JP2001045023A - Video server system and video data distribution method - Google Patents

Abstract

(57) [Summary] [Problem] To provide a video server system in which no sound interruption occurs on a terminal side even when a failure occurs in a server device during distribution processing. SOLUTION: A server status monitoring means 33a uses a backup server 40 as information on video data distribution processing as time slot management information for each time slot cycle.
Send to If the server status monitoring means 43a cannot receive the time slot management information, the main server 3
0 is considered to have failed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to video data (moving image data, audio data, etc.) stored in a server device.
Video server system that distributes
The present invention relates to a video server system employing a cluster.

[0002]

2. Description of the Related Art A cluster is an Ethernet or FD.
A loosely coupled system in which a plurality of server devices are connected by a DI (Fiber Distributed Data Interface) or the like.

FIG. 9 is a schematic functional block diagram of a video server system employing the above-described cluster (hereinafter, referred to as a “cluster-structured video server system”). The configuration will be described below together with a video data distribution processing procedure.

[0004] First, when the terminal 10 issues a distribution request by designating video data, the transmission / reception means 31 of the server device 30 that has received the distribution request via the network device 20 transmits a cluster video file to extract the video data. Instruct system 32. Upon receiving this instruction, the cluster video file system 32 extracts the video data from the storage device 50 such as a hard disk, and transfers the video data to the transmission / reception unit 31. The video data passed in this way is returned to the terminal 10 by the transmission / reception means 31, and is reproduced by reproduction means (not shown) on the terminal 10.

Here, the server device 30 in the operating state
The server status monitoring means 33 (hereinafter referred to as “main server 30”) monitors the status of its own device, and when detecting a sign of a failure, notifies the server status monitoring means 43 of the server device 40 of the fact. After that, the own device is switched from the operating state to the standby state.

On the other hand, when the server status monitoring means 43 of the server device 40 in the standby state (hereinafter referred to as “backup server 40”) receives the notification that the main server 30 has detected a sign of a failure as described above. Then, the own device is switched from the standby state to the operating state.

[0007] Thereafter, the distribution request issued by the terminal 10 is accepted by the transmission / reception means 41 of the backup server 40 and processed by the backup server 40 in the same procedure as in the case of the main server 30. The method of taking over the processing in this way is called failover and has been widely used conventionally.

As described above, if a cluster is adopted, even if a failure occurs in the main server 30, the backup server 40 takes over and executes the distribution processing, so that a video server system with high scalability and reliability can be provided.

[0009] Here, two server devices 30.
Although the cluster is configured by 40, the number of server devices that configure the cluster can be three or more (when the total number of server devices that configure the cluster is N and the standby number is M). , N + M reliability).

[0010]

Here, in the above-mentioned conventional video server system, when a sign of a failure is detected in the main server 30 which is not performing the distribution processing, the distribution processing is taken over by the backup server 40 without any trouble. . However, when a failure occurs in the main server 30 during the distribution process (a failure may occur during the distribution process even if no sign of the failure is detected), the terminal 10 may play back the video data. There was a problem that sound interruption occurred.

That is, according to the conventional video server system, when a failure occurs in the main server 30 during the distribution process, the terminal 10 that has detected that the video data is not distributed from the main server 30 issues a retry request. , The backup server 40 receiving this retry request
Is a procedure that takes over the distribution process, so that a sound break occurs on the terminal 10 side.

The present invention has been proposed on the basis of the above-mentioned conventional circumstances, and provides a video server system in which no sound interruption occurs at the terminal side even when a failure occurs in a server device during distribution processing. It is intended to do so.

[0013]

The present invention employs the following means to achieve the above object. That is, FIG.
As shown in (1), when a failure occurs in the main server 30, a video server system having a cluster configuration in which the backup server 40 takes over the subsequent distribution processing is assumed.

The server status monitoring means 3 of the main server 30
3a transmits to the backup server 40 information relating to video data distribution processing as time slot management information for each time slot cycle. Further, the server status monitoring means 43a of the backup server 40
If the time slot management information cannot be received from 0, it is considered that a failure has occurred in the main server 30. If it is determined that a failure has occurred, the transmission / reception unit 41 of the backup server 40
a starts taking over the distribution process based on the last received time slot management information.

In this way, even if a failure occurs in the main server 30 during the distribution process, a failover can be performed two time slots after the failure occurs, so that two time slots If the terminal 10 pre-reads the video data distributed during the period, it is possible to guarantee that no sound interruption will occur.

