided by the embodiments of the present application;

[0031] Figure 4 The second flow diagram of a content distribution method provided by the embodiments of the present application;

[0032] Figure 5 The structural schematic diagram of a content distribution apparatus provided by the embodiments of the present application;

[0033] Figure 6 The block diagram of an electronic device for implementing a content distribution method provided by the embodiments of the present application. DETAILED DESCRIPTION

[0034] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will combine the drawings in the embodiments of the present application to make a clear and complete description of the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort should belong to the protection scope of the present application.

[0035] It should be noted that the terms "first", "second", "target", and "original" and the like in the description and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include", "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0036] Figure 1 A first flowchart of a content distribution method provided by an embodiment of the present application, the embodiment can be applicable to the case of determining a target edge node to be accessed by a user based on an interest propagation model, and then distributing content based on the target edge node. The content distribution method provided by the embodiment of the present application can be executed by a content distribution device provided by the embodiment of the present application, which can be realized by software and / or hardware and integrated in an electronic device executing the method. The overall process of the content distribution of the present application can be divided into four levels from bottom to top, including a preprocessing layer, a clustering layer, a model layer and a decision layer.

[0037] Referring to Figure 1 The method of the embodiment includes but is not limited to the following steps:

[0038] S110, receiving an operation of triggering a content distribution function by a user, and acquiring a historical edge node accessed by the user.

[0039] The log of the user accessing certain application software is stored in the form of text in the historical edge node, including attribute information of the video, webpage or product browsed by the user, on-demand, review, collection, rating, access time and other interactive behavior record information (i.e. text data).

[0040] In the embodiment of the present application, this step is the preprocessing layer in the overall process of content distribution. When the user opens the content distribution function on the electronic device or the user agrees to receive the content distribution, the preprocessing layer acquires which historical edge nodes are accessed by the user, and acquires the text data accessed by the user from the historical edge nodes. The preprocessing layer then pre-processes the text data of the user by data cleaning or normalization to obtain standard data.

[0041] S120, constructing an interest propagation model for a content distribution network (CDN) based on the historical edge node.

[0042] The interest propagation model is determined based on a center node, a time window node and an edge node, the center node and the edge node being adjacent nodes of the time window node.

[0043] The application considers that the interest of a user has a time limit, adds a time window node in the interest propagation model, avoids that the content to be distributed by the edge node is not the content interested by the user, can accurately predict the content interested by the user in a certain time window in the future, makes the prediction more accurate and unitized, and thus solves the problems of network congestion, slow response speed of user access and low content hit rate.

[0044] In the embodiment of the application, the step is a clustering layer in the overall flow of content distribution, and is used for analyzing the standard data obtained by the preprocessing layer, and then constructing the interest propagation model. Specifically, first, after the text data in the edge node is preprocessed by data cleaning or normalization and the like to obtain standard data, the standard data is analyzed by clustering to obtain a plurality of center nodes corresponding to the text data; optionally, the K-means algorithm can be used for clustering analysis. For example, the determination process of the center node can be as follows: the text content of the standard data is clustered, and the number of center nodes can be determined by specifying the K value. The distance between all objects in the standard data and the K center nodes is calculated, the center node with the minimum distance is selected, and is added to the area represented by the center node, until all objects are allocated to the K center nodes. Preferably, the new clustering center of each center node can be recalculated, and the distance sum at this time is calculated. It is judged whether the change of the distance sum of two consecutive iterations is greater than a set value, if yes, the iteration is continued, if not, the iteration is exited, the algorithm is ended, and the final clustering result is output.

[0045] Then, the videos, webpages or products browsed by the user are divided into a plurality of clusters (i.e. center nodes) according to certain themes by clustering analysis, the theme words of the text data of each center node are extracted, and the theme words are respectively stored in the edge nodes.

[0046] Finally, as shown in FIG. 1, the interest propagation model is obtained. Figure 2 As shown in FIG. 2, it is a schematic diagram of the interest propagation model, wherein C1 and C2 represent center nodes, S 1,1 , S 1,2 , S 2,1 and S 2,2 represent time window nodes, and N1, N2, N3 and N4 represent edge nodes. The center nodes, the time window nodes and the edge nodes are taken as vertices of the interest propagation model. The corresponding weight values are set for the vertices, and thus the interest propagation model is obtained. Optionally, one day can be divided into T time windows in units of hours.

