WO2010043146A1

WO2010043146A1 - Method and device, server cluster for media file storage processing and service processing

Info

Publication number: WO2010043146A1
Application number: PCT/CN2009/074114
Authority: WO
Inventors: 韩磊; 李晋; 梁超; 刘勇; 基思·罗斯
Original assignee: 华为技术有限公司
Priority date: 2008-10-14
Filing date: 2009-09-22
Publication date: 2010-04-22
Also published as: CN101729357A; CN101729357B

Abstract

A method and device, server cluster for media file storage processing and service processing are disclosed in the embodiments of the present invention. Wherein, the method for media file storage processing includes: determining a heat level of a media file according to a heat of the media file; determining the amount M of servers corresponding to the heat level of the media file according to the correspondent relationship between the preset heat level and the amount of servers storing the same media file; encoding the media files into M sub-streams of the same bandwidth; and in accordance with the principles of the server load balancing, storing the M sub-streams of the same bandwidth on M servers of the server cluster respectively. By means of the embodiments of the present invention, load sharing and load balancing can be realized among servers in the server cluster, and the service ability of each server is used effectively.

Description

Media file storage processing and business processing method and device, server cluster

The present invention relates to a load balancing technology, and more particularly to a media file storage processing and service processing method and apparatus, and a server cluster. Background technique

Internet Protocol Television (hereinafter referred to as: ΙΡΤ V) is a home TV, personal computer (PC) or other electronic equipment for the display network, providing TV programs and other value-added services based on TV programs. A wide range of audio-visual broadband IP multimedia information services, such as live TV and on-demand TV.

In the IPTV service system, the server is responsible for storing media files such as movies. When the server receives a media file request sent by the client, it provides the client with the media file it requested. The serviceability of each server is limited. For example: A server can access 500 to 1000 users at the same time, or can generate 500M bandwidth at the same time. When the server's serviceability is full, there is no way to provide more services. For example: A server can access 500 users at the same time. After 500 users have access, the service request of the 501st user will be rejected.

Server clustering technology can bring together several servers and provide services like a server. In a cluster of servers that are clustered together, one or several servers act as scheduling servers to schedule servers in the server cluster, and select appropriate servers to process media file requests for service to users. For example: A server cluster consists of 20 servers, each of which can access 500 users at the same time. The server cluster can access 10,000 users at the same time. The media files requested by each user may be different when accessed. If the 20 servers in the cluster store Movies 8, B, and CT, respectively, the dispatch server will schedule the server that stores the corresponding movie to provide the user with the requested movie. At this time, because the heat of different media files is different, that is: popular with users. Different degrees, the server that causes the media files with higher storage heat is accessed by users more frequently, and the load is larger. However, some servers that store media files with lower heat are accessed by users with less frequency and less load, which may occur. The ability to store servers with hot media files cannot meet the user's service needs and cannot serve more users, while the ability to store less hot media files is underutilized. To solve this problem, you can redundantly store some hot media files between servers, for example: Store the hottest videos on all servers. Thus, when the service capability of a server reaches the limit, the user can go to other servers to obtain the requested media file. In addition, redundant storage between servers also implements backup of media files. When a server fails, the redundant storage server can serve as a backup server to replace the failed server to provide services to users. Normally, when the load on the server is small, the failure rate is also small. Conversely, when the load is high, the failure rate is high. Therefore, in order to ensure the stability of the server, load sharing scheduling between redundantly stored servers is required to achieve load balancing between servers of redundant storage.

However, when redundant storage in a server cluster, it is inevitable to reduce the types of media files that can be provided by the entire server cluster. Therefore, the redundant storage of server clusters and the utilization efficiency of service capabilities are a pair of contradictory factors. When the redundancy of storage is small, the service capabilities of the server cluster cannot be effectively utilized. When the server cluster's service capabilities are effectively utilized, the storage redundancy is greater. Therefore, in a server cluster, how to achieve a balance between redundant storage and service utilization efficiency is an urgent problem to be solved.

In one of the prior art, each edge server (Edge Server, hereinafter referred to as ES) in the existing IPTV system is organized into a server cluster to provide media file services at the level of the IPTV system. Each ES serves a user in a certain area. When the user requests a media file, the ES of the area where the user is located is selected to provide services for the user. The media file request of the user in the ES local area of the user's area is selected, and the media file of the user's interest is selected and stored locally and provided to the user. There is no sharing between ESs in the server cluster. Users at different times may be unbalanced at different times. This results in a heavy ES load in a certain area. The ES load in a certain area is light and cannot be achieved between ESs. Load sharing and negative Load balancing. In addition, the ES of each area needs to store all the media files required by users in the local area. If a small number of users interested in some less hot media files are scattered in different areas, ESs in different areas will satisfy a small number of users in the domain. The request needs to save the corresponding media file, which inevitably leads to unnecessary redundant storage, thereby reducing the kind of media files that the entire server cluster can provide.

In another prior art, the ESs are organized in a Peer-to-Peer (P2P) manner to form a P2P ES cluster, and resource sharing is implemented between the ESs. Through the P2P query in the P2P ES cluster, the distribution of media files in all ESs can be obtained. The P2P server

The ES cluster provides a method for querying media files. Compared with the previous prior art, the content sharing between ESs is implemented, and a fully distributed structure is adopted in the P2P ES cluster, thereby avoiding a single point of failure for the entire The impact of the cluster system increases the robustness of the system. However, load sharing and load balancing between ESs is still not possible in a P2P ES cluster. For example: If a hot movie is only stored on one of the ESs, all requests for that movie will still be concentrated on the ES, causing a hot spot effect.

In the process of implementing the present invention, the inventors have found that: In the prior art, in a P2P-based server cluster, although internal scheduling of the server is implemented, an appropriate server can be selected to serve the user. Among them, the P2P-based server cluster is also referred to as a P2P server cluster. However, how to perform redundant storage between servers to achieve load sharing and load balancing between servers and to effectively utilize the service capabilities of each server is not possible in a P2P server cluster. Summary of the invention

The embodiments of the present invention provide a media file storage processing and service processing method and device, and a server cluster, which implement load sharing and load balancing between servers in a server cluster, and effectively utilize the service capabilities of each server.