As shown in FIG. 2, a proxy device having a cluster configuration including a main proxy 60 and a backup proxy 70 is connected to the terminal 10 and the server device 30.
40 may be provided. The data acquisition unit 62 of the main proxy 60 acquires the video data from the backup proxy 70 when the video data requested to be distributed from the terminal 10 is not stored in the data storage unit 64 of the own device, When the obtained video data is not stored in any of the data storage means 64 and the backup proxy 70 of the own device, the video data is obtained from the main server 30.

As described above, according to the configuration in which the proxy devices 60 and 70 are added, the video data can be distributed from the data storage means 64 and 74 which can access at high speed, so that the response time to the terminal 10 can be shortened. In addition, the time slot cycle varies depending on the access performance of the device in which the video data is stored. According to the above configuration, the time slot cycle is shortened, so that the video data is reproduced without interruption on the terminal 10 side. The buffer capacity required for the operation can be reduced.

The storage device selection means 34 of the main server 30 shown in FIG. 3 determines whether at least one of the RAID device 36 of its own device and the RAID device 46 of the backup server 40 is to be taken out of the video data requested to be distributed from the terminal 10. Or choose. In addition, R of the main server 30
The AID management unit 35 combines the video data received from the storage device selection unit 34 to generate video data requested to be distributed from the terminal 10.

According to such a procedure, video data can be recovered even if a part of the hard disk fails.

Further, each unit shown in FIG. 4 may be added between the cluster video file system 32 and the storage device selecting unit 34 shown in FIG. For example, RAM
The temporary data storage unit 37e temporarily stores data extracted in an optimum size from at least one of the RAID device 36 of its own device and the RAID device 46 of the backup server 40. And the extraction destination determining means 3
7a determines one of the data temporary storage means 37e and the storage device selection means 34 as a fetch destination of the video data requested to be distributed from the terminal 10.

According to such a procedure, the response time to the terminal 10 can be shortened because data can be efficiently extracted.

[0022]

Embodiments of the present invention will be described below in detail with reference to the drawings. (First Embodiment) FIG. 1 is a schematic functional block diagram of a video server system to which the present invention is applied. Hereinafter, only the configuration of the video server system which is different from the above-described conventional video server system will be described.

First, the transmission / reception means 31 returns the video data requested for distribution to the terminal 10 while the main server 30
The server status monitoring means 33a transmits information relating to distribution processing (distribution file name, distribution video data offset, transfer speed, distribution destination address, etc.) to the backup server 40 as time slot management information for each time slot cycle described later. (FIG. 5, steps S1 → S2 → S3
→ S4). The distribution file name is the file name of the video data requested to be distributed, and the distribution destination address is the network address of the terminal 10 that has issued the distribution request.

Here, the server status monitoring means 43a of the backup server 40 determines the failure status of the main server 30 based on whether or not the time slot management information is received every time slot period. In other words, when the time slot management information cannot be received in each time slot cycle, the server status monitoring unit 43a regards that the main server 30 has failed and determines that the main server 30 has failed (immediately before failure occurs) in the next time slot cycle. (Step S14 → S15 in FIG. 5) based on the received time slot management information.

As described above, even if a failure occurs in the main server 30 during the distribution process, a failover is possible two time slots after the failure occurs. If the terminal 10 pre-reads the video data distributed during the period, it is possible to guarantee that no sound interruption occurs.

The technique of prefetching video data is as follows.
Since the technique is widely known from the related art, the description is omitted here. (Second Embodiment) The time slot is a time unit when the video data requested to be distributed is read out from the device in which the video data is stored. Depends on the access performance of the device in which is stored (the higher the access performance, the smaller the value). That is, the time slot cycle when video data is distributed from the hard disk (storage device 50) as in the first embodiment is 13
The time slot period is about 0 ms, but the time slot cycle can be reduced to about 50 ms if a configuration is adopted in which video data is distributed from a RAM or the like that can be accessed at high speed.

Therefore, in this embodiment, the proxy devices 60 and 70 having the cluster configuration shown in FIG. 2 are installed between the terminal 10 and the server devices 30 and 40, and the proxy devices 60 and 70 are provided with, for example, a RAM or the like. Data storage means 6
The video data is distributed from 4.74,
Hereinafter, only the configuration different from the first embodiment will be described.