[0047] Further, the edge node is connected to the center node and the time window node by an edge E C,SThis represents the weight value propagated from the central node to the time window node, affecting E. C,S Key attributes include: the total time within the time window t, the total number of visits, and the total number of visitors. The proportion of the sum of these three key attributes to the total population is used as E. C,S The value; using E S,C E represents the weight value propagated from the time window node to the central node. S,N E represents the weight value propagated from the time window node to the edge node. N,S This represents the weight value propagated from the edge node to the time window node. Optional, E can be adjusted and compared. S,C E S,N and E N,S All are initialized to 1. (Use E) N,N This represents the weight value for propagation from edge node 1 to edge node 2. Since the probability of edge node 1 obtaining resources from the central node is affected by the cost of the propagation from edge node 1 to edge node 2, the goal is to minimize the overall cost. Therefore, the higher the cost from edge node 1 to edge node 2, the lower the probability of selecting edge node 2 to provide services. Hence, E is set to... N,N Transmission overhead needs to be considered. The transmission overhead function indicates that if the content accessed by the user is not stored in this edge node, then access requests for that content will be provided by other edge nodes.

[0048] S130. Determine the target time window node and predict the target edge node that the user will visit at the target time window node based on the interest propagation model.

[0049] In this embodiment, this step is the model layer in the overall content distribution process. The idea behind CDN distribution based on the interest propagation model is as follows: Given a target center node and a target time window node, the interest propagation model sets the edge weights according to the target time window node, and then initializes the interest value of the target center node. After initializing the interest value, the interest propagates through different paths. After several propagations, the edge node receives the interest propagated from different paths. During the interest propagation process, the overhead between the time window node and the edge node can change the interest value. Finally, the interest value of the edge node is obtained based on the sum of the interests propagated from different paths, thereby determining the target edge node that the user should access.

[0050] Specifically, the target edge node to be accessed by the user at the target time window node is predicted based on the interest propagation model, including: first, determining the candidate center node as the target center node corresponding to the target time window node one by one; further, the determination process of the candidate center node is: determining a plurality of adjacent nodes (the adjacent nodes include the center node and the edge node) corresponding to the target time window node; determining the node in the plurality of adjacent nodes which does not belong to the historical edge node and does not belong to the global node set as the candidate center node. Further, the determination process of the global node set is: traversing the nodes (the nodes include the center node, the time window node and the edge node) in the interest propagation model, judging whether each node belongs to the preset time window node corresponding to the other center node and the edge node adjacent to at least one time window node of the historical center node one by one; if not, the node is added to the global node set, wherein the other center node is the center node in the interest propagation model except the historical center node; the nodes include the center node, the time window node and the edge node.

[0051] Then, the candidate edge node adjacent to the target time window node and satisfying the loss standard is determined, the loss standard is that the candidate edge node is adjacent to the edge node of at least one time window node of the historical center node, and the historical center node is the center node corresponding to the historical edge node in the interest propagation model. The benefit of setting the loss standard in the present application is that the transmission consumption cost of the edge node can be controlled, which embodies the principle of accessing nearby.

[0052] Finally, the initial interest value is assigned to the target center node, and the propagation path meeting the propagation rule is selected based on the interest propagation model; the initial interest value is propagated based on the propagation path to obtain the final interest value of the candidate edge node, and the candidate edge node corresponding to the final interest value with the largest value is taken as the target edge node.

[0053] S140, determining the content to be distributed corresponding to the target edge node, and distributing the content to be distributed to the user based on the target edge node.

[0054] In the embodiment of the present application, this step is the decision layer in the overall process of content distribution. After the target edge node to be accessed is determined through the above steps, the topic to be accessed is obtained from the target edge node, the corresponding content to be distributed is determined based on the topic to be accessed, the content distribution service is provided to the user through the target edge node, and the content to be distributed is distributed to the user.

