Content Distribution Architecture Service Caching Method for Space-Ground Collaboration of Low-Earth-Orbit Satellite Internet

Abstract

The present invention discloses a content distribution architecture service caching method for low-orbit satellite Internet satellite-ground cooperation, which relates to the field of wireless communication. The method includes the following steps: for the satellite side, collect the content request information of users, unload part of the content to the ground server according to the request information, and the ground server responds to the request for this part of the content, and the remaining requested content is directly distributed by the satellite. For the ground server, in order to minimize the delay in the distribution process, it is necessary to optimize the cache status of the ground server to maximize its cache hit rate. Taking the maximization of the cache hit rate as the optimization goal, the optimization problem is transformed into a "0-1" knapsack problem. The present invention proposes an improved binary particle swarm optimization (EBPSO) algorithm for the above optimization problem to output the best cache decision of the ground server. The present invention can reduce the delay in the distribution process of this architecture.

Description

Technical Field

[0001] The present invention mainly relates to the storage resource allocation in the field of resource-constrained wireless communication in the satellite-ground collaborative network, and specifically studies a content distribution architecture service caching method for low-orbit satellite Internet satellite-ground collaboration. Background Art

[0002] In the future 6G era, the vision of integrating space, air, and ground will be realized. With the continuous development of information networks and satellite communication technologies, satellite networks and terrestrial networks are gradually deeply integrated, giving rise to a new type of satellite-ground collaborative network. Compared with traditional single satellite or terrestrial networks, this network has a vast coverage area and excellent anti-interference ability.

[0003] In the content distribution architecture based on the satellite-ground collaborative network, since the direct distribution delay of the ground server is lower than that of the satellite direct distribution, and during the distribution process, the satellite will unload some content to the ground server. If the ground server caches the requested content, it will directly distribute it, and other requests will be distributed by the satellite. At this time, the delay of the entire distribution process is determined by the cache hit rate of the ground server. When its hit rate is higher, it means that the number of content distributed by the ground server is more, so the delay of the entire distribution process is smaller. Therefore, it is very necessary to design an effective ground server caching method to maximize its cache hit rate. Summary of the Invention

[0004] The purpose of the present invention is to provide a content distribution architecture service caching method for low-orbit satellite Internet satellite-ground collaboration. For the satellite side, by collecting user request information, the numbers and request times of each requested content are counted. According to the request information, some content is unloaded to the ground server, and the ground server responds to the requests for this part of the content, and the remaining requested content is directly distributed by the satellite. To minimize the delay during the distribution process, it is necessary to optimize the cache state of the ground server to maximize its cache hit rate.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0006] A content distribution architecture service caching method for low-orbit satellite Internet satellite-ground collaboration, comprising the following steps:

[0007] S1: Establish a content distribution architecture model based on the satellite-ground collaborative network, including low-orbit satellites and ground servers; wherein, both low-orbit satellites and ground servers have caching and distribution functions, and the ground server has a computing function and can perform content version conversion;

[0008] S2: The satellite collects the content request information of each user in the network and makes statistics;

[0009] S3: Establish a cache optimization model, that is, minimize the latency in the distribution process. Convert the goal of minimizing latency into maximizing the cache hit rate. The cache hit rate refers to the proportion of the content cached in the ground server among the content requested by users, and establish the cache optimization model as a "0-1" knapsack problem;

[0010] S4: Use the binary particle swarm optimization algorithm to solve the "0-1" knapsack problem to obtain the optimal cache decision for the ground server;

[0011] S5: The satellite unloads some of the content in the statistical information to the ground server cache according to the optimal cache decision. The content request information cached by the ground server is distributed by the ground server, and other requests are directly distributed by the satellite.

[0012] Furthermore, the satellite caches all the content request information in the network, and the ground server only caches the basic version of the content request information. And the content request information is only transcoded from the basic version into other versions, and the other versions do not convert to each other.

[0013] Furthermore, the content request information collected by the satellite includes the content numbers requested by all users, and counts the request times of each number.

