JP2003167813A - Stream data storing and distributing method and system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide the stream data storing distributing method and system capable of shortening a waiting time for a user while inhibiting the lowering of the storage capacity required to a server. <P>SOLUTION: When a part of the stream data S is stored in a downstream server 20 of a second stage or lower, a process [1A] for dividing the i stream data into a leading part data and the remaining part data by a server of the i stage, a process [1B] for transmitting the leading part data of the i stream data to a server of the (i+1) stage by the server of the i stage, and a process [1C] for holding the received leading part data by the server of the (i+1) stage as the (i+1) stream data, are performed. In the process [1A], P<SB>i</SB>H≥T<SB>i</SB>T is satisfied when a replay time by a client, of the leading part data of the i stream data is P<SB>i</SB>H, and a time necessary for transmitting the remaining part data of the i stream data from the server of the i stage to the server of the (i+1) stage is T<SB>i</SB>T. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention distributes time series data (hereinafter referred to as "stream data") such as moving image data and audio data to a plurality of servers (also referred to as "storage" or "cache"). The present invention relates to a stream data storage / delivery method and a stream data storage / delivery system that are delivered to clients via a network.

[0002]

2. Description of the Related Art Recently, a content distribution service in which moving image data and audio data compressed by a data compression technique such as MPEG are stored in a storage device (storage device) of a server and delivered to a client (user) via a network. Are available. In addition, the networks used by users have become more broadband, and a constant connection environment has been established. Therefore, it is expected that network traffic will continue to increase due to the distribution of large-capacity contents such as movies and live images.

Although the backbone of the network is being enhanced with the increase in traffic volume, the transmission capacity of the network is limited. Therefore, in order to maintain the transfer quality satisfying the user's request, it is necessary to distribute the content data to a plurality of servers on the network to reduce the traffic amount in the backbone network. Such a technique has already been adopted in a content distribution service using the Internet, and is disclosed in, for example, US Pat. No. 6,108,703.

[0004]

However, by accumulating (caching) the stream data in a large number of cache servers (edges of the network) close to the user, the response to the user is speeded up and the amount of backbone traffic is reduced. In order to reduce it, a large-capacity storage device must be installed in each cache server, and the price of each cache server (as a result, the facility cost of the entire storage / delivery system) becomes very high. On the contrary, if the storage capacity of the storage device of each cache server is reduced, the price of each cache server (as a result, the facility cost of the entire storage / delivery system) can be reduced. The hit rate (hereinafter also referred to as “cache hit rate”) decreases, and the frequency of receiving stream data from an upstream server increases, which increases the backbone traffic volume and delays the response to the user. In summary, if the storage capacity of the cache server is increased in order to improve the hit rate and the waiting time, the equipment cost increases, and conversely, if the storage capacity of the cache server is decreased to reduce the equipment cost. There is a problem that the amount of traffic increases and the waiting time of the user becomes long.

Therefore, the present invention has been made in order to solve the above-mentioned problems of the prior art, and an object of the present invention is to reduce the waiting time of the user while keeping the storage capacity of the cache server low. An object of the present invention is to provide a stream data storage / delivery method and a stream data storage / delivery system.

[0006]

A method for accumulating / delivering stream data according to the present invention is applied to servers from the first stage to the m-th stage (m is an integer of 2 or more) arranged hierarchically on a network. A method of accumulating stream data and delivering it to a client via a network, comprising a step of accumulating a part of the first stream data held by the first stage server in a downstream server of the second stage and below, The accumulation step is performed for each of i = 1, 2, ..., M−1. [1A] When the (i + 1) th stage server exists downstream of the i-th stage server, the i-th stage server A process of dividing the i-th stream data held by itself into the head partial data and the remaining partial data, [1B] a process in which the i-th stage server transmits the head partial data of the i-th stream data to the (i + 1) -th stage server And, [1C] (i + 1) th
The stage server includes a process of holding the head part data of the received i-th stream data as the (i + 1) th stream data, and in the process [1A], the reproduction time by the client of the head part data of the i-th stream data is set to P _{If iH} and the time required to transfer the remaining part of the i-th stream data from the i-th stage server to the (i + 1) _-th stage server is T _iT , then P _iH ≧ T
It is characterized by satisfying _iT .