A media file storage processing method provided by an embodiment of the present invention includes: Determining the popularity level of the media file according to the popularity of the media file;

Determining, according to a correspondence between a preset heat level and a number of servers storing the same media file, a number of servers corresponding to the heat level of the media file;

Encoding the media file into M sub-streams occupying the same bandwidth;

According to the principle of server load balancing, the M sub-streams occupying the same bandwidth are respectively stored on M servers in the server cluster.

A media file storage processing apparatus provided by an embodiment of the present invention includes:

a heat identification module, configured to determine a heat level corresponding to the heat of the media file according to the correspondence information between the heat and the heat level set in advance;

a determining module, configured to determine, according to a correspondence between a preset heat level and a number of servers storing the same media file, a number M of servers corresponding to a heat level of the media file; and a coding unit, configured to encode the media file The M sub-streams occupying the same bandwidth; the storage allocation module is configured to send the M sub-streams with the same occupied bandwidth to the M servers in the server cluster according to the server load balancing principle.

A server cluster provided by an embodiment of the present invention includes:

a media file storage processing device, configured to determine a heat level corresponding to the heat of the media file according to the correspondence information between the heat and the heat level set in advance, according to a preset heat level and a number of servers storing the same media file Corresponding relationship, determining the number M of servers corresponding to the heat level of the media file, and encoding the media file into M sub-streams with the same occupied bandwidth, and according to the principle of server load balancing, the M sub-streams occupying the same bandwidth Sent to M server storage;

L servers, the M servers of the L servers are respectively configured to store one of the M sub-streams with the same occupied bandwidth encoded by the media file according to the server load balancing principle, and receive the media sent by the terminal. a file obtaining request, and sending, to the terminal, a substream of the media file corresponding to the file identifier in the media file obtaining request, where M is an integer greater than 1, and M is less than or equal to L. Another server cluster provided by the embodiment of the present invention includes L servers, and one of the L servers is configured to determine the heat corresponding to the heat of the media file according to the correspondence information between the preset heat and heat levels. Level, determining a number M of servers corresponding to the heat level of the media file according to a correspondence between a preset heat level and a number of servers storing the same media file, and encoding the media file into M occupied bandwidths Sub-flow, and according to the principle of server load balancing, storing M sub-streams having the same occupied bandwidth respectively on M servers in the L servers, M is an integer greater than 1, and M is less than or equal to L; The M servers of the L servers are respectively configured to store one of the M sub-streams with the same occupied bandwidth encoded by the media file according to the server load balancing principle, and receive the media file acquisition request sent by the terminal, and Subsequent to the stored media file corresponding to the file identifier in the media file acquisition request, To the terminal.

A media file service processing method provided by the embodiment of the present invention includes:

Receiving a media file acquisition request sent by the terminal, where the media file acquisition request carries a media file identifier and an end user identifier;

The M servers in the server cluster respectively send the substreams of the media files stored to the terminal to the terminal according to the end user identifier, where the media files are encoded into M occupied bandwidths according to the heat degree. The substreams are respectively stored on the M servers according to the server load balancing principle, and M is an integer greater than 1.

A device for processing a media file service according to an embodiment of the present invention includes:

The first receiving module is configured to receive a media file obtaining request sent by the terminal, where the media file obtaining request carries the media file identifier and the terminal user identifier;

a querying module, configured to query, according to the media file identifier, an address of the M servers that store the M sub-streams of the media file in the server cluster; and an acquiring module, configured to: according to the addresses of the M servers, respectively Obtaining N substreams of the media file in the M servers, where N is an integer greater than 1 and N is less than or equal to M;

a sending module, configured to be separately from the M servers according to the terminal user identifier The N sub-streams of the media file that are obtained and corresponding to the media file identifier are sent to the terminal; wherein the media file is encoded into M sub-streams with the same occupied bandwidth according to the heat, and is load-balanced according to the server. The principles are stored on the M servers, respectively, and are integers greater than one.

Another server cluster provided by the embodiment of the present invention includes:

L servers, the servers of the L servers are respectively used to store one substream of the same sub-stream with the same occupied bandwidth encoded by the same media file according to the principle of server load balancing, and Μ is an integer greater than 1. , and Μ is less than or equal to L;

The media file processing device is configured to receive a media file acquisition request sent by the terminal, where the media file acquisition request carries the media file identifier and the terminal user identifier, and is stored in the server cluster according to the terminal user identifier. The sub-stream of the media file corresponding to the media file identifier is sent to the terminal. According to the heat of media files, according to the principle of load balancing between servers in the server cluster, the media files are encoded into two sub-streams with the same bandwidth and stored in one server in the server cluster. When the server cluster receives the media file acquisition request sent by the terminal, respectively, sending the sub-streams of the media files stored on the server to the terminal, and the terminal may restore the multiple sub-streams according to the received sub-flows. Media files, which realize load sharing and load balancing between servers in a server cluster, so that the service capabilities of each server can be effectively utilized.

The technical solution of the present invention will be further described in detail below through the accompanying drawings and embodiments. DRAWINGS

1 is a flowchart of an embodiment of a media file storage processing method according to the present invention;

2 is a schematic structural diagram of an embodiment of a media file storage processing apparatus according to the present invention; FIG. 3 is a flowchart of an embodiment of a media file service processing method according to the present invention;

4 is a flowchart of another embodiment of a method for processing a media file service according to the present invention; 5 is a flowchart of still another embodiment of a media file service processing method according to the present invention; FIG. 6 is a schematic structural diagram of an embodiment of a media file service processing apparatus according to the present invention;

8 is a schematic structural diagram of another embodiment of a server cluster according to the present invention;

9 is a schematic structural diagram of still another embodiment of a server cluster according to the present invention;

FIG. 10 is a schematic structural diagram of still another embodiment of a server cluster according to the present invention. detailed description

The media file storage processing method provided by the embodiment of the present invention may encode the media file into M sub-streams with the same occupied bandwidth according to the heat of the media file, and according to the server load balancing principle, the M occupied bandwidths are the same. The substreams are stored separately on M servers in the server cluster. Specifically, in the following embodiments of the present invention, the media file may be encoded by a practical network coding method. Because the media files are encoded according to the heat, the M sub-streams with the same bandwidth are encoded by the practical network coding method, and are stored on the M servers in the server cluster according to the principle of load balancing among the servers in the server cluster. When receiving the media file acquisition request sent by the terminal, the M substreams of the media files stored on the M servers are respectively sent to the terminal, thereby realizing load sharing and load balancing between the servers in the server cluster, so that each server The service capabilities are effectively utilized.