[0055] The technical scheme provided by the embodiment comprises the following steps: receiving an operation of triggering a content distribution function by a user, and obtaining a historical edge node accessed by the user; constructing an interest propagation model for a CDN based on the historical edge node; determining a target time window node, and predicting a target edge node to be accessed by the user at the target time window node based on the interest propagation model; determining content to be distributed corresponding to the target edge node, and distributing the content to be distributed to the user based on the target edge node. The application analyzes interest preferences from historical edge nodes accessed by the user, and considers the timeliness of the interest, and then constructs an interest propagation model. The application predicts the probability of the user accessing other edge nodes at a future time window in combination with transmission cost in the CDN network, and pre-distributes content to a target edge node with the highest interest value, so as to provide content distribution services to the user through the target edge node. The application adds a time window node in the interest propagation model, can accurately predict content interested by the user at a future time window, makes the prediction more accurate and unitized, and can improve network congestion and improve user access response speed and content hit rate. The content distribution method of the application optimally utilizes resources through intelligent scheduling and allocation of resources compared with the current general nearest access method.

[0056] There are many possible paths between the center node and the edge node. Enumerating all the propagation paths is very time-consuming and reduces the hit rate of the cache. Selecting a suitable propagation path that meets the propagation rule from all possible paths can improve the hit rate and efficiency of the cache. The interest always propagates from the center node, then reaches a time window node, and then reaches an edge node adjacent to the time window node, which indicates that the user of the center node accesses the edge node at the time window. The propagation rule can be: (1) the same propagation path does not contain repeated nodes, and if the propagation path passes through repeated nodes, it may cause a dead loop in the propagation process. (2) the propagation path only contains an edge node and a time window node that have been accessed in a region, and if the propagation path contains multiple edge nodes or time window nodes that have been accessed in a region, the length of the path will be very long. (3) the propagation path is terminated when it encounters an edge node that has not been accessed. In order to select a propagation path that meets the above propagation rule, a relatively intuitive method can be used, such as using a breadth-first search algorithm to traverse the interest propagation model.

[0057] In summary, the propagation path in the above step S130 comprises a first propagation path, a second propagation path and a third propagation path, and the first propagation path is determined by connecting the first center node, the first time window node, the first edge node and the second edge node into the first propagation path; the first time window node is a time window node corresponding to the first center node, the first edge node is an edge node corresponding to the first time window node, and the second edge node is an edge node adjacent to the first edge node and not accessed. For example, Figure 3AThis is a schematic diagram of the first propagation path. As can be seen from the diagram, the first propagation path consists of the first central node C1 and the first time window node S. 1,2 The first edge node N2 and the second edge node N3 are connected, as shown in the propagation path abc in the figure.

[0058] The second propagation path is determined as follows: the first center node, the first time window node, the first edge node, the third time window node, and the third edge node are connected to form the second propagation path; the third time window node is the time window node corresponding to another center node adjacent to the first edge node, and the third edge node is the edge node corresponding to the third time window node. For example... Figure 3B This is a schematic diagram of the second propagation path. As can be seen from the diagram, the second propagation path consists of the first central node C1 and the first time window node S. 1,2 First edge node N2, third time window node S 2,1 It is connected to the third edge node N3, as shown in the propagation line abcd in the figure.

[0059] The third propagation path is determined as follows: the first central node, the first time window node, the first edge node, the third time window node, the third central node, the fourth time window node, and the fourth edge node are connected to form the third propagation path; the third central node is the central node corresponding to the third time window node, the fourth time window node is the time window node other than the third time window node corresponding to the third central node, and the fourth edge node is the edge node corresponding to the fourth time window node. For example... Figure 3C This is a schematic diagram of the third propagation path. As can be seen from the diagram, the third propagation path consists of the first central node C1 and the first time window node S. 1,2 First edge node N2, third time window node S 2,1 Third central node C2, fourth time window node S 2,2 It is connected to the fourth edge node N4, as shown in the propagation path abcdef in the figure.

[0060] The content distribution method provided in the embodiments of this application is further described below. Figure 4 This is a schematic diagram of the second process of a content distribution method provided in an embodiment of this application. This embodiment is an optimization based on the above embodiment, specifically optimized as follows: the process for determining the final interest value of candidate edge nodes in this embodiment (i.e., the above...) Figure 1 The specific implementation steps of S130 in the corresponding embodiment will be explained in detail.