Further, when the client makes a reproduction distribution request for reproducing the first stream data to the j-th (j is an integer of 2 or more and m or less) stage server, the requesting client makes the first request.
[2A] j-th stage, in which the j-th stream data [2A] the server itself distributes to the requesting client, A process in which the server requests and receives the remaining part data of each of the stream data from the (j-1) th to the first (j-1) th upstream servers, and [2C] the process [2A ]Followed by,
The j-th stage server receives the remaining part of each of the received (j-1) to 1st stream data as the first
It is possible to include a process of delivering the stream data to the request source client in order according to the time series.

Further, according to another aspect of the present invention, there is provided a method for storing / delivering stream data, wherein the stream data is stored in a server arranged hierarchically on the network from the first stage to the m-th stage (m is an integer of 2 or more). A method of accumulating and delivering to a client via a network, comprising: an accumulating step of accumulating a part of the first stream data held by the first stage server in a downstream server of the second stage and below. But i
= 1, 2, ..., M−1,
[3A] A process of dividing the i-th stream data into a plurality of fragment file data when a (i + 1) th stage server exists downstream of the i-th stage server, [3A]
B] A process of dividing each of the plurality of fragment file data divided in the process [3A] into a head part data and a remaining part data, [3C] the process [3B]
For transmitting to the (i + 1) th stage server the leading part data of a plurality of fragmented file data divided in [1], and [3D] the (i + 1) th stage server received the leading part of the plurality of fragmented file data A process of accumulating data as a plurality of (i + 1) th stream data is performed by the client of the leading part data of the plurality of fragment file data obtained by dividing the i-th stream data in the process [3B]. Playback time is P
_iH1 , P _iH2 , ..., P _iHn, and it is necessary to transfer the remaining partial data of the plurality of fragment file data obtained by dividing the i-th stream data from the i-th stage server to the (i + 1) _-th stage server. The time is _{Ti T1} ,
T _iT2, ..., in the case of the _{T iTn, P iH1} ≧ _T
iT1 , P _iH2 ≧ T _iT2 , ..., P _iHn ≧ T _iTn
It is characterized by satisfying.

When the client makes a normal reproduction distribution request for the first stream data to the j-th (j is an integer of 2 or more and m or less) stage server, the first stream data is normally distributed to the requesting client. Has a regeneration process,
In the normal reproduction step, [4A] The j-th stage server requests the leading part data of the temporally earliest fragment file data among the leading part data of the plurality of fragment file data held by the request source client. Process to deliver to [4
B] A process in which the j-th stage server requests and receives the remaining part data of the plurality of fragment file data from the upstream servers in the (j-1) -th stage and above, and [4C] the process [4]
[A], the j-th stage server sets the remaining part data of the plurality of fragment file data, and the first part data of the plurality of fragment file data which the j-th stage server holds by itself. It may include a process of delivering the stream data to the requesting client in order according to the time series.

Further, when the client makes a jump reproduction distribution request which is a reproduction start request from the specified position of the first stream data to the j-th (j is an integer of 2 or more and m or less) stage server, the request source client Has a first trick play step of delivering the first stream data to the first trick play step, and the first trick play step [5A] is the beginning of a plurality of fragment file data held by the j-th stage server. A process of selecting the fragment file data that is temporally closest to the specified position among the partial data and delivering the head partial data of the selected fragment file data to the requesting client, [5B] the j-th stage server, The remaining part data of the selected fragment file data and the fragment which is later in time than the selected fragment file data in the upstream server of the (j-1) th stage or higher. Requesting receiving processing remainder data Airudeta, and, [5C] The process [5
[A], the j-th stage server receives the remaining partial data of the received fragment file data and the first partial data of the plurality of fragment file data held by the j-th server itself in the first stream. It may include a process of delivering the data to the requesting client in order according to the time series of the data.

Furthermore, when there is a fast-forward reproduction distribution request from the client to the j-th (j is an integer of 2 or more and m or less) stage server, the second special reproduction for distributing the first stream data to the requesting client. And a second step
The trick play process may include a process in which the j-th stage server sequentially distributes the head part data of the plurality of fragment file data held by itself to the requesting client in order from the temporally earlier one.

Further, when a jth (j is an integer of 2 or more and m or less) stage request from the client to the fast-reverse reproduction distribution request, the third special reproduction for distributing the first stream data to the requesting client. In the third trick play step, the j-th stage server sequentially reverses the temporal flow of the first partial data of the plurality of fragment file data held by itself from the later one. It may include a process of delivering to the requesting client.

