Cache and routing policy optimization and network content distribution methods, apparatuses, and devices

Abstract

The application discloses a kind of cache and routing strategy optimization and network content distribution method, device and equipment.The method comprises: constructing network content popularity model and network structure model;Maximum traffic offloading model under network popularity model and network structure model is constructed, and maximum traffic offloading model is used to convert network resource allocation task into offloading network content request traffic task;According to the traffic of each network content request in traffic offloading model under the offloading of cache strategy and routing strategy is solved;According to the offloading traffic solution result, cache strategy and routing strategy are optimized, to realize the network content distribution that can be based on cache strategy and routing strategy to maximize the traffic of offloading content request.This scheme can realize the optimization of cache and routing strategy for network content distribution under the collaborative work of cloud, edge node and terminal equipment, reduce the traffic load pressure of core network, improve the service quality of network content request.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method, apparatus, and device for optimizing caching and routing strategies and distributing network content. Background Technology

[0002] The explosive growth of mobile internet traffic due to the rise of emerging network application technologies (such as virtual reality and augmented reality) has placed higher service demands on next-generation wireless networks, such as lower network transmission latency, higher link bandwidth capacity, and support for large-scale user access.

[0003] The centralized service model of mobile cloud computing (MCC) is difficult to cope with the above challenges. Improving content delivery and ensuring network service quality (QoS) by effectively allocating heterogeneous network resources is a key issue that needs to be addressed in future wireless networks.

[0004] Mobile edge computing (MEC) integrates the computing and caching capabilities at the edge of mobile networks to provide flexible connectivity, real-time response, and traffic offloading to meet user service needs. However, due to the heterogeneity and limitations of edge resources, it cannot effectively distribute network content and generates significant traffic load pressure. This makes it difficult for MEC to effectively improve content delivery and meet the differentiated service needs of massive mobile users, thus reducing the quality of service for network content requests. Summary of the Invention

[0005] This invention provides a method, apparatus, and electronic device for optimizing caching and routing strategies and distributing network content, which can optimize caching and routing strategies used for network content distribution, reduce the traffic load pressure on the core network, and improve the quality of service for network content requests.

[0006] According to one aspect of the present invention, a caching and routing strategy optimization method is provided, comprising:

[0007] Construct a network content popularity model and a network structure model. The network structure model includes terminals, base stations, and the cloud. Base stations are connected to the cloud through the core network, and each base station has a terminal attached to it.

[0008] A maximized traffic offloading model is constructed based on a network popularity model and a network structure model. The maximized traffic offloading model is used to transform the network resource allocation task into an offloading network content request traffic task.

[0009] Based on caching and routing strategies, the traffic offloaded by each network content request under the maximized traffic offload model is calculated.

[0010] The caching and routing strategies are optimized based on the results of the unloading traffic calculation to maximize the network content distribution optimization by maximizing the traffic of unloading content requests.

[0011] According to one aspect of the present invention, another network content distribution method is provided, employing a caching strategy and a routing strategy obtained by any of the caching and routing strategy optimization methods described in this embodiment, the method comprising:

[0012] Load the optimized caching and routing strategies;

[0013] The network content is distributed using the cache and routing results indicated by the optimized caching and routing strategies.

[0014] According to another aspect of the present invention, a caching and routing policy optimization apparatus is provided, comprising:

[0015] The network model construction module is used to build a network content popularity model and a network structure model. The network structure model includes terminals, base stations, and the cloud. The base stations are connected to the cloud through the core network, and each base station has a terminal attached to it.

[0016] The unloading model construction module is used to construct a maximized traffic unloading model under the network popularity model and network structure model. The maximized traffic unloading model is used to transform the network resource allocation task into an unloading network content request traffic task.

[0017] The unloading traffic determination module is used to calculate the unloading traffic of each network content request under the maximized traffic unloading model based on caching and routing strategies.

[0018] The strategy optimization module is used to optimize the caching strategy and routing strategy based on the offload traffic solution results, so as to maximize the network content distribution optimization by maximizing the traffic of offloaded content requests.

[0019] According to another aspect of the present invention, a network content distribution apparatus is provided, employing a caching strategy and a routing strategy obtained by any of the caching and routing strategy optimization methods described in this embodiment, the apparatus comprising:

[0020] The loading module is used to load the optimized caching and routing strategies;

[0021] The distribution module is used to distribute network content using the cache and routing results indicated by the optimized caching and routing strategies, in order to maximize the unloading of content request traffic.

[0022] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0023] At least one processor; and

[0024] A memory communicatively connected to the at least one processor; wherein,

[0025] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the caching and routing strategy optimization or network content distribution method according to any embodiment of the present invention.

[0026] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the caching and routing strategy optimization or network content distribution method according to any embodiment of the present invention.

[0027] The technical solution of this invention constructs a network content popularity model and a network structure model. The network structure model includes terminals, base stations, and the cloud. The base stations are connected to the cloud through the core network, and each base station hosts a terminal. A maximum traffic offloading model is constructed under the network popularity model and the network structure model. The maximum traffic offloading model is used to convert the network resource allocation task into an offloading task of network content request traffic. The traffic offloaded by each network content request under the traffic offloading model is solved according to the caching strategy and the routing strategy. The caching strategy and the routing strategy are optimized according to the offloading traffic solution to optimize network content distribution by maximizing the offloading of content request traffic. The network traffic offloading problem under the cloud-edge-device collaboration condition is introduced. The caching and routing strategies used for network content distribution are optimized under the collaborative work of the cloud, edge nodes, and terminal devices. The optimized caching and routing strategies can greatly alleviate the pressure on the mobile core network, effectively reduce redundant transmission within the network and traffic load within and between network domains, and further improve the content distribution efficiency of the entire network.