As shown in FIG. 1 , it is a flowchart of an embodiment of a media file storage processing method according to the present invention, which includes the following steps:

Step 101: Determine a heat level of the media file according to the heat of the media file.

Specifically, the hotness of the media file may be determined by counting the number of times the media file is accessed by the user terminal for a certain period of time, and the heat level of the media file is further determined according to the heat range included in the preset heat level.

Step 102: Determine, according to a correspondence between a preset heat level and a number of servers storing the same media file, the number M of servers corresponding to the heat level of the media file. As a specific embodiment of the present invention, the heat may be sequentially decreased according to the heat level such as the first heat level, the second heat level, and the third heat level, that is, the higher the heat level, the lower the heat. The number of servers storing the media files of the first heat level with the highest heat is the number L of servers constituting the server cluster, L is an integer greater than or equal to M; the number of servers storing media files of other heat levels other than the first heat level is P , P is an integer less than L. Specifically, P may take an integer value in L/Q or L/2Q- ¹ , and Q is a heat level.

The higher the heat level, the more server resources the media files occupy. The more media files with higher heat levels are encoded into more sub-flows and stored on each server, which can make the servers in the server cluster better. Load sharing and load balancing, and encoding the media files with lower heat levels into fewer sub-flows and storing them on each server, can reduce the waste of server resources caused by redundant storage in the server cluster, so that each The server's service capabilities are fully utilized.

Step 103: The media file is encoded into M sub-streams with the same occupied bandwidth by using a practical network coding method.

Step 104: According to the server load balancing principle, the M sub-streams with the same occupied bandwidth are respectively stored on the M servers in the server cluster.

After the step 104 of the embodiment shown in FIG. 1, the address of the M servers storing the substreams of the media file may be recorded, so that after receiving the media file acquisition request sent by the user terminal, the server cluster queries and stores the substreams of the media file. The address of the server, which gets the subflow of the media file from the corresponding server. Specifically, the address of the M servers storing the substreams of the media file may be recorded in a manner corresponding to the correspondence between the media file identifier and the address of the server, which may be implemented by: establishing a media file identifier and an address of the server. Corresponding relationship information, and storing the correspondence information in each server in the server cluster; establishing correspondence information between the media file identifier and the address of the server, and dividing the correspondence information into multiple parts, Stored on a plurality of different servers in the server cluster; establish correspondence information between the media file identifier and the address of the server, and store the correspondence information on one or more specific servers in the server cluster, or Distributed across multiple specific servers. Because different media files have different enthusiasm. Depending on the degree of heat, the heat of the media files can be graded, and the different media files are divided into a plurality of different heat levels, for example: the first heat level, the second heat level and the third heat level have three heat levels. It is possible to set the media file of the first heat level to be the hottest, the media file of the second heat level to be hot, and so on. Suppose there are L servers in a server cluster, and L is an integer greater than 1. The following is an example in which the media files are divided into three heat levels, and the media files are encoded, and the encoded M sub-streams are respectively stored on the M servers according to the server load balancing principle. For other heat levels, reference may be made. deal with.

First, the media files are divided into different heat levels. Specifically, the popularity level of each media file may be determined according to the frequency at which each media file is accessed by the user for a certain period of time. For example: a frequency greater than or equal to 100 is the first heat level, a frequency between 50 and 100 is the second heat level, and a frequency less than or equal to 50 is the third heat level. In addition, the heat level of the media file may be divided according to the frequency of the media files accessed by the user in a certain period of time and the preset number of media files included in each heat level, wherein the number of media files included in each heat level may be based on The actual situation needs to be updated. For example: The preset first heat level contains 10 media files, the second heat level contains 20 media files, and the remaining media files belong to the third heat level, then sorted according to the frequency of user accesses from high to low for a certain period of time. The first 10 media files belong to the first heat level, and the 11th to 30th media files sorted according to the user access frequency from high to low belong to the second heat level, and the remaining media files belong to the third heat level. After dividing the heat level of the media file, it is assumed that there are X media files of the first heat level, Y media files of the second heat level, and Z media files of the third heat level.

Each of the X media files of the first heat level is encoded into L sub-streams occupying the same bandwidth and stored on L servers in the server cluster. According to the principle of the practical network coding, the terminal only needs to obtain a plurality of substreams of the same media file from the L servers, and the media files can be restored by encoding and combining the substreams. In this way, even if the server in the server cluster is down, as long as it is not a large area, for example: After more than eighty servers are down, the user terminal can obtain enough substreams of the same media file to restore the media files.

Since the sub-stream of the media file of the first heat level is stored on each server in the server cluster, when the end user requests to obtain the media file, a sub-flow of the media file may be separately obtained from each server in the server cluster. Since the server bandwidth occupied by each sub-flow is 1/L of the bandwidth occupied by the media file, the bandwidth consumed by each server to support one media file acquisition request only accounts for 1/L of the bandwidth required for the transmission of the media file. L servers together provide complete media file transfer bandwidth. For each media file acquisition request, load sharing and load balancing are implemented between servers in the server cluster. In general, the media file acquisition request between the servers for the first level of heat is always load-sharing, and the load balancing between servers is realized.

