Adaptive bit rate video caching method and system

Abstract

The present invention provides an adaptive bitrate video caching method and system. The method primarily includes: establishing a heterogeneous mobile edge computing network with a macro base station and multiple small base stations; calculating user content access delays; calculating the capacity constraints of the macro base station and small base stations, as well as the constraints of user video requests; calculating the average delay cost of all video requests; obtaining cache information for each server and user request scheduling information through a cache placement method and a video request scheduling method; and finally, the mobile edge computing server caches the corresponding video variants and video request scheduling method. Compared to existing technologies, the present invention improves cache hit rates and reduces content access delays by transforming the adaptive bitrate video caching problem into a caching problem and a scheduling problem.

Description

Technical Field

[0001] The present invention relates to an adaptive bit rate video caching method and system, and belongs to the technical field of edge computing and video caching. Background Art

[0002] Mobile Edge Computing (MEC) enables not only edge caching but also edge processing. Furthermore, designing a separate caching scheme for each base station is not only cumbersome but also fails to fully utilize cache resources. To address these challenges, collaborative caching schemes have been proposed to improve network performance. While collaborative caching improves cache utilization compared to non-collaborative caching, it still has some drawbacks due to its architecture.

[0003] In wireless video streaming, due to the highly dynamic nature of network conditions, coupled with the heterogeneity of user processing capabilities, users may have different preferences and demands for specific quality and formats of videos. For example, users with high-performance devices and fast network connections generally prefer high-resolution videos, while users with low processing capabilities or low-bandwidth connections may not be able to enjoy high-quality videos because of large delays and the video may not be suitable for the device's display. In ABR video streaming, different bitrate versions of the video can be generated and transmitted to the user based on the user's device capabilities, network connection, and specific requests. Based on actual conditions, the present invention aims to minimize the user's access latency, which is a practical issue, taking into account cache collaboration and adaptive video streaming.

[0004] In view of this, it is indeed necessary to propose an adaptive bit rate video caching method and system to solve the above problems. Summary of the Invention

[0005] The object of the present invention is to provide a method and system for caching video at an adaptive bit rate, thereby minimizing the completion time of a task.

[0006] To achieve the above object, the present invention provides an adaptive bit rate video caching method, which mainly includes the following steps:

[0007] Step 1: Establish a heterogeneous mobile edge computing network with a macro base station and multiple small base stations. Deploy a mobile edge computing server on the macro base station and each small base station, preset the cache placement method, and transcode videos with different bit rates.

[0008] Step 2: Calculate the user's content access delay;

[0009] Step 3: Calculate the capacity constraints of the macro base station and the small base station and the constraints of the user's video request;

[0010] Step 4: Calculate the average delay cost of all video requests;

[0011] Step 5: Obtain cache information of each server and user request scheduling information through the cache placement method and video request scheduling method;

[0012] Step 6: The mobile edge computing server caches the corresponding video variant and video request scheduling method according to the cache placement method obtained in step 5.

[0013] As a further improvement of the present invention, in step 3, in the heterogeneous mobile edge computing network, an edge cache C is deployed in the small base station n. n , the cache capacity is M n , deploy an edge cache C in the macro base station N+1 , the cache capacity is M N+1 , the cache placement method is expressed as C={C1,C2,...C N ,C N+1}, then the capacity constraint of small base station n is:

[0014]

[0015] Capacity constraints of macro base stations:

[0016]

[0017] in, v l Indicates a video that the user can request, r l represents the storage size of each video variant, and V represents all video variants that the user can request.

[0018] As a further improvement of the present invention, in step 3, the constraints of the user's video request can be expressed as

[0019]

[0020] in, Indicates that the user on the small base station n sends a video v l The request can be served by the small cell m without the need for transcoding; Indicates that the user on the small base station n sends a video v l The request can be served by the small cell m, but needs to be transcoded; =1 means that the user on the small base station n sends a video v l The request can be served by the macro base station without the need for transcoding; Indicates that the user on the small base station n sends a video v l The request can be served by the macro base station, but needs to be transcoded; Indicates that the user on the small base station n sends a video v l The request can be serviced by the original content server.

[0021] As a further improvement of the present invention, in step 4, the average delay cost is

[0022]

[0023] in, It is the set of video types requested by users during the use of small cell n.