[0028] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1This is a flowchart of a caching and routing strategy optimization method provided by an embodiment of the present invention;

[0031] Figure 2 This is a schematic diagram of a network structure model applicable to an embodiment of the present invention.

[0032] Figure 3 This is a schematic diagram illustrating the traffic offloading performance of different strategies applied according to embodiments of the present invention under different cache capacities.

[0033] Figure 4 This is a schematic diagram illustrating the traffic offloading performance of different strategies applied according to embodiments of the present invention under different content popularity levels;

[0034] Figure 5 This is a schematic diagram illustrating the traffic offloading performance of different strategies applied according to embodiments of the present invention under different content categories.

[0035] Figure 6 This is a schematic diagram illustrating the traffic offloading performance of different strategies applied according to embodiments of the present invention under different request arrival rates;

[0036] Figure 7 This is a schematic diagram of the convergence curves of the DQN reward function under different cache capacities applicable according to embodiments of the present invention;

[0037] Figure 8 This is a flowchart of a network content distribution method provided according to an embodiment of the present invention;

[0038] Figure 9 This is a schematic diagram of a caching and routing strategy optimization device according to an embodiment of the present invention;

[0039] Figure 10 This is a schematic diagram of the structure of a network content distribution device according to an embodiment of the present invention;

[0040] Figure 11 This is a schematic diagram of the structure of an electronic device that implements the caching and routing strategy optimization method or the network content distribution method of the present invention. Detailed Implementation

[0041] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0042] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0043] In some initial trials, a multi-layered edge network node collaboration framework and edge-end collaboration strategy were proposed, which can alleviate core network traffic pressure, reduce network content transmission latency, and respond promptly to most content requests from mobile users. In time-varying and complex environments, artificial intelligence technology can be applied to mobile edge computing (MEC) to further improve network processing performance. Simultaneously, reinforcement learning was proposed to be applied to the MEC framework to optimize resource allocation in the wireless system and promote content delivery in MEC-assisted networks. However, given the heterogeneity and limitations of edge resources, MEC struggles to effectively improve content delivery and meet the differentiated service needs of a massive number of mobile users.

[0044] Figure 1 This invention provides a flowchart of a caching and routing policy optimization method. This embodiment is applicable to situations where network distribution can be better performed in a collaborative manner involving the cloud, edge nodes, and terminal devices. The method can be executed by a caching and routing policy optimization device, which can be implemented in hardware and / or software and can be configured in any electronic device with network communication capabilities. Figure 1 As shown, the method may include:

[0045] S110. Construct a network content popularity model and a network structure model. The network structure model includes terminals, base stations, and the cloud. The base stations are connected to the cloud through the core network, and each base station has a terminal attached to it.

[0046] See Figure 2 To optimize network resource allocation and content distribution efficiency, this paper comprehensively considers the joint allocation of caching, computing, and communication resources in a hierarchical heterogeneous network consisting of mobile user terminals, base stations, and the cloud. The constructed network architecture model can include mobile user terminals (MCs), base stations (BSs), and the cloud. The base stations (BSs) are connected to the cloud through a complex core network, and each base station (BS) can host a number of terminal MCs.

[0047] As an optional but not limited implementation, constructing a web content popularity model can include the following processes:

[0048] A network content popularity model is constructed based on Zipf's law, which is used to describe the request probability of network content.

[0049] This case demonstrates how to design a popularity model for online content based on Zipf's law. Assume the number of content types in the system is... Available Indicates the first The probability of requesting this type of content is:

[0050]

[0051] in, Representing the The ranking of this type of online content in descending order of popularity among all online content; Zipf skewness coefficient. It represents the popularity of network content requests. The larger the value, the more concentrated the content requested by the user. This represents the number and set of content types. In the formula above, 'i' represents the index of the web content item.

[0052] As an optional but not limited implementation, constructing a network structure model can include the following processes:

[0053] Based on a layered heterogeneous network of cloud, edge, and terminal, a layered heterogeneous network structure model is constructed. In this network structure model, base stations are directly connected to each other, and the network content cached in the directly connected base stations can be transmitted to each other at high speed through optical fiber. Terminals access the base stations through wireless connection.

[0054] See Figure 2 To effectively analyze the traffic offloading problem in a cloud-edge-device collaborative heterogeneous network, a hierarchical heterogeneous network structure model was constructed. This model includes mobile user terminals (MCs), base stations (BSs), and the cloud. In the network structure model, the cloud is connected to multiple base stations (BSs) through a complex core network. However, due to channel conditions and geographical limitations, each base station (BS) and its coverage area... Each terminal MC is connected.

[0055] The network architecture model constructed in this case considers the scenario where base stations (BSs) are directly connected to each other, and resources within directly connected base stations can be transmitted at high speed via fiber optic networks. Terminals (MCs) connect wirelessly to their respective base stations (BSs). Within this network architecture model, the cloud provides all the necessary network content to the terminal MC, while both the base stations (BSs) and terminal MCs have limited cache capacity and request queue lengths. The cache capacity and request queue length of each base station (BS) can be expressed as follows: and Similarly, the first The first base station BS under the first The MC cache capacity and request queue length of each terminal can be expressed as follows: and .use and Represents a node With nodes The bandwidth capacity of the forward link and the bandwidth capacity of the backlink.