[8A] When the server delivers the stream data to the requesting client in response to the delivery request from the client, the delivering server sequentially notifies the upstream server of the usage frequency of the stream data. Then, [8B] this upstream server updates the usage frequency of the corresponding stream data that it has based on the notified usage frequency, and [8C] the updated usage frequency is notified to the upstream server, and [8D]. The processing [8
[B] and [8C] are repeated until there is no upstream server, and [8E] When the usage frequency of any of the servers reaches the threshold value of its own, the upstream of the corresponding stream data that has reached the threshold value is reached. It is also possible to store data in the downstream server that is in the server but not in the downstream server.

Further, in parallel with the process [8E], the corresponding stream data can be accumulated in other servers than the one downstream.

[10A] When the server delivers the stream data to the requesting client in response to the delivery request from the client, the delivering server sequentially notifies the upstream server of the usage frequency of the stream data. [10B] The upstream server updates the usage frequency of the stream data of its own based on the notified usage frequency, and [10C] the updated usage frequency is further notified to the upstream server, and [10D] Processing [1
0B] and [10C] are repeated until there are no upstream servers, and [10E] each server has a plurality of thresholds and a plurality of fragment file data information corresponding to each of the plurality of thresholds. Amount of fragment file data that exists in the upstream server but does not exist in the downstream server is determined by the reached threshold value each time the usage frequency of It can also be stored in the downstream server only.

Further, in parallel with the process [10E], the corresponding stream data can be stored in a server other than the one downstream server.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION <1> First Embodiment <1-1> Configuration of First Embodiment FIG. 1 shows a stream data storage / delivery system (first embodiment) according to a first embodiment of the present invention. FIG. 1 is a configuration diagram schematically showing a system for implementing a method of storing / delivering stream data according to an embodiment of the present invention.

As shown in FIG. 1, the stream data storage / delivery system according to the first embodiment includes all content data for delivery (including stream data such as moving image data and audio data). Root server (Root Server) 10 that stores the caches, and a plurality of cache servers (Ca) that are hierarchically arranged between the root server 10 and the client (Client) 30.
che Server) 20. The client 30 is a user terminal that browses content, for example, a personal computer. The route server 10, cache server 20, and client 30 are connected by a network. The cache server 20
The content data held by the route server 10 is cached (stored) and delivered to the client 30. In the following description, the route server 10 and the cache server 20 are collectively referred to as "server". Further, although FIG. 1 illustrates the case where the number of route servers 10 is one, the number of route servers 10 may be plural.
Further, although FIG. 1 illustrates a case where the cache server 20 constitutes a two-tiered hierarchy of an upstream side (a side closer to the root server 10) and a downstream side (a side closer to the client 30), the number of layers is Any number of stages may be used as long as it is one or more. Furthermore, although FIG. 1 illustrates a configuration in which one server is connected to two downstream servers, the number of downstream servers may be one or three or more.

FIG. 2 is a block diagram schematically showing the server according to the first embodiment.

As shown in FIG. 2, the cache server 20 includes a storage device 21 for storing content data,
It has a storage management device 22 that manages the storage data stored in the storage device 21. Similarly to the cache server 20, the route server 10 also has a storage device 11 that stores content data and a storage management device 12 that manages the storage data stored in the storage device 11. However, accumulation of stream data according to the first embodiment
In the distribution system, the capacity of the storage device 21 of the cache server 20 can be made smaller than the capacity of the storage device 11 of the route server 10. Further, in the stream data storage / delivery system according to the first embodiment, the capacity of the storage device 21 of the cache server 20 can be made smaller as the downstream cache server.

<1-2> Operation of the First Embodiment FIG. 3 is a flowchart showing the distributed arrangement operation (accumulation step) of stream data in each server of the stream data accumulation / distribution system according to the first embodiment. Is.

As shown in FIG. 3, when each server 10 or 20 receives new stream data (step 101), it stores it in the storage device 11 or 21 of the server (step 102), and checks whether or not there is a downstream server. It is determined (step 103), and if a downstream server exists, it is determined whether the received stream data has been transmitted to this downstream server (step 104). If not transmitted, the storage management device 12 or 22 of the server divides the stream data into the head partial data and the remaining partial data according to a predetermined processing method described later, and transmits the head partial data to the downstream server. (Step 105). Each server 10 and 20
However, by executing the processing shown in FIG. 3, a part (head portion data) of the stream data is distributed and arranged (that is, cached) in the storage devices 21 of the cache server 20.