[0014] Furthermore, the specific steps of S3 are as follows:

[0015] S31: Establish a cache optimization model as a "0-1" knapsack model. The cache capacity of the ground server is regarded as the knapsack capacity, and the size and request times of the content are regarded as the weight and value of the item. It is necessary to achieve the goal of maximizing the cache hit rate with limited capacity;

[0016] S32: Use the requested content and version as the grouping basis to divide users into multiple multicast groups G = {g 1,1 ,…,g V,E}; By counting the content request information, obtain the number of users in each multicast group The calculation formula for the total number of requested users is: Among them, g i,j is the multicast group composed of users who have requested version j of content i; is the number of users in the multicast group g i,j , V is the number of content request information in the network, and E is the number of versions of the content request information;

[0017] S33: Add up the request times of all versions of content i. The calculation formula is: In the formula, val i is the value of content i, and obtain the value of each content: val = (val1,…,val V );

[0018] S34: Obtain the cache status of the ground server, C = (c1, …, c V ), where c i ∈{0, 1}. If the value is 0, it means that content i is not cached; if the value is 1, it means that i is cached; make the total content value val sum maximum, val sum = C T ·val.

[0019] Furthermore, the said S4 specifically includes the following steps:

[0020] S41: In the EBPSO algorithm mechanism, a group of particles forms a population, where each particle represents a cache status; initialize the position x i and velocity v i of particle i; use the total content value val sum calculated by the cache status represented by each particle as the fitness of each particle, and record the position P of the particle with the current best fitness value and the historical best position of the i-th particle

[0021] S42: For particles that violate the storage capacity constraint of the ground server, set the fitness value to 0. If the fitness of any particle in the particle swarm is greater than the current best fitness value, replace P with the fitness of the corresponding particle;

[0022] S43: Traverse each particle in the particle swarm and re-initialize the position of the particle that reaches the local optimal fitness;

[0023] S44: Introduce the inertia weight factor w d , α1 and α2. Each particle adjusts its position according to P and velocity v i as shown in the following formula:

[0024]

[0025] where, and are the velocities of the i-th particle at the t-th and (t - 1)-th iterations respectively, is the local optimum of the i-th particle at the (t - 1)-th iteration, P t-1 is the global optimum at the (t - 1)-th iteration, is the position where the i-th particle is located at the (t - 1)-th iteration, rand1 and rand2 are random numbers uniformly distributed between 0 and 1, w d is the fine-tuning weight exponent, and α1 and α2 are weight ratios;

[0026] S45: Determine whether the termination condition is met. If it is met, end the iteration, output the cache state C represented by the particle with the best fitness value, otherwise update the velocity and position parameters of the particle, and return to S41.

[0027] Further, the S5 specifically includes the following process:

[0028] The satellite unloads the base version vid of the corresponding content i according to the optimal cache decision i,1 to the ground server, and the ground server then obtains multiple versions through transcoding calculation; the ground server and the satellite respectively send the cached content to the users requesting the corresponding content.

[0029] The beneficial effects of the present invention are as follows:

[0030] An improved binary particle swarm optimization content caching algorithm is adopted to propose a content distribution architecture service caching method for low-orbit satellite Internet space-ground collaboration. This method takes into account the cache capacity of the ground server, the size of the content, and the number of requests, dynamically adjusts the cache state of the ground server, maximizes the cache hit rate of the ground server during the distribution process, and thus minimizes the delay of the entire distribution process. At the same time, the EBPSO algorithm has the advantages of not requiring relaxation, fast convergence, and avoiding falling into local optima, which helps to obtain the optimal cache decision. Description of the Drawings

[0031] Figure 1 It is a schematic diagram of the content distribution architecture model based on the space-ground collaborative network.

[0032] Figure 2 It is a schematic diagram of the overall process of the present invention.

[0033] Figure 3 It is a schematic diagram of the overall process of the improved binary particle swarm optimization (EBPSO) algorithm. Specific Embodiments

[0034] The following will clearly and completely describe the technical solutions in combination with the embodiments of the present invention. It should be noted that the described embodiments are only a part of the present invention, not all embodiments. And based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0035] Such as Figure 1The following is a content distribution architecture model based on a satellite-ground collaborative network, including satellites, ground servers, and users. The satellites and ground servers have caching and content distribution functions, and the ground servers also have computing and transcoding functions. Among them, the satellites cache all content request information in the network, and the ground servers only cache the basic versions of the content request information. Moreover, the content request information is only transcoded from the basic version into other versions, and no mutual conversion is performed between the other versions. The content request information collected by the satellites includes the content numbers requested by all users, and the request times for each number are counted.