[0056] By adopting the above approach, this solution addresses existing shortcomings and proposes a network content popularity model and network structure model based on cloud-edge-device resource collaboration. This facilitates further analysis of maximizing traffic offloading under heterogeneous wireless network conditions, based on the network content popularity model and network structure model. This enables optimization of caching and routing strategies, which can then be used for network content analysis, effectively alleviating the core network traffic load pressure, reducing redundant transmission in the network, and improving the efficiency of network resource collaboration and content distribution.

[0057] S120. A maximized traffic offloading model is constructed based on the network popularity model and the network structure model. The maximized traffic offloading model is used to transform the network resource allocation task into an offloading network content request traffic task.

[0058] The maximized traffic offloading model can include a terminal traffic offloading model and a base station traffic offloading model. The constructed mobile user terminal traffic offloading model can include a local MC traffic offloading model and a non-directly connected MC traffic offloading model. The constructed base station traffic offloading model can include a local BS traffic offloading model and a directly connected BS traffic offloading model. The terminal traffic offloading model outputs network content request traffic offloaded at the terminal, and the base station traffic offloading model outputs network content request traffic offloaded at the base station. By establishing the maximized traffic offloading model, the optimal network resource allocation problem is transformed into an offloading user request traffic load problem. By constructing the maximized traffic offloading model, the allocation of network resources and the best routing path can be optimized based on the offloaded traffic, achieving better cloud-edge-device resource collaboration in network content distribution.

[0059] This paper analyzes the offloading traffic under different request processing methods and divides the traffic offloading model into mobile user terminal traffic offloading model (including local MC traffic offloading model and non-directly connected MC traffic offloading model) and base station traffic offloading model (including local BS traffic offloading model and directly connected BS traffic offloading model). The optimization objective of the optimal resource allocation problem is also given.

[0060] As an optional but not limited implementation, the maximization of traffic offloading model built on the network popularity model and network structure model may include the following steps B1-B2:

[0061] Step B1: Construct a terminal traffic offloading model for the mobile terminal. The terminal traffic offloading model includes a local terminal traffic offloading model and a non-directly connected terminal traffic offloading model. The terminal traffic offloading model is used to output the network content request traffic offloaded at the terminal.

[0062] A mobile user terminal traffic offloading model is constructed, namely the local MC traffic offloading model and the non-directly connected MC traffic offloading model.

[0063] For the local terminal MC traffic offloading model: if the first The network content requested by the terminal MC , of which The terminal MC is mounted on the If a request at a base station (BS) receives a response at the local terminal (MC), then the local terminal MC traffic offloading model can be represented as:

[0064]

[0065] in, This represents the network content requested by the m-th terminal MC mounted on the i-th base station BS. Network content request traffic that is offloaded at the local terminal (here, the m-th terminal MC). It is a Boolean variable. Indicates network content Is the cached data already mounted on the i-th base station (BS)? At each terminal MC, A value of 1 indicates that the content can be cached. A value of 0 indicates that the cache is not cached; It is a Boolean variable. Indicates network content Is it in the i-th base station BS mounted at the i-th time? In the queue of each terminal MC, a value of 1 indicates that it is in the queue, and a value of 0 indicates that it is not in the queue; Indicates the i-th base station BS mounted on the i-th base station BS. Network content at each terminal MC Request delivery rate Indicates network content The data size.

[0066] For the non-directly connected terminal MC traffic offloading model: if the network content request is not satisfied at the local terminal MC, the network content request will be routed to the next terminal according to the routing policy. At other non-directly connected terminals (MCs) under a base station (BS), if a request is made at the 1st... If a response is received at a non-directly connected terminal MC, then the non-directly connected terminal MC traffic offloading model can be expressed as:

[0067]

[0068] in, This represents the network content requested by the m-th terminal MC mounted on the i-th base station BS. In the The first base station BS Network content request traffic that is offloaded at a non-directly connected terminal. It is a Boolean variable. Indicates network content Is it already cached in the first... At the MC (Multi-Channel Controller) of a non-directly connected terminal, a value of 1 indicates that network content can be cached. A value of 0 indicates that the cache is not cached; It is a Boolean variable. Indicates network content Is it in the first In the queue of a non-directly connected terminal MC, a value of 1 indicates that it is in the queue, and a value of 0 indicates that it is not in the queue. Let be a Boolean variable, representing the network content request for network content k, the th... The MC and the first Does a non-directly connected link exist between the MCs? A value of 1 indicates the existence of a non-directly connected link, while a value of 0 indicates the absence of a non-directly connected link. Furthermore, Indicates the first The number and set of terminal MCs connected to each base station (BS).

[0069] Step B2: Construct a base station traffic offloading model, which includes a local base station traffic offloading model and a directly connected base station traffic offloading model. The base station traffic offloading model is used to output network content request traffic offloaded at the base station.

[0070] For the local base station (BS) traffic offloading model, if the first... When a base station (BS) receives a network content request for network content K from a terminal (MC) attached to it, if the network content request is responded to at the local base station (BS), then the local base station (BS) traffic offloading model can be written as:

[0071]

[0072] in, Indicates the first When a base station BS receives a network content request for network content k from a terminal MC attached to it, the local base station (referring to the first base station BS) performs the following: Network content request traffic offloaded at each base station (BS) It is a Boolean variable. Indicates network content Is it already cached in the first... At each base station (BS), a value of 1 indicates that network content can be cached. A value of 0 indicates that the cache is not cached; It is a Boolean variable. Indicates network content Is it in the first In the queue at each base station (BS), a value of 1 indicates that the data is already in the queue, while a value of 0 indicates that the data is not in the base station queue.