FIG. 4 is an explanatory diagram showing an operation when the stream data is distributed and arranged in the stream data storage / delivery system according to the first embodiment.

As shown in FIG. 4, the route server 10
Stores stream data (for example, moving image data) S. Next, the route server 10 uses the storage management device 12
In accordance with a predetermined processing method according to, the accumulated stream data S is divided into head partial data S _H and remaining partial data (that is, rear partial data) S _T , and the head partial data S _H is stored in the downstream cache server 20. Send. The process of dividing the stream data S into the head partial data S _H and the remaining partial data S _T is performed as follows. When the reproduction time when the head part data S _H of the stream data S is reproduced by the client 30 is set to P _H , the remaining part data S _T of the stream data S is transferred (cached) from the route server 10 to the downstream cache server 20. the time required for the case of a _T T, P _H ≧ T
_The stream data S is divided into head partial data S _H and remaining partial data S _T so that _T is satisfied. Server, the playback time P _H, obtained based compression method of the data, the data size, and the like. In addition, the server is required to transfer the time T
_T is calculated based on various factors such as the size of data, the transfer capability of the server, and the transmission capacity of the network connecting the servers. Thus by dividing the stream data S, the head partial data S _H cache server 2 stream data S
If stored in 0, transmitted from the cache server 20 the head portion data S _H to the client 30, the client 3
With 0, the remaining partial data S _T of the stream data S can be transmitted from the route server 10 to the cache server 20 while reproducing the image of the leading partial data S _H.

Stream data S from the route server 10
Cache server 2 which has received the first partial data S _H
The storage management device 22 of 0 stores the data in the storage device 21 of its own. Here, the beginning portion data S _H of the top portion data S if _{the H} is other cache server further downstream than the cache server 20 having received the present, stream data S received from the upstream of the stream data S, the route server By a method similar to the distributed arrangement operation by 10, the head part data S _HH and the remaining part data S _HT are divided, and the head part data S _HH is transmitted to the cache server further downstream.

Each cache server 20 continues until there is no other cache server downstream from itself.
By the same method, the operation of dividing the accumulated stream data and distributing and distributing it to the downstream cache servers is repeated. By the above distributed arrangement operation, each cache server 2
In the storage device 21 of 0, only a part of the stream data S held by the route server 10 is arranged. Although depending on the transfer rate of the network, the amount of data stored in the cache server 20 can be set to a fraction of one to several tenths of that of the route server 10. Further, the deeper the hierarchy of the cache server 20, the smaller the amount of data stored.

FIG. 5 is a flowchart showing the distribution operation of stream data in the stream data storage / distribution system according to the first embodiment.

As shown in FIG. 5, the client 30
Requests the stream data to be browsed from the cache server 20 at the closest position (step 11).
1). Upon receiving the request, the cache server 20 determines whether or not the head portion data of the corresponding stream data is stored in its own storage device 21 (step 112), and if it is stored, the head portion is immediately sent to the requesting client. The data is transmitted (step 114), and if it is not accumulated, the head part data is acquired from the upstream server (step 113) and the data is transmitted (step 114). In parallel with the transmission of the head portion data, it is judged whether or not the remaining portion data of the stream data is in the storage device 21 of the cache server 20 itself that has received the request (step 11).
5) If there is no remaining partial data, the remaining partial data is acquired from the upstream server (step 116) and transmitted to the request source client (step 117). These processes are repeated until the transmission of the requested data to the requesting client is completed.

FIG. 6 shows the distribution operation of stream data in the stream data storage / distribution system (system for distributing stream data from a root server to a client via one cache server) according to the first embodiment. It is an explanatory view shown. The storage / delivery system of FIG. 6 is a simplified system to facilitate understanding of the present invention.