[0024] As a further improvement of the present invention, in step 4, the objective function of the optimal solution for minimizing the average delay cost within each time t is:

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036] As a further improvement of the present invention, in step 5, the cache information of the server includes cache placement decision information of the macro base station and the small base station, and the request scheduling information includes conversion between video variants.

[0037] As a further improvement of the present invention, the cache placement method includes a reactive LRU cache placement method and a proactive cache placement method.

[0038] As a further improvement of the present invention, the active cache placement method has the following specific steps:

[0039] Step A1: Initialize and place P = {P1, P2, ..., P n ,P N+1}, and cache placement decision set C placement set C = {c1,c2,.c n .,c N.+1},,

[0040] Step A2: Place the decisions that meet the conditions in set P into the corresponding cache set C;

[0041] Step A3: Redefine the objective function as h(C), that is make

[0042] Step A4: Determine whether the placement set P is empty. If it is not empty, find the largest cache placement decision.

[0043] Step A5: If Then Delete from C and replace P n Delete from set P;

[0044] Step A6: If Then Delete from C and replace P n Delete from set P;

[0045] Step A7: Repeat steps A2 to A6 until the set P is empty.

[0046] As a further improvement of the present invention, the video request scheduling method has the following specific steps:

[0047] Step B1: For each video request v arriving at the small cell n (n∈N) l ,if Then the file v l Transmitted from the cache of small base station n to the user;

[0048] Step B2: Define video v on small base station n l The closest transcodeable variant is defined as T(n,v l )=v h , and h satisfies the following constraints:

[0049] Step B3: If T(n,v l )≠0 and Then the video T(n,v l ) is converted to v l and transmit it to the user;

[0050] Step B4: If Then the video v lTransmitted from the macro base station to the small base station n and then transmitted to the user;

[0051] Step B5: If Then the video is transmitted to the cache of the small base station m and then transmitted to the user through the macro base station and small base station n;

[0052] Step B6: If T(N+1,v l )≠0 and Then the video in the macro base station cache is transcoded and transmitted to the small base station n and then transmitted to the user;

[0053] Step B7: If If it is not an empty set, then calculate the remaining computing resources of all qualified small base stations

[0054] Step B8: Find the small base station with the largest remaining processing resources, i.e. Under the constraint of S(m)≥0, if it is not an empty set, T(f,v l ) and transcode, transmit the transcoded video to the macro base station, then transmit it to the small base station n, and then transmit it to the user;

[0055] Step B9: If all the above conditions are met, obtain v from the original content server. l , and then transmitted to users through macro base stations and small base stations n.

[0056] To achieve the above object, the present invention further provides an adaptive bit rate video caching system, which applies the above-mentioned adaptive bit rate video caching method.

[0057] The beneficial effects of the present invention are as follows: the present invention can improve the cache hit rate and reduce content access delay by converting the adaptive bit rate video cache problem into a cache problem and a scheduling problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a structural diagram of the adaptive bit rate video caching system of the present invention. DETAILED DESCRIPTION

[0059] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] It should be noted here that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions of the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0061] In addition, it should be noted that the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements that are inherent to such process, method, article, or apparatus.

[0062] like Figure 1 As shown, the present invention proposes an adaptive bitrate video caching method and system based on a heterogeneous mobile edge computing network. The adaptive bitrate video caching method can take into account the communication delay between each base station, aiming to minimize the user's access delay, and solve the video cache placement decision and user scheduling decision. The method mainly includes the following steps:

[0063] Step 1: Establish a heterogeneous mobile edge computing network with a macro base station and multiple small base stations. Deploy a mobile edge computing server on the macro base station and each small base station, preset the cache placement method, and transcode videos with different bit rates.

[0064] Step 2: Calculate the user's content access delay;

[0065] Step 3: Calculate the capacity constraints of the macro base station and the small base station and the constraints of the user's video request;

[0066] Step 4: Calculate the average delay cost of all video requests;

[0067] Step 5: Obtain cache information of each server and user request scheduling information through the cache placement method and video request scheduling method;

[0068] Step 6: The mobile edge computing server caches the corresponding video variant and video request scheduling method according to the cache placement method obtained in step 5.

[0069] Steps 1 to 6 are described in detail below.