[0073] For the direct-connected base station (BS) traffic offloading model: if the request is not satisfied at the local BS, it will be routed to other directly connected BSs according to the routing policy. If a response is received at a directly connected BS, then the direct-connect BS traffic offloading model can be written as follows:

[0074]

[0075] in, Indicates the first When a base station (BS) receives a network content request for network content k from a terminal (MC) attached to it, at the 1st... Network content request traffic offloaded at a directly connected base station. It is a Boolean variable. Indicates network content Is it already cached in the first... At each directly connected base station (BS), a value of 1 indicates that network content can be cached. A value of 0 indicates that the cache is not cached; It is a Boolean variable. Indicates network content Is it in the first In the queue of each directly connected base station (BS), a value of 1 indicates that it is already in the queue, and a value of 0 indicates that it is not in the queue. Let be a Boolean variable, representing the first... The BS and the first Does a direct link exist between the BSs? A value of 1 indicates the existence of a direct link, while a value of 0 indicates the absence of a non-direct link. Furthermore, Indicates the relationship with the first The number and set of base stations directly connected to each BS.

[0076] To improve network resource collaboration and content distribution efficiency, this study designs a cloud-edge-device collaborative offloading scheme, which can achieve integrated allocation of resources such as caching, computing, and communication, as well as joint optimization between network caching and routing. Based on the above analysis, the goal of the optimal resource allocation problem is to maximize the offloading of user request traffic; therefore, the maximization traffic offloading model can be expressed as:

[0077]

[0078] Among them, optimization objectives Indicates the first The traffic offloaded when the m-th terminal MC attached to a base station BS requests network content k is calculated using the following formula:

[0079]

[0080] Among these constraints, - This means that the cached content of MC and BS cannot exceed their maximum cache capacity. and ; - This means that the number of content requests processed by MC and BS cannot exceed their maximum queue length. and ; This means that the link traffic between the local MC and the non-directly connected MC cannot exceed their maximum available bandwidth. ,in Represents a node With nodes Maximum available link bandwidth, Represents a node With nodes The maximum available link bandwidth; This means that the link traffic between the local BS and the directly connected BS cannot exceed its maximum available bandwidth. ; This indicates a direct connection to the BS complementary storage network to improve cache hit rate. Additionally, This represents the number and set of BSs within the network. See Table 1 for the symbols and meanings of the core parameters used in the traffic offloading model of this case, and see Figure 2 for a comparison of the characteristics of different caching strategies in this case.

[0081] Table 1

[0082]

[0083] Table 2

[0084]

[0085] Using the above approach, this case transforms the optimal resource allocation problem into the problem of unloading user request traffic load, resulting in a model that maximizes traffic unloading. Furthermore, caching and routing strategies are used to find the optimal routing path for obtaining content cache, further optimizing the system. Simulation analysis is conducted on the designed model based on caching and routing strategies, comprehensively considering various factors affecting traffic unloading: cache capacity, content popularity, number of content types, and request arrival rate. Through discussion, optimal content caching and maximized traffic unloading are achieved, thereby optimizing the caching and routing strategies.

[0086] S130. Based on the caching strategy and routing strategy, calculate the traffic offloaded by each network content request under the maximized traffic offload model.

[0087] As an optional but not limited implementation, the routing strategy can be used to describe the routing path of network content requests. The routing path is constructed in the order of local terminal, local base station, non-directly connected terminal, directly connected base station, and cloud. The local terminal is the terminal that makes the network content request, the local base station is directly connected to the local terminal, the non-directly connected terminal is not directly connected to the local terminal but is mounted under the same local base station, and the directly connected base station is directly connected to the local base station.

[0088] Based on the designed routing strategy and the aforementioned maximum traffic offloading model, the traffic offloaded by a request under this model can be obtained. To further improve the collaboration efficiency between the terminal MC, base station BS, and cloud, and to ensure that the content transmission distance of user-issued network content requests is minimized, a user request routing strategy under a cloud-edge-device collaborative architecture that considers the aggregation effect of the same content is proposed. This strategy aims to ensure that user requests are processed at the edge and terminal as much as possible. This routing strategy determines under what circumstances a user's network content request is routed to which node for processing. The overall routing strategy makes routing decisions in the following order: local MC, local BS, non-directly connected MC, directly connected BS, and cloud.

[0089] As an optional but not limited implementation, the traffic offloaded by each network content request under the traffic offload model is calculated based on caching and routing strategies. This may include: treating the same network content request as the same network content request; and calculating the traffic offloaded by each network content request under the traffic offload model based on caching and routing strategies.

[0090] This case considers the aggregation effect of requests for the same content. Requests for the same content do not need to be queued repeatedly at the base station or terminal, nor do they need to repeatedly occupy network bandwidth. This case treats requests for the same content as a single request, and then divides them into multiple content tasks to be downloaded to the user terminals that require this content. This aggregation effect of the same content can significantly improve system performance, increase network content distribution efficiency, and reduce redundant network content transmission.

[0091] Furthermore, edge caching plays a dominant role in the MEC paradigm. Popular content can be cached on MEC nodes to respond promptly to user content requests. However, the cache capacity of MEC nodes is limited, thus requiring an efficient caching mechanism to make reasonable use of MEC node cache. Based on the above considerations regarding aggregation effects and caching strategies, the caching strategy in this case can be generated by combining at least one of the following strategies: a aggregation-based no-caching strategy, an LRU-based distributed online caching strategy, an LFU-based distributed online caching strategy, an LRFU-based online collaborative caching strategy, a content popularity-based offline collaborative caching strategy, and a DQN-based online collaborative caching strategy.