First, when the terminal 10 issues a distribution request by designating video data, the transmission / reception means 61 of the main proxy 60 that has accepted the distribution request instructs the data acquisition means 62 to retrieve the video data. Upon receiving this instruction, the data acquisition unit 62 extracts the video data from the data storage unit 64 and passes it to the transmission / reception unit 61. The video data passed in this manner is transmitted to
1 and is reproduced by the reproducing means on the terminal 10 (FIG. 6, step S21 → S2).
2).

Here, while the transmission / reception means 61 returns the video data requested for distribution to the terminal 10, the proxy status monitoring means 63 of the main proxy 60 transmits the time slot management information described in the first embodiment. Send to backup proxy 70 and backup proxy 7
When the time slot management information cannot be received, the proxy status monitoring unit 73 of 0 starts taking over the distribution process. Such a takeover is similar to the takeover performed by the server status monitoring means 33a and 43a, and
Since this is as described in the embodiment, detailed description thereof will be omitted here.

As described above, according to the configuration in which the proxy devices 60 and 70 are added, the video data can be distributed from the data storage means 64 and 74 that can access at high speed, so that the response time to the terminal 10 can be shortened. As the time slot period is shortened, the capacity of the buffer necessary for reproducing video data on the terminal 10 side without interruption can be reduced.

By the way, it is usual that only specific video data (for example, only video data with a high distribution frequency) is stored in the proxy device.

Therefore, there are cases where the video data requested for distribution is not stored in the data storage means 64, and in this case, the data acquisition means 62 of the main proxy 60.
Issues a request to acquire the video data to the backup proxy 70 (FIG. 6, step S21 → S2).
3). Then, the data acquisition unit 72 of the backup proxy 70 receiving the acquisition request retrieves the video data from the data storage unit 74 and retrieves the video data.
(FIG. 6, steps S24 → S25 → S22).

However, if the video data is not stored in the data storage unit 74 of the backup proxy 70 (described later), the data acquisition unit 62 of the main proxy 60 does not send the video data to the backup proxy 70 but to the main server 30. Issue the video data acquisition request (step S24 → S26 in FIG. 6). Then, the transmission / reception means 31 of the main server 30 receiving the acquisition request retrieves the video data from the storage device 50 via the cluster video file system 32 and returns it to the main proxy 60 (FIG. 6, step S27 → S2).
2).

Here, the proxy status monitoring means 63 of the main proxy 60 checks the status of the data storage means 64 (name of stored files, attribute information, distribution frequency information, etc.).
Is periodically transmitted to the backup proxy 70, and the proxy status monitoring means 73 of the backup proxy 70 periodically transmits the status (same as above) of the data storage means 74 to the main proxy 60.
In this way, the data acquisition unit 62 of the main proxy 60 can determine whether the video data is stored in the data storage unit 74 of the backup proxy 70 as described above.

Here, the data storage means 64 of the main proxy 60 and the data storage means 74 of the backup proxy 70 store video data individually, but the present invention is not limited to this. Absent. That is, when one of the data storage units is updated, the updated content may be reflected on the other data storage unit. In this case, since both data storage means 64 and 74 store the same video data, the state monitoring procedure described above can be omitted.

In this case, two proxy devices 60
A cluster is configured at 70, but the number of proxy devices that configure the cluster is three or more as in the case of the server device (a cluster configuration having N + M reliability).
You can also. (Third Embodiment) In the storage device 50 in the first and second embodiments, two or more hard disks are connected and used as one logical disk (RA).
ID) is preferably employed. If RAID is adopted, video data can be stored redundantly, so that video data can be restored even if a part of the hard disk fails.

Therefore, in the server devices 30 and 40 of the present embodiment, the RAID device shown in FIG.
D) (storage devices employing D) 36 and 46 store video data. The configuration of the video data will be described below, focusing only on the differences from the first embodiment.

First, the cluster video file system 32 instructed by the transmission / reception means 31 to retrieve the video data requested to be distributed, instructs the storage device selection means 34 to retrieve the video data, and the storage device receiving the instruction. The selecting means 34 converts the video data into a RAID
It is extracted from the device 36 and returned to the cluster video file system 32.

However, when a part of the RAID device 36 is out of order, the video data generated in the following procedure is returned to the cluster video file system 32. That is, the storage device selection unit 34 that has detected that a part of the RAID device 36 has failed extracts video data from the non-failed hard disk and passes it to the RAID management unit 35, while The above video data is taken out from the RAID device 46 of the backup server 40 and passed to the RAID management means 35. Then, the RAID management unit 35 combines the two pieces of video data passed as described above and returns them to the cluster video file system 32.