[0070] In step 1, a heterogeneous mobile edge computing (MEC) network is established. The heterogeneous MEC network consists of a macro base station (MBS) and multiple small base stations (SBSs). An MEC server is deployed on the MBS and each SBS, and a cache placement method is preset to transcode videos with different bit rates.

[0071] N={1, 2, 3...n} represents different SBS sets. The nth SBS is SBS n. They are deployed in different areas, and their coverage areas do not overlap with each other.

[0072] N+1 represents MBS.

[0073] The set of videos available for user request is defined as V = {1, 2, ...v}, and it is defined that all videos have the same playback time, and each video has different bitrate variants. Different bitrates of the same video can be converted to each other, but only from high bitrate to low bitrate.

[0074] Let L = {1, 2, ..., l} represent all bit rate levels, b l Indicates the bit rate corresponding to the video. A higher bit rate level has a higher bit rate, that is, b1<b2<...<b l .

[0075] For any This means that the video can be converted from bit rate h to l, using p hl It represents the cost (CPU cycles) of converting the same video from bit rate h to l, let p hl =p l The size of each video variable l is proportional to its bit rate and is expressed as r l Indicates the storage size of each video variant. l Represents a video that a user can request, then all video variants that a user can request can be represented as:

[0076] V={v l |v∈F,l=1,2...L} (1).

[0077] In step 2, calculate the user's content access delay: Assuming that the delay cost of the user obtaining the video file from the SBS n is 0, then

[0078] represents the delay cost of transferring the file from MBS to SBS n, which is set to 7ms;

[0079] The delay cost of transferring the file from the small base station m to the MBS is also 7ms;

[0080] represents the delay cost of transmitting the file from the small base station m to the SBS n via the MBS, that is,

[0081]

[0082] Calculated The latency cost is 14ms.

[0083] The delay cost incurred when transferring files from the original content server to the MBS is 80ms.

[0084] represents the delay cost of transferring files from the original content server to SBS n, i.e.

[0085]

[0086] Calculated The latency cost is 87ms.

[0087] It can be seen that: That is, transferring files from the original content server to the MBS incurs a higher latency cost compared to transferring files between the MBS and the SBS.

[0088] In step 3, set the variable When its value is 1, it means v l Cache is on SBS n or MBS. It does not exist when it is equal to 0.

[0089] In a heterogeneous MEC network, an edge cache is deployed in SBS n, denoted as C n , the cache capacity is M n .

[0090] Deploy an edge cache in MBS, denoted as C N+1 , the cache capacity is M N+1 A feasible cache placement method is represented as C = {C1, C2, ... C N ,C N+1},

[0091] SBS n must satisfy the capacity constraint:

[0092]

[0093] MBS must meet capacity constraints:

[0094]

[0095] The state of the video is represented by a binary variable, where

[0096] Indicates that the user on SBS n sends a video l The request can be served by the small base station m without the need for transcoding.

[0097] Indicates that the user on SBS n sends a video l The request can be served by the small base station m, but transcoding is required.

[0098] Indicates that the user on SBS n sends a videol The request can be served by MBS without transcoding.

[0099] Indicates that the user on SBS n sends a video l The request can be served by MBS, but needs to be transcoded.

[0100] Indicates that the user on SBS n sends a video l The request can be serviced by the original content server.

[0101] In order to satisfy each user's video request, the following constraints need to be met:

[0102]

[0103] Step 4: Calculate the average latency cost of all video requests.

[0104] Joint Collaborative Caching and Scheduling in Mobile Edge Computing Networks,An optimization problem is formulated that aims to minimize the average delay cost of all video requests.

[0105] Given the available resources (cache storage and video processing power), the goal is to jointly determine the cache placement method C and the video request scheduling method To minimize the average delay cost, it is calculated as follows: in, It is the collection of video categories requested by users during the use of SBS n.

[0106] In order to consider the user's dynamics, It represents the set of video requests at SBSn at each time t whenever a new video request arrives, that is, Therefore, the optimal solution that minimizes the average delay cost is obtained by solving the static joint cooperative caching and processing problem at each time t. The objective function is

[0107]

[0108]

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115]

[0116]

[0117]

[0118] In step 5, we design a cache placement method and video request scheduling method for the edge server, obtaining cache information for each server and user request scheduling information. The server cache information includes cache placement decisions for macro and small cell base stations, and the request scheduling information includes the conversion between video variants.