[0092] For the aggregation-based no-caching strategy, this strategy considers the aggregation effect of requests with the same content. At the user terminal, base station, or cloud, arriving requests first check all request content types in the waiting queue. If there are requests with the same content, they are merged into one processing step. Meanwhile, during content download, requests with the same content only occupy bandwidth once, and upon arrival at the terminal, they are divided into multiple content tasks and distributed to the user. Since this strategy does not set up caching on edge devices, all user requests retrieve content data from the cloud. It is easy to predict that this strategy offloads the least amount of traffic.

[0093] The Least Recently Used (LRU) based distributed online caching strategy refers to using the LRU algorithm to adjust the cached content of edge nodes in real time. The basic principle of the Least Recently Used (LRU) algorithm is to replace content that has not been accessed for a relatively long time. Specifically, this algorithm considers recently accessed content to be likely to be accessed again soon; while content that has not been accessed for a long time is considered unlikely to be accessed in the near future and is replaced or evicted. This strategy uses the LRU algorithm to record the user content requests responded to and processed by edge nodes within a time slot, and adjusts and improves the cached network content within the edge nodes in real time accordingly. The real-time caching strategy ensures the response efficiency of popular requests within a certain period and is easy to implement in practical engineering.

[0094] The Least Frequently Used (LFU) based distributed online caching strategy refers to using the LFU algorithm to adjust the cached content of edge nodes in real time. The basic principle of the Least Frequently Used (LFU) algorithm is to replace content accessed less frequently by users. Specifically, this algorithm considers content accessed more frequently recently as "hot" content; while content accessed less frequently is considered unlikely to be accessed in the near future and is replaced or evicted. This strategy uses the LFU algorithm to record the user content requests responded to and processed by edge nodes within a time slot, and adjusts and improves the cached network content within the edge nodes in real time accordingly. This strategy, along with the LRU-based distributed online caching strategy, can serve as important references for measuring the impact of the frequency and duration of user access to content on the edge cache hit rate.

[0095] For the LRFU-based online collaborative caching strategy, this strategy refers to using the LRFU algorithm to adjust the cached content of edge nodes in real time. The Least Recently Frequently Used (LRFU) algorithm is a cache replacement algorithm that comprehensively considers the frequency and duration of user access to content. It combines the advantages of both LRU and LFU algorithms, thereby further improving the hit rate of edge cache.

[0096] For content popularity-based offline collaborative caching strategies, this is theoretically the optimal caching deployment method. This strategy stores content in the edge node cache according to its popularity ranking, from highest to lowest. Mobile user terminals are given the highest priority, followed by base stations for complementary storage, and finally, the cloud stores all possible requested content. This caching mechanism ensures that highly popular content is processed at the edge, while less popular content is more easily transmitted to the cloud for processing. Furthermore, according to the Zipf distribution, the higher the popularity of a requested content, the higher the proportion of requests it receives. Therefore, the probability of a request being hit at the edge is greater. This strategy maximizes traffic offloading from a caching perspective and is an important metric for evaluating the performance of DQN-based solutions.

[0097] For the DQN-based online collaborative caching strategy, which combines deep learning and reinforcement learning (DRL) to make optimal caching and routing decisions based on currently available network resources, specifically, it aims to find the optimal caching and routing decisions based on the currently available network state space at any given time. The network state space includes the network topology, the number of nodes reachable by the current request, the set of cache states, the set of request queue states, and the set of maximum available link bandwidth. The DRL strategy, which combines deep learning (DL) and reinforcement learning (RL), can further improve resource allocation and content placement in heterogeneous networks. Deep Q-Network (DQN), as a branch of DRL, utilizes deep neural networks to automatically learn low-dimensional feature representations, effectively handling the dimensionality problem brought about by complex network environments. This strategy refers to using the DQN algorithm to make optimal caching and routing decisions based on the currently available network resources in the system.

[0098] DQN usage parameters are: The neural network acts as an evaluation network, outputting action values. At a given time... The input and output at time are states and action value To explore unknown information in the environment, and based on maximizing the use of training results, we utilize... A greedy strategy selects an action after the neural network outputs its action value. Greedy strategy is Probability choice The action, or a random selection with a probability of... The process involves the neural network in DQN updating relevant parameters through backpropagation and gradient descent algorithms to minimize the deviation between the label and the output. The mean squared error (MSE) is used as the loss function for the neural network in DQN, which can be written as:

[0099]

[0100] in, The parameter is The target Q-value output by the target network. As the attenuation factor, The reward value obtained for performing this action, The parameter is The evaluation network outputs the evaluation Q-value. To improve training stability and convergence, the target network is set to a fixed label. Simultaneously, the target network and the evaluation network have the same initial parameters, but the parameters of the target network are different. The parameters of the network are updated after each step to evaluate its performance. It will be updated after a fixed number of steps.

[0101] Furthermore, at a given time network state space including network topology Number of nodes that the current request can reach. Cache state set Request queue status set The set of maximum available link bandwidth and Therefore, the network state space can be written as At a given moment Actions at the location The purpose is to update the cache state in the next training cycle. and select the next jump. The routing node. Therefore, it can be written as The corresponding reward function during the process. It can be written as:

[0102]

[0103] in, Represents a node With nodes The maximum available link bandwidth, This indicates the request content traffic that was unloaded. Indicates the first The probability of requesting this type of content. This represents the discount factor for reward values ​​from past training cycles, ranging from 0 to 1. Therefore, the DQN algorithm for this strategy aims to find a state-space-based reward distribution at any given time t. Optimal caching and routing decisions To maximize reward value .

[0104] S140. Optimize the caching strategy and routing strategy based on the offload traffic solution results to maximize the network content distribution optimization by maximizing the traffic of offloaded content requests.