As shown in FIG. 6, in the stream data storage / delivery system according to the first embodiment, the route server 10 stores the stream data S while waiting for a delivery request from the client 30. , the cache server 20 is storing the first partial data S _H stream data S (column after distributed). When there is a request from the client 30, the cache server 20 sends to the client 30 the head partial data _SH.
Is started, and acquisition of the remaining partial data S _T from the route server 10 is started (column at start of distribution). When the reproduction of the beginning portion data S _H is completed at the client 30 has completed the transmission of the remainder data S _T from the route server 10 to the cache server 20, distributed to the client 30, the remaining partial data of the stream data S It switched to the delivery of S _T (column in the delivery _{(S H} delivered at the end)). After that, the distribution of the remaining part data S _T of the stream data S ends (column at the end of distribution). In the example of FIG. 6, the data amount of the head partial data S _H stored in the storage device of the cache server 20 continues to be reproduced at least during the transfer time of the remaining partial data S _T from the root server 10. The amount of data that can be used. In this way, the distribution of the stream data S to the request source client 30 can be started immediately after the request and can be executed without interruption.

FIG. 7 shows the distribution operation of stream data in the stream data storage / distribution system (system for distributing stream data from a root server to two clients via two cache servers) according to the first embodiment. It is an explanatory view shown. The storage / delivery system of FIG. 7 is a simplified system for easy understanding of the present invention.

As shown in FIG. 7, in the first embodiment,
In the storage / distribution system of such stream data
Is waiting for the delivery request from the client
Means that the route server 10 stores the stream data S,
The cache server 20 is the head part of the stream data S.
Data S_HCache server 20 that accumulates
Is the data S_HHead part data S of_HHIs accumulating (minutes
Column after scattered placement). Distributed by request from client 30
When the communication is started, the downstream cache server 20
Is the head partial data S for the requesting client 30.
_HHStarts the distribution of the upstream cache server
Remaining data S from 20_HTStart to get (delivery
Column at the beginning). In addition, the upstream cache server 20
Data S from the route server 10_TWhen you start to get
The data S for the downstream cache server 20._Tof
Start sending (column at the time of delivery start). First part data S
_HHWhen the playback of the
Remaining data S in the bus 20_HTHas been transferred,
The remaining data S is delivered to the client 30.
_HTSwitch to delivery (During delivery (S_HH(At the end of delivery)
Column). Remaining part data S_{H T}When the playback of
Is the remaining partial data S in the downstream cache server 20. _T
Has been transferred and is delivered to the client 30.
Is the remaining data S _TSwitch to the delivery of
(S_HTAt the end of delivery) column). Then the remaining data
S_TThe delivery of is finished (column at the time of delivery end). In addition,
Data transfer when the cache server has a hierarchical structure with three or more stages
The method is performed similarly. Note that in the example of FIG.
The top stored in the storage device of the current cache server 20
Partial data S_HHData volume of at least the upstream
Remaining data S from the cache server 20_HTWhen transferring
The amount of data that can be continuously played during the period
It In addition, the head portion data S_HData volume (ie
Ta S_HHAnd data S_HTTotal amount of data)
Remaining data S of server 10_TCache server 2
You can keep playing for the entire time it takes to transfer to 0.
Data amount. In this way, the client 30
Can be started immediately after requesting delivery of stream data S to
It can be executed without interruption.

<1-3> Effects of First Embodiment In the stream data storage / delivery system and storage / delivery method according to the first embodiment described above,
Since the cache server 20 is only accumulating only the head part data of the stream data (or the head part data when the divided head part data is regarded as one stream data), the accumulation of the cache server 20 The storage capacity of the device 21 may be reduced, and the facility cost required for the storage / delivery system can be suppressed.

Further, since the head portion data of the stream data requested by the client 30 is cached in the cache server 20 closest to the client, the time from the transmission of the distribution request from the client 30 to the start of reproduction is greatly increased. Can be shortened to

Further, the data amount of the head portion data stored in the storage device 21 of the cache server 20 is set to be the data amount that can be continuously reproduced at least during the transfer time from the upstream server, so the client 30
Stream data can be delivered to the stream without interruption.

<2> Second Embodiment In the first embodiment, a stream in which almost all (or all) stream data can be reproduced at high speed by using a cache server having a small storage capacity of the storage device. The data storage / delivery system and storage / delivery method were explained. However, in the first embodiment, special reproduction (fast forward, rewind, jump, etc.) is not considered. That is, the stream data storage / delivery system according to the first embodiment is a method of obtaining the remaining partial data from the upstream server while transmitting the first partial data of the stream data to the client.
When the client issues a request such as fast-forward reproduction or jump reproduction, the cache server 20 cannot immediately respond to the request (that is, the transfer of the remaining part of the stream data from the upstream server is not in time). The stream data storage / delivery system and storage / delivery method according to the second embodiment enable immediate response to a trick play delivery request from a client.