For each of the Y media files of the second heat level, if the number of servers in the server cluster is even, that is, L is an even number, all are encoded into L/2 sub-streams occupying the same bandwidth, each will The media files are stored on L/2 servers in the server cluster. Specifically, the L/2 sub-streams of the first media file in the second heat level may be respectively stored in any L/2 servers in the server cluster, and the L/2 sub-streams of the second media file are separately stored. On another L/2 servers in the server cluster; L/2 subflows of the third media file are stored on any L/2 servers in the server cluster, and L/2 subflows of the fourth media file They are stored separately on the other L/2 servers in the server cluster, and so on. If the number of servers in the server cluster is an odd number, that is, L is an odd number, the Y media files of the second heat level are coded into (L-1)/2 and (L+1)/2 occupied bandwidths. Subflow, which stores each media file evenly on (L-1)/2 and (L+1)/2 servers in the server cluster. Specifically, the (L-1)/2 substreams of the first media file in the second heat level may be respectively stored in any (L-1)/2 servers in the server cluster, and the second media file is The encoded (L+1)/2 substreams are stored separately on the other (L+1)/2 servers in the server cluster; the third media file is encoded into (L-1)/2 substreams Any (L-1) stored in the server cluster 12 On the server, the (L+1)/2 substreams into which the fourth media file is encoded are stored separately on the other (L+1)/2 servers in the server cluster, and so on. According to the principle of the practical network coding, the terminal only needs to obtain a plurality of substreams of the same media file from the L servers, and the media files can be restored by encoding and combining the substreams. In this way, even if the server in the server cluster is down, as long as it is not a large-scale downtime, for example: more than 80% of the servers that store the same media file substream are down, the user terminal can get enough of the same media file. Substreams, thereby restoring media files.

For the media file acquisition request of the second heat level, there are L/2 or (L-1)/2 or (L+1)/2 servers storing the media file substream to provide complete media file transmission. Bandwidth, for each media file acquisition request, load sharing and load balancing between L/2 or (L-1) /2 or (L+1) /2 servers in the server cluster. From a global perspective, since the heat of different media files in the same heat level is different, for example: the heat of media file A is high, and 1000 end users request to obtain the media file, then the 1000 media file acquisition request Provided by 10 servers where the sub-stream of the media file A is located, each server needs to serve the bandwidth of 100 media files, the heat of the media file B is low, and 800 end users request to obtain the media file, then the 800 media The file acquisition request is provided by the 10 servers where the media file B is located. Each server needs to serve the bandwidth of 80 media files. In addition, the overall bandwidth occupied by each media file may also be different, which makes it possible to store different media of the same heat level. Load differences and load differences between servers in the file subflow. Therefore, when placing the second heat level media file substream on the server, the total heat of the media file and the required occupied bandwidth should be considered, and the total heat of the media files on each server and the required occupied bandwidth are approximately equal to achieve Better load sharing and load balancing.

Relative to the media file of the first heat level, because only L/2 or (Ll) /2 or (L+1) /2 servers store substreams of the same media file, the same media can be provided to the end user. The server size of the file is reduced, and the impact of some server downtime on the quality of the media file is relatively large. Therefore, the robustness of the server cluster system is reduced, but due to the second level of heat The degree of access to the media files is relatively low. Compared with the media files of the first heat level, the redundant storage reduces the waste of the server storage space resources, and the service capabilities of the servers are effectively utilized.

For the Z media files of the third heat level, view the remaining capacity of each server in the server cluster. The remaining capacity may be the total bandwidth supported by the server and the total bandwidth occupied by the media stream sub-streams, and the remaining capacity is selected. A server with a capacity greater than a preset threshold or a few servers with the highest remaining capacity, assuming R. The Z media files of the three-heat level are respectively encoded into R sub-streams and stored in R servers. According to the principle of the practical network coding, the terminal only needs to obtain a plurality of substreams of the same media file from the R servers, and the media files can be restored by encoding and combining the substreams. In this way, even if the server in the server cluster is down, as long as it is not a large-scale downtime, for example: more than 80% of the servers that store the same media file substream are down, the user terminal can get enough of the same media file. Substreams, thereby restoring media files.

Obviously, the media file of the third heat level, compared with the media files of the first heat level and the second heat level, the number of servers serving the end user cannot be guaranteed, and correspondingly, the load sharing and the system between the servers The robustness effect is also relatively poor. However, due to the low frequency of access to media files of the third heat level, the impact on load sharing between the entire servers is not very large. Moreover, since the third-level hot media file is accessed at a relatively low frequency, the redundant storage reduces the waste of the server storage space resources compared to the media files of the first heat level and the second heat level, so that each The server's service capabilities are effectively utilized.

FIG. 2 is a schematic structural diagram of an embodiment of a media file storage processing apparatus according to the present invention. The media file storage processing apparatus of this embodiment may be used to implement the foregoing embodiment of the media file storage processing embodiment of the present invention, which includes a popularity identification module 201, a determination module 202, an encoding module 203, and a storage allocation module 204. The heat identification unit 201 is configured to determine a heat level corresponding to the heat of the media file according to the correspondence information between the heat and the heat level set in advance. Indeed The determining module 202 is configured to determine the number M of servers corresponding to the heat level of the media file according to the correspondence between the preset heat level and the number of servers storing the same media file. The encoding module 203 is configured to encode the media file into M sub-streams with the same occupied bandwidth. The storage allocation module 204 is configured to send the M sub-streams with the same occupied bandwidth encoded by the encoding module 203 to the M servers in the server cluster according to the server load balancing principle.

As a specific embodiment of the present invention, the heat may be sequentially decreased according to the heat level such as the first heat level, the second heat level, and the third heat level, that is, the higher the heat level, the lower the heat. The number of servers storing the media files of the first heat level with the highest heat is the number L of servers constituting the server cluster, L is an integer greater than or equal to M; the number of servers storing media files of other heat levels other than the first heat level is P , P is an integer less than L. Specifically, P may take an integer value in L/Q or L/2Q- ¹ , and Q is a heat level.

Further, as shown in FIG. 2, the media file storage processing apparatus of the embodiment of the present invention may further include a recording module 205 for recording an address of the M servers storing the substreams of the media file.