[0061] Before introducing the determination process of the final interest value of the candidate edge node, the defects of the existing breadth-first search algorithm need to be introduced. If a node is allowed to be accessed multiple times, the existing method will cause the interest return problem, which will cause the interest value of the edge node to be incorrect, and further cause the target edge node to be selected inaccurately. Based on the defect, the subtraction backtracking mechanism is introduced to optimize the breadth-first search algorithm, which can solve the interest return problem caused by the multiple access of a node. The edge node with the maximum interest value is selected to provide services for the user, which not only optimizes the network performance, but also improves the application efficiency and integrates the content resources. The following is the introduction of the breadth-first search algorithm of the subtraction backtracking mechanism applied to the interest propagation model.

[0062] Referring to Figure 4 The method of the embodiment includes but is not limited to the following steps:

[0063] S210, determining a first interest value corresponding to a second node according to a weight value of the first node propagated to the second node and an interest value of the first node.

[0064] The first node and the second node are nodes in a propagation path, the first node is a center node, a time window node or an edge node, and the second node is a center node, a time window node or an edge node. The second node is a neighboring node of the first node.

[0065] In the embodiment of the application, interest is propagated from the first node v to its neighbor node, the second node v', and the interest value corresponding to the second node is r v ·w v,v' , wherein r v is the interest value of the first node, and w v,v' is the weight value of the first node propagated to the second node. If the breadth-first search algorithm is used and the node is allowed to be accessed multiple times, the second node v' collects the interest propagated from different paths of all its neighbor nodes at the same time, and the first interest value corresponding to the second node is , wherein r v' is the first interest value of the second node, r v″ is the interest value of the neighboring node of the second node, w v″,v' is the weight value of the neighboring node of the second node propagated to the second node, and A v' represents the set of the neighboring nodes of the second node. It should be noted that the first node v is included in A v' .

[0066] S220, determining a second interest value corresponding to the interest return of the second node, and determining the interest value corresponding to the second node based on the first interest value and the second interest value.

[0067] In this embodiment, when the second node v' performs the next propagation, it will propagate the collected first interest value to its neighbor node A. v' And one of the neighboring nodes A v' It will be the first node v, which will cause the interest backhaul problem.

[0068] To avoid partial interest backpropagation, interest is propagated from the second node v' to its neighbor node A. v' Previously, it was necessary to start from r v' Subtract the interest r that propagates from the first node v to the second node v' v .w v,v' The interest value corresponding to the second node v' is obtained, thus avoiding duplicate nodes in the same propagation path due to partial interest back transmission. The interest value corresponding to the second node v' is calculated using the following formula (1):

[0069]

[0070] In the formula, v″ is the neighboring node of the second node; A v' Let r be the set of adjacent nodes of the second node; v' r is the first interest value of the second node. v″ The interest values ​​of the neighboring nodes of the second node; w v″,v' It is the weight value propagated from the neighboring nodes of the second node to the second node; r v The interest value of the first node; w v,v' It is the weight value propagated from the first node to the second node; w v',v It is the weight value propagated from the second node to the first node.

[0071] The breadth-first search algorithm uses a subtraction backtracking mechanism and allows nodes to be visited multiple times. After the interest propagates from the first node v to its neighbor node v', the interest propagated from the first node v to the second node v' should be subtracted from the sum of the interests propagated from all the neighbors of the second node. This can avoid some interest being propagated back, and then the next propagation can be carried out.

[0072] S230. Take the second node as the new first node, take the next node of the second node in the propagation path as the new second node, and repeatedly execute the operation of determining the first interest value corresponding to the second node based on the weight value propagated from the first node to the second node and the interest value of the first node, until the interest value of the last node in the propagation path is determined.

[0073] In this embodiment, the last node in the propagation path is a candidate edge node. After propagating all nodes in the propagation path according to steps S210-S220 above, the final interest value of the candidate edge node can be obtained.