[0105] Deep reinforcement learning algorithms optimize their caching and routing strategies based on the results of each unloading traffic calculation in order to maximize the unloading content request traffic.

[0106] According to the network structure model established in this invention, the network has a three-layer topology. The top layer represents the cloud; the middle layer represents the BS layer; and the bottom layer represents the MC layer. The number of MCs is set to 12, the number of BSs to 6, and the number of cloud nodes to 1. Requests can be transmitted within connected nodes via connected links. Uplink transmission is for computation requests, which consumes less bandwidth; downlink transmission is for user-requested content, which consumes more bandwidth. In this topology, the terminal MC, the base station BS, and the cloud all have computation and caching capabilities. In the simulation, the Zipf distribution skewness coefficient is assumed to be... The variation range is 0.4-1.6. Simultaneously, the cache capacity is set as a percentage, which is the relative size of the cache capacity compared to the number of content types, and its variation range is 0.1%-1%.

[0107] To demonstrate the system performance of this invention, the system performance is discussed based on the following strategies: a cacheless strategy considering aggregation, a distributed online caching strategy based on LRU, a distributed online caching strategy based on LFU, an online collaborative caching strategy based on LRFU, an offline collaborative caching strategy based on content popularity, and an online collaborative caching strategy based on DQN, thereby obtaining the optimal solution for maximizing the traffic offloading model.

[0108] Figure 3 This section describes the traffic offloading performance of different strategies under varying cache capacities. As the cache capacity increases, more network content can be stored in the BS and MC, improving the efficiency of system resource allocation and content delivery. It's clear that the "DQN-based online collaborative caching strategy" outperforms other solutions. Furthermore, the performance gap widens as the cache capacity increases. This is because DQN can make optimal caching and routing decisions for recently arrived content requests based on request history and currently available network resources. The "LRFU-based online collaborative caching strategy" outperforms both the "LRU-based distributed online caching strategy" and the "LFU-based distributed online caching strategy," but falls short of the "content popularity-based offline collaborative caching strategy." This is because network nodes in online caching strategies store some short-lived popular network content based on user request behavior, resulting in worse performance than offline caching strategies. However, the LRFU algorithm considers both the time and frequency characteristics of network requests to adjust cached content, reducing the negative impact of short-term request behavior, thus making the system performance closer to that of offline caching strategies. The "no-caching strategy" in the BS and MC stores no content, causing all user requests to be downloaded from the cloud, resulting in the worst system performance.

[0109] Figure 4This study assesses the traffic offloading performance of different strategies under varying content popularity. As content popularity increases, users send more requests for more popular online content, improving traffic offloading performance and bridging the performance gap between various strategies. Simultaneously, the performance of the "no-caching strategy" is improved because the convergence effect offloads more traffic from the system. The "DQN-based online collaborative caching strategy" exhibits fast convergence and outperforms other solutions. Frequent cache replacements as popularity increases are also improved, narrowing the gap between offline and online caching strategies.

[0110] Figure 5 This study compares the traffic offloading performance of different strategies under varying numbers of content types. As the number of content types increases, the performance of the online caching strategy based on the replacement algorithm significantly decreases, while the performance of the "no-caching strategy" is only slightly affected. This is because each request must be routed to the cloud to obtain the corresponding content. The "DQN-based online collaborative caching strategy" outperforms the "content popularity-based offline collaborative caching strategy" and other online caching strategies. This is because the DQN algorithm can make optimal caching and routing decisions based on request history and available network resources in the cloud edge environment.

[0111] Figure 6 The performance of traffic offloading for different strategies under varying request arrival rates was observed. The traffic offloading performance of six solutions at different request arrival rates was examined. As the request arrival rate increased, the traffic offloaded by all solutions increased at a relatively slow rate. When the request arrival rate continued to increase and exceeded a certain value, the performance of the solution with internal network caching almost stopped improving. This is because there are more user requests that are lost and retransmitted during routing, thus balancing the impact of packet loss on cache hit rate and request aggregation effect. However, the performance of the "DQN-based online collaborative caching strategy" was significantly better than other solutions because the DQN algorithm reduces the probability of packet loss by optimizing the allocation of network resources and cached content. For the "no-caching strategy," the increased number of user requests strengthened the aggregation effect of nodes, resulting in improved performance.

[0112] Figure 7 The simulation curves show the convergence curves of the DQN reward function under different cache sizes. Observation of the simulation curves reveals that larger node cache sizes can store more popular content in the cloud edge environment, thereby improving resource allocation and obtaining more reward values. When the cache size changes, the proposed "DQN-based online collaborative caching strategy" can converge quickly and reach a steady state.

[0113] This invention addresses aspects lacking in existing solutions by proposing a network content distribution mechanism based on cloud-edge-device resource collaboration. To optimize network resource allocation and content distribution efficiency, this invention comprehensively considers the joint allocation of resources for caching, computing, and communication in the hierarchical heterogeneous network of mobile user terminals, base stations, and the cloud. The traffic offloading problem is established as a centralized model for maximizing traffic offloading, and theoretical analysis is conducted. Then, reinforcement learning is used to optimize and solve this model, designing an adaptive solution for finding the optimal cached content placement and reasonable routing paths. Finally, simulation analysis is performed on the designed traffic offloading model under various strategies, discussing the optimal content caching and maximized traffic offloading by comprehensively considering multiple factors affecting traffic offloading. Simulation results show that the solution proposed in this invention significantly outperforms existing cloud-edge collaborative solutions, providing a feasible solution for the high traffic load pressure brought about by dense base station deployment and the growth in the number of users and the scale of communication data.