For example, four hard disks (HDD1)
RAI composed of HDD2, HDD3, HDD4)
When the HDD 3 of the D device 36 is out of order, the storage device selecting unit 34 extracts the data 1, 2, and 4 on the HDD 1, HDD 2, and HDD 4 and passes it to the RAID management unit 35, while the RAID device of the backup server 40 Data 3 (insufficient data) is extracted from 46 and passed to the RAID management means 35. Then, the RAID management unit 35 generates video data requested to be distributed by combining the data 1, 2, 3, and 4 passed as described above, and returns the video data to the cluster video file system 32.

According to the above procedure, even if a part of the hard disk fails, the video data can be recovered, and only the video data on the failed hard disk is taken out from the backup server 40. Therefore, it is possible to minimize the amount of data transfer when a failure occurs.

In the above description, "the cluster video file system 32 simply instructs the storage device selecting means 34 to extract video data". However, such an instruction is not required for data of a fixed size (for example, 8 KB). This is realized by issuing the retrieval request of the document a plurality of times. However, it is preferable that the data retrieved from the hard disk has an optimum size (for example, 64 KB) which is the most efficient retrieved data size. Therefore, in the present embodiment, each unit shown in FIG. 4 is added between the cluster video file system 32 and the storage device selecting unit 34, so that data can be extracted at the optimum size.

That is, upon receiving the 8 KB data retrieval request from the cluster video file system 32, the retrieval destination determining means 37a sets the retrieval magnification "8" described later together with the retrieval request to the data retrieval means 37d.
(FIG. 7, steps S31 → S32 → S3
3). Upon receiving the notification, the data extracting unit 37d
Data of a size obtained by multiplying the size of the requested data “8 KB” by the take-out magnification “8” (ie, 64K
B) via the storage device selecting means 34
After being taken out from the device 36, the first 8 KB of data is returned to the cluster video file system 32, and the remaining data (that is, 56KB of data) is stored
Temporarily stored in the data temporary storage means 37e (FIG. 7,
Steps S33 → S34 → S35).

In this way, the data requested to be fetched thereafter (second to eighth times) can be fetched from the data temporary storage means 37e capable of high-speed access (FIG. 7, steps S31 → S32 → S36). → S37 → S
38 → S31) As a result, the response time to the terminal 10 can be reduced.

The take-out magnification is determined by the take-out magnification storage means 37b based on the size of the video data requested to be taken out from the cluster video file system 32 and the optimum size previously stored in the optimum size storage means 37c. decide. For example, the size of the video data requested to be extracted is 8 KB, and the optimal size is 64 KB.
In the case of KB, the takeout magnification storage unit 37b stores “6
4 (KB) / 8 (KB) ", and the result of the operation," 8 ", is taken out and stored as the magnification.

In the above description, the main server 30
Although the procedure in which the backup server 40 and the backup proxy 70 take over the distribution process when a failure occurs in the main proxy 60 has been described, the distribution process is not limited to the case where the failure occurs. That is, the server status monitoring unit 33a and the proxy status monitoring unit 63 monitor the load of the own device,
When this load exceeds a predetermined value, the distribution process is taken over by the backup server 40 and the backup proxy 7.
0 may be requested.

Further, in the above description, the installation environment of the cluster-structured proxy device and the cluster-structured server device is not particularly mentioned. For example, as shown in FIG. The same effect as described above can be obtained even if the devices are dispersed and installed.

[0048]

As described above, according to the present invention, even if a failure occurs in the main server during the distribution process, the failure occurs two time slots after the failure occurs.
Since the overrun is possible, if the terminal pre-reads the video data distributed between two time slot periods, it is possible to guarantee that the sound will not be interrupted.

According to the configuration in which the proxy device having the cluster configuration is added, the video data can be distributed from the data storage means capable of high-speed access, so that the video server can be distributed to more terminals.

Further, according to the above configuration, the response time to the terminal can be shortened, and the capacity of the buffer necessary for reproducing the video data on the terminal side without interruption can be reduced.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a video server system according to a first embodiment.

FIG. 2 is a functional block diagram of a video server system according to a second embodiment.

FIG. 3 is a functional block diagram of a video server system according to a third embodiment.

FIG. 4 is a functional block diagram of a main part of a video server system according to a third embodiment.

FIG. 5 is a sequence diagram showing takeover of distribution processing in the present invention.

FIG. 6 is a flowchart showing a data acquisition procedure in the present invention.

FIG. 7 is a flowchart showing a data fetching procedure in the present invention.