[0036] Figure 2 The overall process schematic diagram of the present invention is given. The satellites collect user request information, model the caching decision of the ground servers as a "0-1" knapsack problem, use an improved binary particle swarm optimization (EBPSO) algorithm to find the optimal caching decision, and unload the content to the ground servers according to the decision, and the content distribution is completed through their collaboration. Specifically, it includes the following steps:

[0037] S1: Establish a content distribution architecture model based on a satellite-ground collaborative network, including low-earth orbit satellites and ground servers; among them, both the low-earth orbit satellites and the ground servers have caching and distribution functions, and the ground servers have computing functions and can perform content version conversion.

[0038] S2: The satellites collect the content request information of each user in the network and perform statistics.

[0039] S3: Establish a caching optimization model, that is, minimize the delay in the distribution process, convert the goal of minimizing the delay into maximizing the caching hit rate. The caching hit rate refers to the proportion of the content cached in the ground servers among the content requested by users, and establish the caching optimization model as a "0-1" knapsack problem; specifically, it includes the following steps:

[0040] S31: Establish the caching optimization model as a "0-1" knapsack model. The caching capacity of the ground servers is regarded as the knapsack capacity, the size and request times of the content are regarded as the weight and value of the items, and the goal of maximizing the caching hit rate needs to be achieved with limited capacity;

[0041] S32: Use the requested content and version as the grouping basis to divide users into multiple multicast groups G = {g 1,1 ,…,g V,E}; by counting the content request information, obtain the number of users in each multicast group The calculation formula for the total number of requested users is: Among them, g i,j is the multicast group composed of users who have requested version j of content i; is the multicast group g i,jThe number of users within, V is the number of content request messages in the network, and E is the number of versions of the content request messages;

[0042] S33: Add up the request counts of all versions of content i. The calculation formula is: In the formula, val i is the value of content i, and obtain the values of each content: val = (val1,..., val V );

[0043] S34: Obtain the cache status of the ground server, C = (c1,..., c V ), where c i ∈{0, 1}. If the value is 0, it means that content i is not cached. If the value is 1, it means that i is cached; make the total content value val sum maximum, val sum = C T ·val.

[0044] S4: Use the binary particle swarm optimization algorithm to solve the "0-1" knapsack problem and obtain the optimal cache decision of the ground server.

[0045] Figure 3 Figure shows the schematic flow chart of the improved binary particle swarm optimization (EBPSO) algorithm. First, set the algorithm parameters, then initialize the positions and velocities of the population particles, and calculate the individual fitness values of the particles in the population. After that, re-initialize the positions of the particles that reach the local optimal fitness value, and continuously iterate and update the particle velocity and position parameters before reaching the termination condition. Finally, terminate the iteration after reaching the condition and output the optimal cache decision of the ground server. S4 specifically includes the following steps:

[0046] S41: In the EBPSO algorithm mechanism, a group of particles form a population, and each particle represents a cache status; initialize the position x i and velocity v i of particle i; use the total content value val sum calculated by the cache status represented by each particle as the fitness of each particle, and record the position P of the particle with the current best fitness value and the historical best position of the i-th particle

[0047] S42: For the particles that violate the storage capacity constraint of the ground server, set the fitness value to 0. If the fitness of any particle in the particle swarm is greater than the current best fitness value, replace P with the fitness of the corresponding particle;

[0048] S43: Traverse each particle in the particle swarm and re-initialize the positions of the particles that reach the local optimal fitness;

[0049] S44: Introducing the inertia weight factor w d , α1 and α2, each particle is based on P and speed v i Adjust the position as shown in the following formula:

[0050]

[0051] in, and are the speeds of the ith particle at the tth and t-1th iterations, respectively. is the t-1th local optimum of the i-th particle, P t-1 is the t-1th global optimal value, is the position of the ith particle at the t-1th iteration, rand1 and rand2 are random numbers uniformly distributed between 0 and 1, and w d is the fine-tuning weight index, α1 and α2 are the weight ratios;

[0052] S45: Determine whether the termination condition is met. If so, end the iteration and output the cache state C represented by the particle with the best fitness value. Otherwise, update the speed and position parameters of the particle and return to S41.

[0053] S5: The satellite caches the base version vid corresponding to content i according to the best cache decision i,1 The content is unloaded to the ground server, which then obtains multiple versions through computational transcoding. The ground server and satellite respectively send the cached content to the users who request the corresponding content.

[0054] The embodiments of the present invention are described above, but the present invention is not limited to the above-mentioned specific implementation modes. The above-mentioned specific implementation modes are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, all of which are protected by the present invention.