[0114] Figure 8 This invention provides a flowchart of a network content distribution method. This embodiment is applicable to situations where network distribution can be better performed in collaboration between the cloud, edge nodes, and terminal devices. The method can be executed by a network content distribution device, which can be implemented in hardware and / or software and can be configured in any electronic device with network communication capabilities. Figure 8 As shown, the method may include:

[0115] S810, load the optimized caching and routing strategies.

[0116] The caching and routing strategies obtained by using any of the caching and routing strategy optimization methods described in the above embodiments.

[0117] S820 distributes network content using the cache and routing results indicated by the optimized caching and routing strategies, in order to maximize the unloading of content request traffic.

[0118] Figure 9 This invention provides a structural block diagram of a caching and routing policy optimization device. This embodiment is applicable to situations where network distribution can be better performed under the collaborative operation of the cloud, edge nodes, and terminal devices. The caching and routing policy optimization device can be implemented in hardware and / or software and can be configured in any electronic device with network communication capabilities. Figure 9 As shown, the device may include: a network model construction module 910, an offload model construction module 920, an offload traffic determination module 930, and a policy optimization module 940. Wherein:

[0119] The network model construction module 910 is used to construct a network content popularity model and a network structure model. The network structure model includes a terminal, a base station, and a cloud. The base station is connected to the cloud through the core network, and each base station has a terminal attached to it.

[0120] The unloading model construction module 920 is used to construct a maximized traffic unloading model under the network popularity model and network structure model. The maximized traffic unloading model is used to convert the network resource allocation task into an unloading network content request traffic task.

[0121] The offloading traffic determination module 930 is used to solve the traffic offloaded by each network content request under the maximized traffic offloading model based on caching and routing strategies.

[0122] The strategy optimization module 940 is used to optimize the caching strategy and routing strategy based on the offload traffic solution results, so as to maximize the network content distribution optimization by maximizing the traffic of offloaded content requests.

[0123] Based on the above embodiments, optionally, a network content popularity model is constructed, including:

[0124] A network content popularity model is constructed based on Zipf's law, which is used to describe the request probability of network content.

[0125] Based on the above embodiments, optionally, a network structure model is constructed, including:

[0126] Based on a layered heterogeneous network of cloud, edge, and terminal, a layered heterogeneous network structure model is constructed. In this network structure model, base stations are directly connected to each other, and the network content cached in the directly connected base stations can be transmitted to each other at high speed through optical fiber. Terminals access the base stations through wireless connection.

[0127] Based on the above embodiments, optionally, the maximized traffic offloading model includes a terminal traffic offloading model and a base station traffic offloading model; the terminal traffic offloading model is used to output the network content request traffic offloaded at the terminal; the base station traffic offloading model is used to output the network content request traffic offloaded at the base station.

[0128] Based on the above embodiments, optionally, the terminal traffic offloading model includes a local terminal traffic offloading model and a non-directly connected terminal traffic offloading model; the base station traffic offloading model includes a local base station traffic offloading model and a directly connected base station traffic offloading model.

[0129] Based on the above embodiments, optionally, the routing strategy is used to describe the routing path of network content requests. The routing path is constructed in the order of local terminal, local base station, non-directly connected terminal, directly connected base station, and cloud. The local terminal is the terminal that makes the network content request, the local base station is directly connected to the local terminal, the non-directly connected terminal is not directly connected to the local terminal but is mounted under the same local base station, and the directly connected base station is directly connected to the local base station.

[0130] Based on the above embodiments, optionally, the caching strategy is generated by combining at least one of the following strategies: a convergence-based no-caching strategy, an LRU-based distributed online caching strategy, an LFU-based distributed online caching strategy, an LRFU-based online collaborative caching strategy, a content popularity-based offline collaborative caching strategy, and a DQN-based online collaborative caching strategy.

[0131] Optionally, based on the above embodiments, the online collaborative caching strategy based on DQN is a deep reinforcement learning-based online collaborative caching strategy that combines deep learning and reinforcement learning, used to find the optimal caching and routing decision based on the currently available network state space at any given time; the network state space includes the network topology, the number of nodes reachable by the current request, the cache state set, the request queue state set, and the set of maximum available link bandwidth.

[0132] Based on the above embodiments, optionally, the traffic offloaded by each network content request under the maximization traffic offload model is calculated according to the caching strategy and routing strategy, including:

[0133] Treat requests for the same network content as the same network content request;

[0134] Based on caching and routing strategies, the traffic offloaded by each network content request under the maximized traffic offload model is calculated.

[0135] The caching and routing policy optimization device provided in the embodiments of the present invention can execute the caching and routing policy optimization method provided in any of the embodiments of the present invention, and has the corresponding functions and beneficial effects of executing the caching and routing policy optimization method. For details, please refer to the relevant operations of the caching and routing policy optimization method in the foregoing embodiments.

[0136] Figure 10 This invention provides a flowchart of a network content distribution device. This embodiment is applicable to situations where network distribution can be better performed in collaboration between the cloud, edge nodes, and terminal devices. The network content distribution device can be implemented in hardware and / or software and can be configured in any electronic device with network communication capabilities. Figure 10As shown, the device may include a loading module 1010 and a distribution module 1020. Wherein:

[0137] Load module 1010, used to load the optimized caching and routing strategies;

[0138] The caching and routing strategies obtained by using any of the caching and routing strategy optimization methods described in the above embodiments.

[0139] The distribution module 1020 is used to distribute network content using the cache and routing results indicated by the optimized caching and routing strategies, in order to maximize the unloading content request traffic.