[0119] Following the "divide and conquer" approach, the original problem of solving a series of problems is decomposed into a cache placement problem and a request scheduling problem, and the corresponding solution is obtained with the goal of minimizing the average delay cost.

[0120] Take the cache placement problem as an example.

[0121] (1) When the content is not popular, a reactive LRU cache placement method is adopted.

[0122] In this case, a cache placement method must be made after each cache miss (i.e., the requested video is neither available in the cache nor can it be transcoded from other videos in the cache).

[0123] LRU is a reactive caching method that fetches videos from other servers and caches them if there is a cache miss. The reactive LRU cache placement method should be simple so that it can quickly respond to the large number of requests arriving at the system. If the cache is full, LRU will replace the least recently used video in the cache.

[0124] (2) When the content is popular, an active cache placement method is used to maximize the monotone submodule function. Specifically, the active cache placement method has the following steps:

[0125] Step A1: Initialize and place P = {P1, P2, ..., P n ,P N+1}, and cache placement decision set C placement set C = {c1,c2,.c n .,c N.+1},,

[0126] Step A2: Place the decisions that meet the conditions in set P into the corresponding cache set C;

[0127] Step A3: Redefine the objective function in (18a) as h(C), i.e. make

[0128] Step A4: Determine whether the placement set P is empty. If it is not empty, find the largest cache placement decision.

[0129] Step A5: If Then Delete from C and replace P n Delete from set P;

[0130] Step A6: If Then Delete from C and replace P n Delete from set P;

[0131] Step A7: Repeat steps A2 to A6 until the set P is empty.

[0132] In the entire system, the content popularity in each cell can be estimated by implementing complex machine learning and data mining algorithms in the MEC server. The content popularity within each SBS n is set to

[0133]

[0134] in, It represents the user request video v within the communication coverage of SBS n l possibility.

[0135] When solving the cache placement problem, a new variable is introduced To represent the request scheduling variable in problem P1, this variable represents the video v at SBS k l Request for servicing SBS j:

[0136]

[0137] Using the conversion factor according to the above formula Reconstructing the constraints (7g-7j) yields the following constraints:

[0138]

[0139] Therefore, constraint (7b) can be rewritten as:

[0140]

[0141] in, They represent the second and third cases in formula (9) respectively.

[0142] For a given information about video V, we can calculate the video v from SBS n l Average access delay cost when request is served:

[0143]

[0144] The cache placement problem is simplified by abstracting the request scheduling variables (i.e., the conversion factors) and omitting the constraints (7e, 7f). The resulting problem is called the proactive cache placement problem, which aims to minimize the user's expected access latency cost and can be expressed as:

[0145]

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152] Each video has only one bitrate variable, such as L=1, and different videos cannot be transcoded with each other.

[0153] In order to facilitate calculation, a cache placement set P = {P1, P2, ...P N , P N+1 , P N+2} represents the set of all locations that can be cached. n represents the set of all videos that can be cached on SBS n, P n ={f 1n , f 2n ,...f Fn}.

[0154] Further transforming the above problem, first, from constraint (13c) we can get

[0155]

[0156] Then, according to the combination of formula (4) and formula (5), and finally substituting it into formula (12), we can get formula (15), which is expressed as follows:

[0157]

[0158] The above formula can be simplified to:

[0159]

[0160] T n (v l ) is expressed as follows:

[0161]

[0162] It can be observed from formula (16) that T n (v l ) is considered as the optimization variable, that is, it is regarded as the delay cost savings when users request video files on SBS n. Therefore, minimize under the constraint Equivalent to maximizing T n (v l ), so the problem of minimizing the average delay cost of the video requested by the user is transformed into the problem of maximizing the average delay cost saving, which can be expressed as follows:

[0163]

[0164] St: (13c-13g)(18b),

[0165]

[0166] This problem is essentially a knapsack problem, so the optimization problem is NP-complete. It is known that the problem described in Equation (18) is essentially the classic problem of maximizing a monotone submodular set function under matroid constraints. Currently, the reactive LRU cache placement method is commonly used due to its high efficiency, low complexity, and good approximation ratio.