<2-1> Configuration of the Second Embodiment FIG. 8 is a configuration diagram schematically showing a server in the second embodiment. The network configuration of the second embodiment is similar to that of FIG. 1, but the server included therein has a configuration different from that of the first embodiment, as shown in FIG.

As shown in FIG. 8, the cache server 40 includes a storage device 41 for storing content data,
The storage management device 42 manages the storage data stored in the storage device 41, and the stream dividing device 43 divides the stream data into fragment file data. In addition, in the second embodiment, the route server 10 also has a storage device 11, a storage management device 12, and a stream division device 13 as with the cache server 40. However, in the stream data storage / delivery system according to the second embodiment, the cache server 40
The capacity of the storage device 41 of the above can be made smaller than the capacity of the storage device 11 of the route server 10. Further, in the storage / delivery system according to the second embodiment, the capacity of the storage device 41 of the cache server 40 can be made smaller as the downstream cache server.

<2-2> Operation of the Second Embodiment FIG. 9 is a flowchart church showing the operation of the stream data storage / delivery system according to the second embodiment when the stream data is distributed (stored). Is. Further, FIG. 10 is an explanatory diagram showing an operation at the time of distributed arrangement of stream data in the stream data storage / distribution system according to the second embodiment.

Each server 10 or 40 has a new stream.
When the system data is received (step 201), the server stores it.
Stored in the stacking device 11 or 41 (step 202), downstream
The presence or absence of a server is determined (step 203), and the downstream server
Stream exists, the received stream data is sent to this downstream server.
(Step 204). here
If not, the stream distribution device 1 of the server
3 or 43 n stream data S (n is 2 or more)
Fragment file data of the upper integer) (step 20)
5), fragment file data S₁, S_Two,…, S_nPredetermined
Processing method (similar to the processing method in the first embodiment)
It is divided according to (step 206) and the fragment file data is
Data S₁, S_Two,…, S_nHead part data S of_1H, S
_2H,…, S_nHTo the downstream server (step 2
07). Each server executes the processing shown in FIG. 9 and FIG.
By executing the stream data S
Single file data S₁, S_Two,…, S_nEach head part of
Data S_1H, S_2H,…, S _nHBut the cache server
40 storage devices 41 are distributed (that is, cached)
Be done.

The storage management device 42 of the server is
As in the case of the above embodiment, the fragment file data S
_k(K = 1, 2, ..., N) as one stream data
Deemed fragment file data S_kIs the head part data S
_kHAnd the remaining part data (rear part data) S_kTSplit into
To do. In addition, the head portion data S_kHThe client 30
Play time in P_kHAnd the head part data S_kTo
Required to transfer (cache) to downstream server
A good time T_kTWhen expressed as P_kH≧ T_kTHolds
Each fragment file data S_kIs the head part data S
_kHAnd the remaining data S _kTSplit into.

By the above processing, the cache server 40
In the storage device 41, only part of the stream data S held by the route server 10 (the head part data of each fragment file data) is arranged. The amount of data stored in the cache server 40 is the same as in the first embodiment, and is much smaller than that of the route server 10.

Next, the distribution operation of stream data according to the second embodiment will be described. The operation during normal reproduction is the same as in the case of the first embodiment, while the first partial data S _kH of each fragment file data S _k is reproduced, the cache server _{reads the} remaining partial data S _kT from the upstream. Obtain. Following reproduction of the head portion data S _kH of the fragment file data S _k , the server stores the remaining portion data of the plurality of fragment file data and the head portion data of the plurality of fragment file data held by the server itself. And are distributed to the request source client 30 in the order of the time series of the stream data S. The normal reproduction operation in the second embodiment is similar to the operation in the first embodiment already described.

FIG. 11 is a flow chart showing the distribution operation of stream data during special reproduction in the stream data storage / distribution system according to the second embodiment.

As shown in FIG. 11, the client 3
0 requests trick play to the cache server 40 at the closest position (step 211). The storage management device 42 of the cache server 40, which has received the special reproduction request,
At that point, the normal reproduction distribution operation is stopped, and the following processing is executed according to the request from the client 30.