[0074] The technical scheme provided by the embodiment determines a first interest value corresponding to a second node according to a weight value propagated from the first node to the second node and an interest value of the first node; determines a second interest value corresponding to interest back transmission of the second node, and determines an interest value corresponding to the second node based on the first interest value and the second interest value; takes the second node as a new first node, takes a next node of the second node in a propagation path as a new second node, and repeatedly performs the operation of determining the first interest value corresponding to the second node according to the weight value propagated from the first node to the second node and the interest value of the first node until the interest value of the last node in the propagation path is determined. An interest propagation algorithm based on breadth-first search is proposed in the application, interest is propagated through nodes, and a subtraction backtracking mechanism is introduced, which solves the problem of interest back transmission caused by allowing a node to be accessed multiple times, pre-distributes content to the edge with the highest interest value, optimizes the performance and bandwidth utilization of the system, and can improve network congestion, improve user access response speed and content hit rate, and improve the experience of user access requests.

[0075] Figure 5 A structural schematic diagram of a content distribution device provided by the embodiment of the application is shown in FIG. 5, which can include: Figure 5

[0076] An information acquisition module 510 is configured to receive an operation of triggering a content distribution function by a user, and acquire a historical edge node accessed by the user.

[0077] A model construction module 520 is configured to construct an interest propagation model for a content distribution network (CDN) based on the historical edge node; the interest propagation model is determined based on a center node, a time window node and an edge node, and the center node and the edge node are adjacent nodes of the time window node.

[0078] A node prediction module 530 is configured to determine a target time window node, and predict a target edge node to be accessed by the user at the target time window node based on the interest propagation model.

[0079] A content distribution module 540 is configured to determine to-be-distributed content corresponding to the target edge node, and distribute the to-be-distributed content to the user based on the target edge node.

[0080] ​Further, the node prediction module 530 can be specifically configured to: determine a target center node corresponding to the target time window node from the candidate center nodes; determine a candidate edge node adjacent to the target time window node and satisfying a loss criterion, the loss criterion being that the candidate edge node is adjacent to an edge node of at least one time window node of a historical center node, the historical center node being a center node corresponding to the historical edge node in the interest propagation model; assign an initial interest value to the target center node, and select a propagation path meeting a propagation rule based on the interest propagation model; propagate the initial interest value based on the propagation path to obtain a final interest value of the candidate edge node, and take the candidate edge node corresponding to the final interest value with the largest value as the target edge node.

[0081] Optionally, the propagation path includes a first propagation path, a second propagation path, and a third propagation path, and the propagation path is determined by: connecting a first center node, a first time window node, a first edge node, and a second edge node into the first propagation path, the first time window node being a time window node corresponding to the first center node, the first edge node being an edge node corresponding to the first time window node, and the second edge node being an edge node adjacent to the first edge node and not visited; connecting the first center node, the first time window node, the first edge node, a third time window node, and a third edge node into the second propagation path, the third time window node being a time window node corresponding to another center node adjacent to the first edge node, and the third edge node being an edge node corresponding to the third time window node; and connecting the first center node, the first time window node, the first edge node, the third time window node, a third center node, a fourth time window node, and a fourth edge node into the third propagation path, the third center node being a center node corresponding to the third time window node, the fourth time window node being a time window node corresponding to the third center node and other than the third time window node, and the fourth edge node being an edge node corresponding to the fourth time window node.

[0082] Further, the node prediction module 530 can be further configured to: determine a first interest value of a second node according to a weight value propagated from a first node to the second node and an interest value of the first node; determine a second interest value of the second node corresponding to an interest backhaul, and determine an interest value of the second node based on the first interest value and the second interest value; take the second node as a new first node, and take a next node of the second node in the propagation path as a new second node, and repeat the operation of determining the first interest value of the second node according to the weight value propagated from the first node to the second node and the interest value of the first node until the interest value of the last node in the propagation path is determined; wherein the first node and the second node are nodes in the propagation path, the last node in the propagation path is the candidate edge node, and the first node and the second node are the center node, the time window node or the edge node.

[0083] Optionally, the candidate center node is determined by: determining a plurality of adjacent nodes corresponding to the target time window node; and determining a node in the plurality of adjacent nodes that does not belong to the historical edge node and does not belong to a global node set as the candidate center node.