A server cluster provided by the embodiment of the present invention can be used to implement the media file storage processing method of the foregoing embodiment, which includes a media file storage processing device and L servers. The M servers of the L servers are respectively configured to store one substream of the M sub-streams with the same occupied bandwidth that the same media file is encoded according to the server load balancing principle, and receive the media file acquisition request sent by the terminal. And sending, to the terminal, a substream of the media file corresponding to the file identifier in the media file obtaining request, where M is an integer greater than 1, and M is less than or equal to L. The media file storage processing device is configured to encode the same media file into M sub-streams with the same occupied bandwidth, and send the M sub-streams with the same occupied bandwidth to the M servers for storage according to the server load balancing principle. Specifically, the media file storage processing device can be implemented by a media center in a server cluster.

Another server cluster provided by the embodiment of the present invention may be used to implement the media file storage processing method of the foregoing embodiment, which includes L servers, and one of the L servers is used to encode the same media file into M occupations. Substreams with the same bandwidth, and load balancing according to the server Then, M sub-streams occupying the same bandwidth are respectively stored on M servers in the L servers, M is an integer greater than 1, and M is less than or equal to L. The M servers of the L servers are respectively configured to store one substream of the M sub-streams with the same occupied bandwidth that the same media file is encoded according to the server load balancing principle, receive the media file acquisition request sent by the terminal, and The sub-streams of the media files corresponding to the file identifiers in the media file acquisition request are respectively sent to the terminal.

A media file service processing method provided by the embodiment of the present invention includes: receiving a media file acquisition request sent by a terminal, where the media file acquisition request carries a media file identifier and a terminal user identifier; M of the M substreams in the server cluster The servers respectively send substreams of the respective stored media files to the terminal according to the end user identifier. The media files are encoded into M sub-streams with the same bandwidth according to the heat, and are respectively stored on the above M servers according to the server load balancing principle, where M is an integer greater than 1. The utility network coding method is a network media coding method, which can encode a media file into a plurality of substreams, and arbitrarily receive a part of the substreams of the plurality of substreams, and decode and combine the partial substreams. The original media file is restored for playback, and the playback quality is not affected. As an embodiment, the server cluster in the embodiment of the present invention is a P2P server cluster formed by organizing a plurality of servers together in a P2P manner.

Because the media files are pre-according to the heat, they are encoded into M sub-streams with the same bandwidth and are stored in the M servers in the server cluster according to the principle of load balancing between servers in the server cluster. When the media file is obtained, the M substreams of the media files stored on the M servers are respectively sent to the terminal, thereby realizing load sharing and load balancing between the servers in the server cluster, so that the service capabilities of the servers are effective. use.

As shown in FIG. 3, it is a flowchart of an embodiment of a method for processing a media file service according to the present invention, which includes the following steps:

Step 301: The server cluster receives a media file acquisition request sent by the terminal, where the media file acquisition request carries the media file identifier and the terminal user identifier. Media file identification in a service The device cluster uniquely identifies a media file, which may be one or more of the name of the media file, the serial number of the media file, the storage time of the media file, and the bandwidth required by the media file.

Step 302: The server that receives the media file acquisition request in the server cluster queries the address of the M servers corresponding to the media file identifier in the media file acquisition request. The media files are encoded into M sub-streams with the same bandwidth according to the heat, and are stored on the above M servers according to the server load balancing principle, where M is an integer greater than 1.

Specifically, the correspondence information between the media file identifier and the address of the server may be separately stored on each server in the server cluster, so that the server that receives the media file acquisition request sent by the terminal queries the media file identifier stored by the server. The correspondence information between the addresses of the servers can obtain the addresses of the M servers corresponding to the media file identifiers in the media file acquisition request. In addition, the correspondence information between the media file identifier and the address of the server may also be divided into multiple parts, which are respectively stored on a plurality of different servers in the server cluster, and the server that receives the media file acquisition request sent by the terminal is in itself When the address of the server corresponding to the media file identifier in the media file acquisition request is not stored, the information is exchanged with other servers, and the addresses of the M servers corresponding to the media file identifier in the media file acquisition request are obtained from other servers. Alternatively, the correspondence information between the media file identifier and the address of the server may be all stored on one or more specific servers in the server cluster, or distributed on multiple specific servers, and the media file sent by the terminal is received. The requested server obtains the address of the M servers corresponding to the media file identifier in the media file acquisition request from the specific server.

Step 303: The server that receives the media file acquisition request in the server cluster forwards the media file acquisition request to the corresponding M servers according to the addresses of the M servers.

If the server receiving the media file acquisition request is one of the M servers, there is no need to forward the media file acquisition request to itself.

Step 304: The M servers that receive the media file acquisition request send the substreams of the stored media files to the terminal according to the terminal user identifier. After receiving the multiple substreams of the media file returned by the server cluster, the terminal may decode and combine the multiple substreams, and then restore the original media file and play it.

As shown in FIG. 4, it is a flowchart of another embodiment of a method for processing a media file service according to the present invention, which includes the following steps:

Step 401: The server cluster receives a media file acquisition request sent by the terminal, where the media file acquisition request carries the media file identifier and the terminal user identifier. The media file identifier uniquely identifies a media file in a server cluster, and may specifically be one or more of a name of the media file, a serial number of the media file, a storage time of the media file, and a bandwidth required for the media file. .

Step 402: The server that receives the media file acquisition request in the server cluster queries the address of the M servers corresponding to the media file identifier in the media file acquisition request. The media files are encoded into M sub-streams with the same bandwidth according to the heat, and are stored on the above M servers according to the server load balancing principle, where M is an integer greater than 1.