When there is a jump reproduction distribution request for starting reproduction from a designated position in the type determination (step 2)
12), the storage management device 42 of the cache server 40
The client 3 _receives the head part data S _kH of the fragment file data S _k of the stream data that is temporally closest to the specified position (jump destination) on the requested stream data S.
0 (step 213), and the normal reproduction operation is executed starting from that position (step 214). The normal reproducing operation is performed by a method of _{inputting / removing the} remaining partial data S _kT from the upstream server while reproducing the leading partial data S _kH .

If there is a high speed reproduction distribution request or a high speed reverse reproduction (rewind) distribution request in the type determination (step 212), the storage management device 42 of the cache server 40.
, If the current transmission to which stream data S segment file data S next fragment file data S _{k k + 1} of the _(fast rewind playback, the previous fragment file data S
The leading part data of ( _k-1 ) is sequentially taken out from the storage device (step 215) and transmitted to the client 30 (step 216), and this step is maintained as long as an instruction such as high speed reproduction is continued (step 217).

<2-3> Effects of the Second Embodiment According to the stream data storage / distribution system and storage / distribution method of the second embodiment described above, the effects of the first embodiment are obtained. In addition, special playback (fast forward playback, rewind playback, jump playback, etc.) can be provided to the user without waiting, without adding resources to the route server 10 and the cache server 30.

The number of divisions when the stream data S is divided into fragment file data S _k does not affect the capacity of the storage device required for each server, the communication volume between each server, etc. Since it is exactly the same as the embodiment, it can be realized at the same cost as the one embodiment.

<3> Third Embodiment The stream data storage / delivery system (and storage / delivery method) according to the first or second embodiment converts stream data into head data and remaining data in advance. By dividing and allocating the leading part data to the cache server in a distributed manner, the user can quickly start playing the video data etc. after a distribution request, and it is possible to perform smooth playback without interruption in the middle of the playback. To do. The stream data storage / delivery system (and storage / delivery method) according to the third embodiment is an improvement of the stream data storage / delivery system (and storage / delivery method) according to the first or second embodiment. is there. The improvement is a function that dynamically changes the arrangement of stream data (how much data is stored in which cache server) according to the frequency (frequency of use) when stream data is used (delivered to clients). Is a point having. In the stream data storage / delivery system according to the third embodiment, the more frequently used stream data S are arranged in the cache server closer to the client (user), the end cache server (that is, The cache hit rate in the cache server close to the client) is increased to reduce the amount of communication to the backbone network. In the following description, when the usage frequency of the upstream server reaches a certain threshold value, the upstream server causes the downstream server to cache the stream data accumulated by itself, which is referred to as “push”.

<3-1> Configuration of Third Embodiment Stream Data Storage / Distribution System According to Third Embodiment (System for Implementing Stream Data Storage / Distribution Method According to Third Embodiment ) Is configured as shown in FIG.
As shown in. The server in the third embodiment has a configuration different from that in the first and second embodiments, as shown in FIG. As shown in FIG. 12, the server 50 according to the third embodiment includes a storage device 51 that stores content data, a storage management device 52 that manages the storage data, and a stream division device 53 that divides stream data. A usage frequency management device 54 that manages the usage frequency (hit rate) of the cache data is provided therein. The route server 10 also has the same configuration as the cache server 50. In the stream data storage / delivery system according to the third embodiment, the stream data S of the route server 10 is distributed to the downstream cache servers 50 (that is, cache) by the same method as in the first or second embodiment. ) Will be done.

<3-2> Operation of the third embodiment The usage frequency management device 54 of the cache server 50 records the usage frequency (the number of transmissions per hour) of the corresponding stream data in its own storage device 51. The value is sequentially notified to the usage frequency management device 54 (or 14) of the upstream server. The upstream server, which has obtained the usage frequency information of the corresponding stream data from the downstream server, adds the value to its own usage frequency information, and if there is an upstream server, manages the value sequentially by the upstream server usage frequency. Notify the device 54. Here, a server has stream data,
If the frequency of use is Fs, then this server will have a frequency of use F
The following operation is performed according to the value of s. The server has a certain threshold T
s is prepared, and when Fs> Ts, the server determines that the usage frequency Fs is the threshold Ts for the storage management device of the downstream server.
Notify you have exceeded.