FIG. 8 is another configuration example of a video server system to which the present invention is applied.

FIG. 9 is a functional block diagram of a conventional video server system.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Terminal 20 Network equipment 30 Main server 31 Main server transmission / reception means 33a Main server server status monitoring means 34 Main server storage device selection means 35 Main server RAID management means 36 Main server RAID device 37e Main server data temporary storage Means 37a Main server extraction destination determination means 40 Backup server 50 Storage device 60 Main proxy 61 Main proxy transmission / reception means 62 Main proxy data acquisition means 63 Main proxy proxy status monitoring means 64 Main proxy data storage means 70 Backup proxy

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B082 DC04 DC05 DE02 5C064 BA01 BA07 BB06 BC16 BC18 BC27 BD13 5K030 HB02 JA10 KA01 KA02 MD02 9A001 BB04 CC07 JJ07 JJ27 LL02 LL09

Claims (12)

[Claims] 1. A video server system having a cluster configuration including a main server and a backup server, and in a case where a failure occurs in the main server, the backup server takes over the subsequent distribution processing. Server status monitoring means for transmitting information relating to video data distribution processing to the backup server as time slot management information for each time slot cycle. If the backup server cannot receive the time slot management information from the main server, Server status monitoring means for determining that the main server has failed;
A video server system comprising: a transmission / reception unit that starts taking over the distribution process based on the time slot management information received last by the server status monitoring unit. 2. The video server system according to claim 1, wherein said terminal prefetches video data distributed during two time slot periods. 3. A proxy device having a cluster configuration including a main proxy and a backup proxy is provided between the terminal and the server device, and the main proxy transmits information on video data distribution processing at each time slot period. Proxy status monitoring means that transmits to the backup proxy as time slot management information, and if the backup proxy cannot receive the time slot management information from the main proxy, the proxy status monitoring that the main proxy is considered to have failed 3. The video server system according to claim 2, further comprising: means for transmitting and receiving the distribution processing based on the time slot management information received last by the proxy status monitoring means. 4. When the video data requested to be delivered from the terminal is not stored in the data storage means of the own device, the main proxy acquires the video data from the backup proxy, and outputs the video data requested to be delivered from the terminal. 4. A data acquisition means for acquiring the video data from the main server when the video data is not stored in any of the data storage means and the backup proxy of the own apparatus.
A video server system according to item 1. 5. The storage device selecting means for the main server to select at least one of a storage device of its own device and a storage device of a backup server as a fetch destination of video data requested to be distributed from a terminal, and the storage device 4. The video server system according to claim 1, further comprising: a RAID management unit that combines the video data received from the device selection unit and generates the video data requested to be distributed. 6. The data temporary storage means, wherein the main server temporarily stores data extracted at an optimum size from at least one of the storage device of its own device and the storage device of the backup server; 6. The video server system according to claim 5, further comprising: an extraction destination determination unit that determines any one of the storage device selection units as an extraction destination of the video data requested to be distributed by the terminal. 7. The video server system according to claim 1, wherein the server device has a cluster configuration having N + M reliability. 8. The video server system according to claim 3, wherein the proxy device has a cluster configuration having N + M reliability. 9. A cluster comprising a plurality of server devices,
If a failure occurs in the main server, the video data distribution request issued by the terminal after that will be processed in the video data distribution method processed by the backup server, and the information on the video data distribution processing will be used as time slot management information for each time slot cycle. The video is transmitted from the main server to the backup server, and if the transmission is not performed, the backup server starts taking over the distribution processing based on the last transmitted time slot management information. Data delivery method. 10. A proxy device having a cluster configuration including a main proxy and a backup proxy is installed between the terminal and the server device, and when a failure occurs in the main proxy, distribution of video data issued by the terminal thereafter. The video data distribution method according to claim 9, wherein the request is processed by a backup proxy. 11. A method in which information relating to video data distribution processing is transmitted from the main proxy to the backup proxy as time slot management information for each time slot period. If the transmission is not performed, the last transmitted time is transmitted. 2. The backup proxy starts takeover of the distribution process based on slot management information.
0. A video data distribution method according to item 0. 12. When the video data requested to be distributed by the terminal is not stored in the main proxy, the video data stored in the backup proxy is distributed to the terminal, and the video data requested to be distributed by the terminal is transmitted to the main proxy. 11. The video data distribution method according to claim 10, wherein the video data stored in the main server is distributed to the terminal when the video data is not stored in any of the backup proxies.