Specifically, the correspondence information between the media file identifier and the address of the server may be separately stored on each server in the server cluster, so that the server that receives the media file acquisition request sent by the terminal queries the media file identifier stored by the server. The correspondence information between the addresses of the servers can obtain the addresses of the M servers corresponding to the media file identifiers in the media file acquisition request. In addition, the correspondence information between the media file identifier and the address of the server may also be divided into multiple parts, which are respectively stored on a plurality of different servers in the server cluster, and the server that receives the media file acquisition request sent by the terminal is in itself When the address of the server corresponding to the media file identifier in the media file acquisition request is not stored, the information is exchanged with other servers, and the addresses of the M servers corresponding to the media file identifier in the media file acquisition request are obtained from other servers. Alternatively, the correspondence information between the media file identifier and the address of the server may be all stored on one or more specific servers in the server cluster, or distributed on multiple specific servers, and the media file sent by the terminal is received. The requested server obtains the address of the M servers corresponding to the media file identifier in the media file acquisition request from the specific server. Step 403: The server that receives the media file acquisition request in the server cluster obtains the N substreams of the media file from the M servers according to the addresses of the M servers, and is an integer greater than 1 and N is less than or equal to M.

Step 404: The server that receives the media file acquisition request sends the N substreams of the obtained media file to the terminal respectively according to the terminal user identifier.

After receiving the multiple substreams of the media file returned by the server cluster, the terminal can decode and combine the multiple substreams to restore the original media file and play it.

As shown in FIG. 5, it is a flowchart of still another embodiment of a media file service processing method according to the present invention, which includes the following steps:

Step 501: The server cluster receives a media file acquisition request sent by the terminal, where the media file acquisition request carries the media file identifier and the terminal user identifier. The media file identifier uniquely identifies a media file in a server cluster, and may specifically be one or more of a name of the media file, a serial number of the media file, a storage time of the media file, and a bandwidth required for the media file. .

Step 502: The server that receives the media file acquisition request in the server cluster queries the address of the M servers corresponding to the media file identifier in the media file acquisition request. The media files are encoded into M sub-streams with the same bandwidth according to the heat, and are stored on the above M servers according to the server load balancing principle, where M is an integer greater than 1.

Specifically, the correspondence information between the media file identifier and the address of the server may be separately stored on each server in the server cluster, so that the server that receives the media file acquisition request sent by the terminal queries the media file identifier stored by the server. The correspondence information between the addresses of the servers can obtain the addresses of the M servers corresponding to the media file identifiers in the media file acquisition request. In addition, the correspondence information between the media file identifier and the address of the server may also be divided into multiple parts, which are respectively stored on a plurality of different servers in the server cluster, and received by the terminal. The server that obtains the media file acquisition request does not store the address of the server corresponding to the media file identifier in the media file acquisition request, and performs information interaction with other servers, and obtains the media file identifier corresponding to the media file acquisition request from other servers. The address of the M servers. Alternatively, the correspondence information between the media file identifier and the address of the server may be all stored on one or more specific servers in the server cluster, or distributed on multiple specific servers, and the media file sent by the terminal is received. The requested server obtains the address of the M servers corresponding to the media file identifier in the media file acquisition request from the specific server.

Step 503: The server that receives the media file acquisition request in the server cluster sends the indication information to the corresponding M servers according to the addresses of the M servers, and instructs the sub-flows of the media files respectively stored by the M servers to be sent to the terminal. The indication information carries an end user identifier.

If the server receiving the media file acquisition request is one of the M servers, it is not necessary to send the indication information to itself.

Step 504: The M servers that receive the indication information respectively send the N sub-streams of the respective stored media files to the terminal according to the terminal user identifier.

FIG. 6 is a schematic structural diagram of an embodiment of a media file service processing apparatus according to the present invention. The media file service processing apparatus of the embodiment may be used to implement the foregoing media file service processing method embodiment shown in FIG. A receiving module 601, a query module 602, an obtaining module 603 and a sending module 604. The first receiving module 601 is configured to receive a media file obtaining request sent by the terminal, where the media file obtaining request carries the media file identifier and the terminal user identifier. The querying module 602 is configured to query, according to the correspondence information between the media file identifier and the address of the server, the addresses of the M servers of the M substreams of the media file corresponding to the media file identifier in the storage media file obtaining request. The obtaining module 603 is configured to obtain N substreams of the media file from the M servers according to the addresses of the M servers queried by the query module 602, which are integers greater than 1 and N. Less than or equal to M. The sending module 604 is configured to send the N substreams of the media file corresponding to the media file identifier acquired by the obtaining module 603 to the terminal according to the terminal user identifier in the media file obtaining request. The media file is encoded into M sub-streams having the same occupied bandwidth according to the heat, and is respectively stored on M servers according to the server load balancing principle, where M is an integer greater than 1.

A server cluster provided by the embodiment of the present invention, the server cluster of the embodiment may be used to implement the foregoing media file service processing method embodiment of the present invention, which includes a media file service processing device 1 and L servers 2. The M servers of the L servers 2 are respectively configured to store one substream of the M sub-streams with the same occupied bandwidth, which is encoded by the same media file according to the server load balancing principle, where M is an integer greater than 1, and M is less than or equal to L. The media file service processing apparatus 1 is configured to receive a media file acquisition request sent by the terminal, where the media file acquisition request carries the media file identifier and the terminal user identifier, and stores the M servers in the server cluster according to the terminal user identifier. The sub-stream of the media file corresponding to the media file identifier is sent to the terminal. As shown in FIG. 7, it is a schematic structural diagram of an embodiment of a server cluster according to the present invention. The server cluster of the embodiment includes five servers, that is, the value of L is 5.

FIG. 8 is a schematic structural diagram of another embodiment of a server cluster according to the present invention. Compared with the server cluster of the embodiment shown in FIG. 7, the server 2 of the embodiment includes a storage module 701, and the media file service processing apparatus 1 includes The first receiving module 601, the querying module 602, the obtaining module 603, and the sending module 604. The storage module 701 is configured to store, among the other servers 2 in the server cluster, one substream of the M sub-streams with the same occupied bandwidth, which are encoded into the same media file according to the load balancing principle. The first receiving module 601 is configured to receive a media file obtaining request sent by the terminal, where the media file obtaining request carries the media file identifier and the terminal user identifier. The querying module 602 is configured to query, according to the correspondence information between the media file identifier and the address of the server, the addresses of the M servers of the M substreams of the media file corresponding to the media file identifier in the storage media file obtaining request. The obtaining module 603 is configured to obtain N sub-streams of the media file from the M servers according to the addresses of the M servers queried by the query module 602, where N is an integer greater than 1 and N is less than or equal to M. The sending module 604 is configured to obtain an end user identifier according to the media file acquisition request. The N substreams of the media files corresponding to the media file identifiers obtained by the obtaining module 603 from the storage modules 701 of the M servers 2 are sent to the terminal.