[0084] Optionally, the global node set is determined by: traversing nodes in the interest propagation model, determining whether the nodes belong to a preset time window node corresponding to other center nodes and the preset time window node is adjacent to an edge node of at least one time window node of the historical center node, the other center nodes being center nodes in the interest propagation model except the historical center node; the nodes including the center node, the time window node and the edge node; and adding the nodes to the global node set if the nodes do not belong to the preset time window node corresponding to the other center nodes and the preset time window node is not adjacent to the edge node of the at least one time window node of the historical center node.

[0085] Further, the model construction module 520 can be further configured to: perform cluster analysis on text data in the historical edge node to obtain a plurality of center nodes corresponding to the text data; extract a subject word of the text data of each center node, and store the subject word in an edge node respectively; take the center node, the time window node and the edge node as vertices of the interest propagation model, and set corresponding weight values for the vertices, thereby obtaining the interest propagation model.

[0086] The content distribution device provided in the embodiment can be applied to the content distribution method provided in any of the above embodiments, and has corresponding functions and advantages.

[0087] Figure 6is a block diagram of an electronic device that implements a content distribution method according to an embodiment of the present application. The electronic device 10 is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular telephones, smart phones, wearable devices (e.g., headsets, glasses, watches, etc.), and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit the implementations of the present application described and / or claimed in this document.

[0088] As shown in Figure 6 The electronic device 10 includes at least one processor 11, and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., connected to the at least one processor 11, where the memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer programs stored in the read-only memory (ROM) 12 or loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0089] Various components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc., an output unit 17, such as various types of displays, speakers, etc., a storage unit 18, such as a magnetic disk, an optical disk, etc., and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0090] The processor 11 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs various methods and processes described above, such as the content distribution method.

[0091] In some embodiments, the content distribution method can be implemented as a computer program tangibly embodied in a computer readable storage medium, e.g., storage unit 18. In some embodiments, parts or all of the computer program can be loaded and / or installed onto electronic device 10 via, e.g., ROM 12 and / or communication unit 19. When the computer program is loaded onto RAM 13 and executed by processor 11, one or more steps of the content distribution method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the content distribution method by way of other means (e.g., by way of firmware).

[0092] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, specially designed application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0093] Computer programs implementing methods of the present application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the computer program, when executed, can implement the functions / acts specified in the flowcharts and / or block diagrams. The computer program can be executed entirely on a machine, partially on a machine, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0094] In the context of this application, a computer readable storage medium can be a tangible medium that can contain or store computer programs for use by or in connection with an instruction execution system, apparatus, or device. Computer readable storage media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer readable storage medium can be a machine readable signal medium. More specific examples of the machine readable storage medium will include a one or more lines of a electrical connection, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0095] To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0096] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), blockchain network, and the Internet.

[0097] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, to solve the defects of large management difficulty and weak business scalability in traditional physical host and VPS service.

[0098] It should be understood that the various forms of flow shown above can be reordered, added to, or have steps deleted. For example, the steps described in this application can be performed in parallel, in series, or in a different order, as long as the desired results of the technical solutions of this application can be achieved, and this application does not limit herein.

[0099] The above detailed description does not constitute a limitation on the scope of protection of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A content distribution method characterized by, The method comprises: receiving an operation of triggering a content distribution function by a user, obtaining a historical edge node accessed by the user; constructing an interest propagation model for a content distribution network (CDN) based on the historical edge node; the interest propagation model is determined based on a center node, a time window node and an edge node, and the center node and the edge node are adjacent nodes of the time window node; determining a target time window node, predicting a target edge node to be accessed by the user at the target time window node based on the interest propagation model; determining a to-be-distributed content corresponding to the target edge node, and distributing the to-be-distributed content to the user based on the target edge node; wherein the constructing of the interest propagation model for the CDN based on the historical edge node comprises: performing cluster analysis on text data in the historical edge node to obtain a plurality of center nodes corresponding to the text data; extracting a subject word of the text data of each center node, and storing the subject word in an edge node respectively; taking the center node, the time window node and the edge node as vertices of the interest propagation model, and setting corresponding weight values for the vertices, thereby obtaining the interest propagation model.