[0167] In the online version of the problem, for a given cache placement decision C, user requests for content are displayed one after another, then the user request scheduling problem at each time t can be expressed as:

[0168]

[0169] Constraints: (7b), (7e)-(7j),

[0170]

[0171] The optimal request scheduling method is to solve problem P2 from the beginning every time a new video request arrives. In this way, the request scheduling decision obtained is optimal at each time t, and will produce the optimal solution in the long run.

[0172] Since P2 is an integer linear programming problem, it can be solved optimally using a standard optimization solver. An optimal request scheduling method is proposed that uses an integer linear programming optimal solver to solve P2 at each time t.

[0173] However, a drawback of this approach is that every time the optimal request scheduling solution is recalculated, it may result in redirecting existing (ongoing) video requests, thereby wasting buffered data at the base station. Furthermore, the complexity of solving problem P2 increases with the number of requests arriving and the number of cache servers, so it is very impractical to re-solve the problem when a large number of requests arrive in a short period of time.

[0174] To overcome the impracticality of optimization methods, an online algorithm for the video request scheduling problem is proposed. Specifically, consider the online version of problem P2, where the video request set Unknown at the beginning, video requests arrive one at a time. At each time t, the request scheduling algorithm must decide how to handle the new incoming request, and once the decision is made, it is irreversible. Once scheduled, the request cannot be redirected. Specifically, the video request scheduling method has the following steps:

[0175] Step B1: For each video request v arriving at SBS n(n∈N) l ,if Then the file v l Transmitted from SBS n’s cache to the user;

[0176] Step B2: Define video v on SBS n l The closest (bitrate) transcodeable variant of is defined as T(n,v l )=v h , and h satisfies the following constraints:

[0177] Step B3: If T(n,v l )≠0 and Then the video T(n,v l ) is converted to v l and transmit it to the user;

[0178] Step B4: If Then the video v l Transmitted from MBS to SBS n and then to the user;

[0179] Step B5: If Then the video is transmitted to the small base station m to the MBS cache and then transmitted to the user through the MBS and SBSn;

[0180] Step B6: If T(N+1,v l )≠0 and Then the video in the MBS cache is transcoded and transmitted to SBS n and then to the user;

[0181] Step B7: If If it is not an empty set, then calculate the remaining computing resources of all eligible SBSs

[0182] Step B8: Find the SBS with the largest remaining processing resources, i.e. Under the constraint S(m)≥0, if it is not an empty set, T(f,v l ) and transcode, transmit the transcoded video to MBS, then transmit it to SBS n, and then transmit it to the user;

[0183] Step B9: If all the above conditions are met, obtain v from the original content server. l , and then transmitted to users through MBS and SBSn.

[0184] Step 6: The mobile edge computing server caches the corresponding video variant and video request scheduling method according to the cache placement method obtained in step 5.

[0185] In summary, the present invention formalizes the adaptive bitrate video caching problem in heterogeneous mobile edge computing networks. First, through the collaborative caching model in heterogeneous mobile edge computing networks, the edge cache of macro base stations and small base stations is mainly used to bring storage resources closer to users; secondly, in order to solve the challenges brought by the original problem, the original problem is divided into two problems using the divide-and-conquer method: the caching problem and the scheduling problem. In the caching problem, when the content popularity is not available, the least recent cache method is adopted; when the content popularity is available, the monotone submodule maximization solution is used to solve the active caching problem in a greedy manner to minimize the cache delay. In the scheduling problem, it is regarded as an online multiple-choice knapsack problem to solve the cache delay in a greedy manner, which ultimately enables the present invention to improve the cache hit rate and reduce content access delay.