[0140] The network content distribution device provided in the embodiments of the present invention can execute the network content distribution method provided in any of the embodiments of the present invention, and has the corresponding functions and beneficial effects of executing the network content distribution method. For details, please refer to the relevant operations of the network content distribution method in the foregoing embodiments.

[0141] Figure 11 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0142] like Figure 11 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0143] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0144] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of 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 suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as caching and routing strategy optimization methods or network content distribution methods.

[0145] In some embodiments, the caching and routing policy optimization method or network content distribution method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the caching and routing policy optimization method or network content distribution method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the caching and routing policy optimization method or network content distribution method by any other suitable means (e.g., by means of firmware).

[0146] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0147] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0148] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0149] To provide interaction with a user, the systems and techniques described herein 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 pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; 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 sound input, voice input, or tactile input).

[0150] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0151] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0152] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0153] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. 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 substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for optimizing caching and routing strategies, characterized in that, The method includes: A network content popularity model and a network structure model are constructed. The network structure model includes terminals, base stations, and the cloud. The base stations are connected to the cloud through the core network, and each base station has a terminal attached. The construction process of the network structure model includes: building a layered heterogeneous network structure model based on the cloud-edge-device layered heterogeneous network. A maximized traffic offloading model is constructed based on a network popularity model and a network structure model. The maximized traffic offloading model is used to transform the network resource allocation task into an offloading task of network content request traffic. The maximized traffic offloading model includes a terminal traffic offloading model and a base station traffic offloading model. The terminal traffic offloading model is used to output the network content request traffic offloaded at the terminal. The base station traffic offloading model is used to output the network content request traffic offloaded at the base station. Based on caching and routing strategies, the traffic offloaded by each network content request under the maximized traffic offload model is calculated. The caching and routing strategies are optimized based on the results of the offload traffic calculation.

2. The method according to claim 1, characterized in that, The construction of the online content popularity model includes: A network content popularity model is constructed based on Zipf's law, which is used to describe the request probability of network content.

3. The method according to claim 1, characterized in that, Constructing a network structure model includes: In the network structure model described, base stations are directly connected to each other, and the network content cached within the directly connected base stations can be transmitted to each other at high speed via optical fiber. Terminals access the base stations via wireless connection.

4. The method according to claim 1, characterized in that, The terminal traffic offloading model includes a local terminal traffic offloading model and a non-directly connected terminal traffic offloading model; the base station traffic offloading model includes a local base station traffic offloading model and a directly connected base station traffic offloading model.

5. The method according to claim 1, characterized in that, The routing strategy is used to describe the routing path of network content requests. The routing path is constructed in the order of local terminal, local base station, non-directly connected terminal, directly connected base station, and cloud. The local terminal is the terminal that makes the network content request. The local base station is directly connected to the local terminal. The non-directly connected terminal is not directly connected to the local terminal but is mounted under the same local base station. The directly connected base station is directly connected to the local base station.

6. The method according to claim 1, characterized in that, The caching strategy is generated by combining at least one of the following strategies: a convergence-based no-caching strategy, an LRU-based distributed online caching strategy, an LFU-based distributed online caching strategy, an LRFU-based online collaborative caching strategy, a content popularity-based offline collaborative caching strategy, and a DQN-based online collaborative caching strategy.

7. The method according to claim 6, characterized in that, The DQN-based online collaborative caching strategy is a deep reinforcement learning-based online collaborative caching strategy that combines deep learning and reinforcement learning. It is used to find the optimal caching and routing decision based on the currently available network state space at any given time.

8. The method according to claim 1, characterized in that, Based on caching and routing strategies, the traffic offloaded by each network content request under the maximization of traffic offload model is calculated, including: Treat requests for the same network content as the same network content request; Based on caching and routing strategies, the traffic offloaded by each network content request under the maximized traffic offload model is calculated.

9. A method for distributing network content, characterized in that, The caching and routing strategies obtained by the caching and routing strategy optimization method according to any one of claims 1-8, wherein the method includes: Load the optimized caching and routing strategies; The network content is distributed using the cache and routing results indicated by the optimized caching and routing strategies.

10. A caching and routing strategy optimization device, characterized in that, The device includes: The network model construction module is used to construct a network content popularity model and a network structure model. The network structure model includes terminals, base stations, and the cloud. The base stations are connected to the cloud through the core network, and each base station has a terminal attached to it. The construction process of the network structure model includes: constructing a layered heterogeneous network structure model based on a cloud-edge-device layered heterogeneous network. The unloading model construction module is used to construct a maximized traffic unloading model under the network popularity model and network structure model. The maximized traffic unloading model is used to transform the network resource allocation task into an unloading network content request traffic task. The maximized traffic unloading model includes a terminal traffic unloading model and a base station traffic unloading model. The terminal traffic unloading model is used to output the network content request traffic unloaded at the terminal. The base station traffic unloading model is used to output the network content request traffic unloaded at the base station. The unloading traffic determination module is used to calculate the unloading traffic of each network content request under the maximized traffic unloading model based on caching and routing strategies. The strategy optimization module is used to optimize the caching strategy and routing strategy based on the offload traffic solution results.

11. A network content distribution device, characterized in that, The caching and routing strategies obtained by the caching and routing strategy optimization method according to any one of claims 1-8, the apparatus comprising: The loading module is used to load the optimized caching and routing strategies; The distribution module is used to distribute network content using the cache and routing results indicated by the optimized caching and routing strategies, in order to maximize the unloading of content request traffic.

12. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the caching and routing policy optimization method of any one of claims 1-8 or the network content distribution method of claim 9.

Images