As shown in FIG. 9 , it is a schematic structural diagram of another embodiment of the server cluster of the present invention. The server cluster of the embodiment can be used to implement the process of the media file service processing method embodiment shown in FIG. 3 . Compared with the server cluster of the embodiment shown in FIG. 7, the media file service processing apparatus 1 in this embodiment includes a first receiving module 601, a querying module 602, and a forwarding module 801. The server 2 includes a second receiving module 802 and a storage module. 701 and sending module 604. The first receiving module 601 is configured to receive a media file obtaining request sent by the terminal, where the media file obtaining request carries a media file identifier and an end user identifier. The querying module 602 is configured to query, according to the correspondence information between the media file identifier and the address of the server, the addresses of the M servers of the M substreams of the media file corresponding to the media file identifier in the storage media file obtaining request. The forwarding module 801 is configured to forward the media file acquisition request to the second receiving module 802 of the M servers 2 according to the addresses of the M servers queried by the query module 602. The second receiving module 802 is configured to receive a media file acquisition request forwarded by the forwarding module 801. The storage module 701 is configured to store, in accordance with the load balancing principle, one sub-stream of the M sub-streams with the same occupied bandwidth, which are encoded by the same media file, with other servers in the server cluster. The sending module 604 stores, according to the terminal user identifier carried in the media file request received by the second receiving module 802, a substream of the media file corresponding to the media file identifier stored in the storage module 701 of the server 2 where the sending module 604 is located. Send to the terminal. Fig. 9 shows only the connection relationship between the media file service processing apparatus 1 and a server 2. For the case where the server cluster includes a plurality of servers, the connection relationship between the media file service processing apparatus 1 and the other servers 2 is the same.

As shown in FIG. 10, it is a schematic structural diagram of another embodiment of a server cluster according to the present invention. The server cluster of this embodiment can be used to implement the process of the media file service processing method embodiment shown in FIG. 5. Compared with the server cluster of the embodiment shown in FIG. 7, the media file service processing apparatus 1 in this embodiment includes a first receiving module 601, a querying module 602, and an indicating module 901. The server 2 includes a second receiving module 802 and a storage module. 701 and sending module 604. The first receiving module 601 The method is configured to receive a media file acquisition request sent by the terminal, where the media file acquisition request carries a media file identifier and an end user identifier. The querying module 602 is configured to query, according to the correspondence information between the media file identifier and the address of the server, the addresses of the M servers of the M substreams of the media file corresponding to the media file identifier in the storage media file obtaining request. The indication module 901 is configured to send the indication information to the second receiving module 802 of the M servers 2 according to the address of the M servers that are queried by the querying module 602, where the indication information includes the terminal user identifier and the media file identifier, and the indication M The sending module 604 of the server 2 sends the sub-stream of the media file corresponding to the media file identifier stored on the server 2 of the sending module 604 to the terminal according to the terminal user identifier. The second receiving module 802 is configured to receive a media file obtaining request forwarded by the forwarding module 801. The storage module 701 is configured to store, among the other servers 2 in the server cluster, one substream of the M sub-streams with the same occupied bandwidth, which are encoded into the same media file according to the load balancing principle. The sending module 604 sends the substream of the media file corresponding to the media file identifier stored in the storage module 701 of the server 2 where the sending module 604 is located, according to the terminal user identifier in the indication information received by the second receiving module 802. terminal. FIG. 10 shows only the connection relationship between the media file service processing apparatus 1 and one server 2. For the case where the server cluster includes a plurality of servers, the connection relationship between the media file service processing apparatus 1 and the other servers 2 is the same.

In the server cluster of the foregoing embodiment of the present invention, the L storage modules 701 of the L servers 2 are respectively used to store L media streams of the same heat level with the highest popularity. A sub-flow, L servers 2 are in accordance with the load balancing principle, wherein the P storage modules 701 are also respectively configured to store P media files of the same heat level other than the first heat level encoded into P sub-streams having the same occupied bandwidth A subflow. As a specific embodiment, P takes an integer value in L/Q or L/2Q- ¹ , and Q is a heat level.

A person skilled in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by using hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, and the program is executed when executed. The steps of the foregoing method embodiments are included; and the foregoing storage medium includes: ROM, RAM, disk or optical disk, etc. The media of the sequence code.

In the embodiment of the present invention, according to the heat of the media file, the media file is encoded into M sub-streams with the same occupied bandwidth according to the load balancing principle of the servers in the server cluster, and are respectively stored on the M servers in the server cluster. When the server cluster receives the media file acquisition request sent by the terminal, respectively, the M substreams of the media files stored on the M servers are sent to the terminal, and the terminal may be restored according to the received multiple substreams of the M substreams. The media files are output, thereby realizing load sharing and load balancing between servers in the server cluster, and considering the storage redundancy and system robustness of the server cluster, so that the service capabilities of the servers are effectively utilized.

It should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention and are not to be construed as limiting. Although the present invention has been described in detail with reference to the preferred embodiments thereof, it will be understood by those skilled in the art that the invention may be modified or equivalently substituted without departing from the invention. The spirit and scope of the programme.

Claims

Rights request

A method for processing a media file storage, comprising:

Determining the popularity level of the media file according to the popularity of the media file;

Encoding the media file into M sub-streams occupying the same bandwidth;

The method according to claim 1, wherein the correspondence between the heat level and the number of servers storing the same media file comprises:

The number of servers storing the hottest first heat level media files is the number of servers constituting the server cluster L, L is an integer greater than or equal to M; is an integer less than L.

3. Method according to claim 2, characterized in that P is taken from L/Q or L/2 ⁽ an integer value in ^, Q is a heat level).