[0186] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An adaptive bit rate video caching method, characterized in that: The main steps include: Step 1: Establish a heterogeneous mobile edge computing network with a macro base station and multiple small base stations. Deploy a mobile edge computing server on the macro base station and each small base station, preset the cache placement method, and transcode videos with different bit rates. Step 2: Calculate the user's content access delay; Step 3: Calculate the capacity constraints of the macro base station and the small base station and the constraints of the user's video request; Step 4: Calculate the average delay cost of all video requests; Step 5: Obtain cache information of each server and user request scheduling information through the cache placement method and video request scheduling method; Step 6: The mobile edge computing server caches the corresponding video variants and video request scheduling method according to the cache placement method obtained in step 5 to minimize access delay; In step 3, in the heterogeneous mobile edge computing network, an edge cache C is deployed in the small base station n. n , the cache capacity is M n , deploy an edge cache C in the macro base station N+1 , the cache capacity is M N+1 , the cache placement method is expressed as C={C1,C2,...C N ,C N+1 }, then the capacity constraint of small base station n is: Capacity constraints of macro base stations: in, n∈N∪{N+1},v l ∈V,v l Indicates a video that the user can request, r l represents the storage size of each video variant, and V represents all video variants that the user can request; In step 3, the constraints of the user's video request can be expressed as in, Indicates that the user on the small base station n sends a video v l The request can be served by the small cell m without the need for transcoding; Indicates that the user on the small base station n sends a video v l The request can be served by the small cell m, but needs to be transcoded; Indicates that the user on the small base station n sends a video v l The request can be served by the macro base station without the need for transcoding; Indicates that the user on the small base station n sends a video v l The request can be served by the macro base station, but needs to be transcoded; Indicates that the user on the small base station n sends a video v l The request can be served by the original content server; In the caching problem, when content popularity is not available, the least recent cache method is adopted; when content popularity is available, the monotone submodule maximization method is used to solve the active caching problem in a greedy manner to minimize the cache delay; in the scheduling problem, it is regarded as an online multiple-choice knapsack problem and solved in a greedy manner to minimize the cache delay.

2. The adaptive bitrate video caching method according to claim 1, wherein: In step 4, the average delay cost is in, It is the set of video types requested by users during the use of small cell n.

3. The adaptive bitrate video caching method according to claim 2, wherein: In step 4, the objective function of the optimal solution to minimize the average delay cost within each time t is 4. The adaptive bitrate video caching method according to claim 1, wherein: In step 5, the cache information of the server includes cache placement decision information of the macro base station and the small base station, and the request scheduling information includes conversion between video variants.

5. The adaptive bitrate video caching method according to claim 1, wherein: The cache placement method includes a reactive LRU cache placement method and a proactive cache placement method.

6. The adaptive bit rate video caching method according to claim 5, wherein: The specific steps of the active cache placement method are as follows: Step A1: Initialize and place P = {P1, P2, ..., P n ,P N+1 }, and cache placement decision set C placement set C = {c1,c2,...,c n ,c N+1 }, Step A2: Place the decisions that meet the conditions in set P into the corresponding cache set C; Step A3: Redefine the objective function as h(C), that is make Step A4: Determine whether the placement set P is empty. If it is not empty, find the largest cache placement decision. Step A5: If Then Delete from C and replace P n Delete from set P; Step A6: If Then Delete from C and replace P n Delete from set P; Step A7: Repeat steps A2 to A6 until the set P is empty.

7. The adaptive bit rate video caching method according to claim 1, wherein: The specific steps of the video request scheduling method are as follows: Step B1: For each video request v arriving at the small cell n, n∈N l ,if Then the file v l Transmitted from the cache of small base station n to the user; Step B2: Define video v on small base station n l The closest transcodeable variant is defined as T(n,v l )=v h , and h satisfies the following constraints: Step B3: If T(n,v l )≠0 and Then the video T(n,v l ) is converted to v l and transmit it to the user; Step B4: If Then the video v l Transmitted from the macro base station to the small base station n and then transmitted to the user; Step B5: If Then the video is transmitted to the cache of the small base station m and then transmitted to the user through the macro base station and small base station n; Step B6: If T(N+1,v l )≠0 and Then the video in the macro base station cache is transcoded and transmitted to the small base station n and then transmitted to the user; Step B7: If If it is not an empty set, then calculate the remaining computing resources of all qualified small base stations Step B8: Find the small base station with the largest remaining processing resources, i.e. Under the constraint of S(m)≥0, if it is not an empty set, T(f,v l ) and transcode, transmit the transcoded video to the macro base station, then transmit it to the small base station n, and then transmit it to the user; Step B9: If all the above conditions are met, obtain v from the original content server. l , and then transmitted to users through macro base stations and small base stations n.

8. An adaptive bitrate video caching system, characterized by: Apply the adaptive bit rate video caching method as described in any one of claims 1 to 7.

Images