2. The content distribution method according to claim 1, characterized by, The predicting of the target edge node to be accessed by the user at the target time window node based on the interest propagation model comprises: determining a target center node corresponding to the target time window node from candidate center nodes; determining a candidate edge node adjacent to the target time window node and satisfying a loss standard, the loss standard being that the candidate edge node is adjacent to an edge node of at least one time window node of a historical center node, and the historical center node being a center node corresponding to the historical edge node in the interest propagation model; assigning an initial interest value to the target center node, and selecting a propagation path meeting a propagation rule based on the interest propagation model; propagating the initial interest value based on the propagation path to obtain a final interest value of the candidate edge node, and taking a candidate edge node corresponding to a final interest value with the largest value as the target edge node.

3. The content distribution method according to claim 2, characterized by, The propagation path comprises a first propagation path, a second propagation path and a third propagation path, and the propagation path is determined by the following method: connecting a first center node, a first time window node, a first edge node and a second edge node into the first propagation path; the first time window node is a time window node corresponding to the first center node, the first edge node is an edge node corresponding to the first time window node, and the second edge node is an edge node adjacent to the first edge node and not accessed; connecting the first center node, the first time window node, the first edge node, a third time window node and a third edge node into the second propagation path; the third time window node is a time window node corresponding to another center node adjacent to the first edge node, and the third edge node is an edge node corresponding to the third time window node; The first center node, the first time window node, the first edge node, the third time window node, the third center node, the fourth time window node and the fourth edge node are connected into the third propagation path; the third center node is a center node corresponding to the third time window node, the fourth time window node is a time window node corresponding to the third center node except the third time window node, and the fourth edge node is an edge node corresponding to the fourth time window node.

4. The content distribution method according to claim 3, characterized by, The propagating of the initial interest value based on the propagation path comprises: determining a first interest value corresponding to a second node according to a weight value propagated from a first node to the second node and an interest value of the first node; determining a second interest value corresponding to interest backhaul of the second node, and determining an interest value corresponding to the second node based on the first interest value and the second interest value; repeating the operation of determining the first interest value corresponding to the second node according to the weight value propagated from the first node to the second node and the interest value of the first node, until an interest value of a last node in the propagation path is determined, wherein the first node and the second node are nodes in the propagation path, the last node in the propagation path is the candidate edge node, the first node is the center node, the time window node or the edge node, and the second node is the center node, the time window node or the edge node. The candidate center node is determined in the following manner:

5. The content distribution method according to claim 2, characterized by, determining a plurality of adjacent nodes corresponding to the target time window node; determining a node in the plurality of adjacent nodes which does not belong to the historical edge node and does not belong to a global node set as a candidate center node. The global node set is determined in the following manner:

6. The content distribution method according to claim 5, characterized by, traversing nodes in the interest propagation model, judging whether the nodes belong to a preset time window node corresponding to other center nodes and the preset time window node is adjacent to an edge node of at least one time window node of the historical center node, the other center nodes being center nodes in the interest propagation model except the historical center node; the nodes including center nodes, time window nodes and edge nodes; if not, adding the nodes to the global node set. The device comprises:

7. A content distribution apparatus characterized by comprising: an information acquisition module configured to receive an operation of triggering a content distribution function by a user, and acquire a historical edge node accessed by the user; a model construction module configured to construct an interest propagation model for a content distribution network (CDN) based on the historical edge node; the interest propagation model is determined based on center nodes, time window nodes and edge nodes, the center nodes and the edge nodes being adjacent nodes of the time window nodes; a node prediction module configured to determine a target time window node, and predict a target edge node to be accessed by the user at the target time window node based on the interest propagation model. ​ The content distribution module is configured to determine content to be distributed corresponding to the target edge node, and distribute the content to be distributed to the user based on the target edge node. The model construction module is specifically configured to perform clustering analysis on the text data in the historical edge nodes to obtain a plurality of center nodes corresponding to the text data; extract a keyword of a theme of the text data of each center node, and store the keyword in an edge node; and set the center nodes, the time window nodes and the edge nodes as vertices of the interest propagation model, and set corresponding weight values for the vertices, so as to obtain the interest propagation model.

8. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory connected to the at least one processor in communication; wherein The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the content distribution method of any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for enabling the processor to implement the content distribution method of any one of claims 1 to 6 when executed.

Images