The method according to any one of claims 1 to 3, wherein after the M sub-streams having the same occupied bandwidth are respectively stored on the M servers in the server cluster, the method further includes:

The addresses of the M servers storing the substreams of the media files are recorded.

A media file storage processing device, comprising:

a determining module, configured to determine, according to a correspondence between a preset heat level and a number of servers storing the same media file, a number M of servers corresponding to a heat level of the media file; and a coding unit, configured to encode the media file M sub-streams occupying the same bandwidth; The storage allocation module is configured to send the M sub-streams with the same occupied bandwidth to the M server storages in the server cluster according to the server load balancing principle.

The device according to claim 5, wherein the number of servers of the media files of the first heat level with the highest heat is the number L of servers constituting the server cluster, L is an integer greater than or equal to M; An integer less than L.

7. The device according to claim 5 or 6, further comprising:

And a recording module, configured to record an address of the M servers storing the substreams of the media file.

8. A server cluster, comprising:

L servers, the M servers of the L servers are respectively configured to store one of the M sub-streams with the same occupied bandwidth encoded by the media file according to the server load balancing principle, and receive the media sent by the terminal. a file obtaining request, and sending, to the terminal, a substream of the media file corresponding to the file identifier in the media file obtaining request, where M is an integer greater than 1, and M is less than or equal to L.

A server cluster, comprising: L servers, wherein one of the L servers is configured to determine a heat level corresponding to the heat of the media file according to the correspondence information between the preset heat and heat levels. Determining, according to a correspondence between a preset heat level and a number of servers storing the same media file, a number M of servers corresponding to a heat level of the media file, and encoding the media file into M sub-bands having the same occupied bandwidth Stream, and press According to the principle of server load balancing, M sub-streams occupying the same bandwidth are respectively stored on M servers in the L servers, M is an integer greater than 1, and M is less than or equal to L;

The M servers of the L servers are respectively configured to store one of the M substreams with the same occupied bandwidth, and receive the media file acquisition request sent by the terminal according to the server load balancing principle. And sending, to the terminal, the substream of the media file corresponding to the file identifier in the media file obtaining request that is stored.

A method for processing a media file service, comprising:

The method according to claim 10, further comprising: after receiving the media file acquisition request sent by the terminal,

Querying an address of the M servers storing the media file;

And forwarding, according to the addresses of the M servers, the media file acquisition request to the M servers respectively; or

And sending, to the terminal, the sub-streams of the media files that are respectively stored, respectively: querying an address of the M servers that store the media files;

Obtaining, according to the addresses of the M servers, N substreams of the media file from the M servers, being an integer greater than 1 and N being less than or equal to M;

Sending N substreams of the media file to the terminal according to the end user identifier; or

And sending, to the terminal, the sub-streams of the media files that are stored separately, respectively: querying an address of the M servers that store the media files; And sending a substream of the media file respectively stored by the M servers to the terminal.

The method according to claim 10 or 11, wherein the number of servers storing the media files of the first heat level with the highest heat is the number L of servers constituting the server cluster, L is greater than or equal to M Integer; is an integer less than L.

The method according to claim 10 or 11, further comprising: receiving, by the terminal, a plurality of substreams of the media file returned by the server cluster, and decoding and combining the plurality of substreams, Restore the media file.

14. A media file service processing device, comprising:

a sending module, configured to send, to the terminal, the N sub-streams of the media files that are respectively obtained from the M servers and corresponding to the media file identifier, according to the terminal user identifier, where The media file is encoded into M sub-streams having the same bandwidth according to the heat, and is respectively stored on the M servers according to the server load balancing principle, and is an integer greater than 1.

15. A server cluster, comprising:

a media file processing apparatus, configured to receive a media file acquisition request sent by the terminal, the media text The device obtains the media file identifier and the terminal user identifier, and the sub-streams of the media files corresponding to the media file identifiers stored on the M servers in the server cluster are sent to the sub-users. The terminal.

The server cluster according to claim 15, wherein the server comprises: a storage module, configured to store, according to a load balancing principle, an M that is encoded by the same media file with other servers in the server cluster. One of the substreams occupying the same bandwidth; accordingly,

The media file processing apparatus includes:

a query module, configured to query an address of the M servers storing the M substreams of the media file;

An obtaining module, configured to acquire, according to the addresses of the M servers, N sub-streams of the media file from the M servers, respectively, being an integer greater than 1 and N is less than or equal to M;

a sending module, configured to send, according to the terminal user identifier, N substreams of the media file corresponding to the media file identifier stored in the storage modules of the M servers to the terminal; or

The media file processing apparatus includes:

a forwarding module, configured to forward the media file acquisition request to the M servers according to the addresses of the M servers;

The server includes:

a second receiving module, configured to receive a media file acquisition request forwarded by the forwarding module; a storage module, configured to store, in accordance with a load balancing principle, one sub-stream of the M sub-streams with the same occupied bandwidth encoded by the same media file, and the other servers in the server cluster;

a sending module, configured to send, according to the terminal user identifier in the media file obtaining request, a substream of the media file corresponding to the media file identifier stored in the storage module, to the terminal; Or

The media file processing apparatus includes:

The indication module is configured to: according to the address of the M servers, respectively, the sending module of the M servers, according to the terminal user identifier, the media file identifier corresponding to the media file identifier stored on the server where the sending module is located a substream of the media file is sent to the terminal; correspondingly, the server includes:

a second receiving module, configured to receive indication information sent by the indication module;

a storage module, configured to store, in accordance with a load balancing principle, one sub-stream of the M sub-streams with the same occupied bandwidth encoded by the same media file;

And a sending module, configured to send, according to the terminal user identifier in the media file obtaining request, a substream of the media file corresponding to the media file identifier stored in the storage module, to the terminal.

The server cluster according to claim 16, wherein the L storage modules of the L servers are respectively used to store the media files of the first heat level with the highest heat, and the L occupied bandwidths are the same. a sub-stream of the sub-flows, the L servers are in accordance with the load balancing principle, wherein the P storage modules are also respectively used to store the media data of the remaining heat levels other than the first heat level, and the P occupied bandwidths are the same. A subflow in the subflow.