Claims

1. A method for caching content distribution architecture services in low-orbit satellite Internet space-ground collaboration, characterized in that, It includes the following steps: S1: Establish a content distribution architecture model based on the satellite-ground collaborative network, including low-orbit satellites and ground servers; among them, both low-orbit satellites and ground servers have caching and distribution functions, and ground servers have computing functions and can perform content version conversion; S2: The satellite collects the content request information of each user in the network and conducts statistics; S3: Establish a caching optimization model, that is, minimize the latency in the distribution process, convert the latency minimization goal into maximizing the cache hit rate. The cache hit rate refers to the proportion of the content cached in the ground server in the content requested by users, and establish the caching optimization model as a "0-1" knapsack problem; S4: Use the binary particle swarm optimization algorithm to solve the "0-1" knapsack problem to obtain the optimal caching decision of the ground server; S5: The satellite unloads some of the content in the statistical information to the ground server cache according to the optimal caching decision. The content request information cached by the ground server is distributed by the ground server, and other requests are directly distributed by the satellite; Among them, the specific steps of S3 are as follows: S31: Establish the caching optimization model as a "0-1" knapsack model. The cache capacity of the ground server is regarded as the knapsack capacity, and the size and request times of the content are regarded as the weight and value of the item. It is necessary to achieve the goal of maximizing the cache hit rate with limited capacity; S32: Group users into multiple multicast groups \(G = \{g 1,1 , \ldots, g V,E \}\) based on the requested content and version; obtain the number of users in each multicast group by counting the content request information The formula for calculating the total number of requesting users is: where \(g i,j \) is the multicast group composed of users who have requested version \(j\) of content \(i\); is the number of users in multicast group \(g i,j \), \(V\) is the number of content request messages in the network, and \(E\) is the number of versions of the content request messages; S33: Add up the request counts of all versions of content i. The calculation formula is: In the formula, val i is the value of content i, and the values of each content are obtained: val = (val1, …, val V ); S34: Obtain the cache status of the ground server, C = (c1, …, c V ), where c i ∈{0, 1}, if the value is 0, it means that content i is not cached, and if the value is 1, it means that i is cached; make the total value of the content val sum maximum, val sum = C T ·val; Among them, the specific steps of S4 are as follows: S41: In the EBPSO algorithm mechanism, a group of particles forms a population, where each particle represents a cache state; initialize the position x of particle i i and velocity v i ; use the total content value val sum calculated from the cache state represented by each particle as the fitness of each particle, and record the position P of the particle with the current best fitness value and the historical best position of the i-th particle S42: For particles that violate the storage capacity constraint of the ground server, record the fitness value as 0. If the fitness of any particle in the particle swarm is greater than the current best fitness value, then replace P with the fitness of the corresponding particle; S43: Traverse each particle in the particle swarm and re-initialize the position of the particle that reaches the local optimal fitness; S44: Introduce the inertia weight factor w d , α1 and α2, and each particle adjusts its position according to P and velocity v i as shown in the following formula: Among them, and are the velocities of the i-th particle at the t-th and (t-1)-th iterations respectively, is the local optimum of the i-th particle at the (t-1)-th iteration, P t-1 is the global optimum at the (t-1)-th iteration, is the position where the i-th particle is located at the (t-1)-th iteration, rand1 and rand2 are random numbers uniformly distributed between 0 and 1, w d is the fine-tuning weight exponent, and α1 and α2 are weight ratios; S45: Judge whether the termination condition is satisfied. If it is satisfied, end the iteration and output the cache state C represented by the particle with the best fitness value. Otherwise, update the velocity and position parameters of the particle and return to S41.

2. The content distribution architecture service caching method for low-orbit satellite Internet space-ground cooperation according to claim 1, characterized in that, The satellite caches all the content request information in the network. The ground server only caches the basic version of the content request information, and the content request information is only transcoded from the basic version to other versions, and the other versions are not converted with each other.

3. The content distribution architecture service caching method for low-orbit satellite Internet space-ground collaboration according to claim 1, wherein The content request information collected by the satellite includes the content numbers requested by all users, and the request times of each number are counted.

4. The content distribution architecture service caching method for low-orbit satellite Internet space-ground cooperation according to claim 1, wherein The specific process of S5 is as follows: The satellite unloads the base version vid of the corresponding content i according to the optimal caching decision i,1 to the ground server, and the ground server obtains multiple versions after transcoding through calculation; the ground server and the satellite respectively send the cached content to the user requesting the corresponding content.

Images