Upon receiving the notification that the usage frequency Fs has exceeded the threshold Ts, the downstream server checks whether or not the corresponding stream data is stored in its own storage device, and all of the corresponding stream data is stored. If not, it requests the upstream server to push the part of the stream data that it does not have. Upon receiving this request, the upstream server pushes the requested data to the downstream server. The downstream server receiving the stream data from the upstream server stores (caches) the data in its own storage device. The threshold value for the usage frequency of stream data can be changed.

FIG. 13 is a flow chart showing the distributed arrangement operation of stream data in the stream data storage / delivery system according to the third embodiment.

As shown in FIG. 13, when the server is waiting to receive a request (step 301) and there is a push request from the downstream server (step 302), the server sends the corresponding stream data to the downstream server. (Step 303).

Further, as shown in FIG. 13, when the server waits for a request (step 301) and a downstream server gives a usage frequency notification (step 302),
The usage frequency management device 54 of the server that has received the notification updates the usage frequency of the corresponding stream data in its own storage device 51 (if there are multiple downstream servers, the upstream server is notified by the multiple downstream servers). The usage frequency is added and recorded as its own usage frequency (step 304), and when the usage frequency Fs of the notified server reaches the threshold Ts (step 305), the downstream server is notified of the threshold crossing (step 305). Step 306).

Further, as shown in FIG. 13, when the server is waiting for a request (step 301), a notification of threshold crossing is received from the upstream server (step 30).
2) The downstream server that has received the notification determines whether or not the corresponding stream data has been accumulated (step 307). If not, a push request is sent to the upstream server to obtain the stream data accumulated by the upstream server. (Step 30
8) The stream data is accumulated (step 30)
9).

If the storage device of the downstream server has no free space to store the data sent from the upstream server, the downstream server must delete the stream data in the storage device in advance. is there. The selection method of the stream data to be deleted is, for example, selecting a stream to be deleted in the order of low usage frequency, selecting stream data to be deleted in the order of the oldest saved time, or deleting the stream in the order of oldest used time. However, the deletion method is not limited to these. Further, the stream data storage / delivery system according to the third embodiment states that "many users tend to browse the same title at the same time (for example, delivery requests tend to be concentrated on popular video contents. The effect can be expected if the assumption "."

<3-3> Effects of Third Embodiment According to the stream data storage / delivery system and storage / delivery method according to the third embodiment described above, the client 30 is assigned to stream data that is used more frequently. Will be pushed to the cache server in a position close to, and the stream data will be arranged very efficiently (that is, the cache hit rate in the cache server will be high). Therefore, it is possible to promptly start the delivery in response to the delivery request from the client, and to provide stable delivery without interruption.

Further, the cache server has a high cache hit rate, and the amount of communication to the backbone network can be reduced.

Furthermore, stream data that was popular in a certain area (for example, a cache server located in the Tokyo area) can be used for cache servers in other areas where the frequency of use is expected to increase in the future (for example, the Osaka area). It is also possible to push it to the cache server located at. That is, when the frequency of use of the upstream server is high, not only is stream data pushed to the downstream server, but some common element (for example, a common element such as a large city area , English-speaking common elements, common country being common elements, combinations of these, and combinations of these and other common elements, etc.). is there. In addition,
The third embodiment is the same as the first and second embodiments except the above.

<4> Fourth Embodiment <4-1> Configuration of Fourth Embodiment In the third embodiment, the “function of dynamically changing the arrangement of stream data according to the frequency of use” has been described. .
Among them, each server aggregates the usage frequency of the corresponding stream data and selects the stream data to be arranged (cached) in the downstream server according to the value.

However, in the third embodiment, when the stream data in the upstream server is transmitted (cached) to the downstream server, for example, when the stream data such as a movie is transmitted to a plurality of downstream servers. Moreover, the load on the communication line cannot be ignored. As a countermeasure for this, a method of suppressing the total amount of communication by using a communication method such as multicast or anycast can be considered, but there is still a demand for transmitting a huge file in a short time. In addition, if you use a deletion method such as “delete files in the order of low usage frequency” or “delete the last used stream data in the order of stream data”, you can secure sufficient storage capacity. There is a possibility that stream data may be deleted and re-cached frequently in a cache server that cannot.

Therefore, in the third embodiment, the unit of data to be pushed (or deleted) is the "stream unit", but in the fourth embodiment, the stepwise stream of the data amount according to the usage frequency is used. It was decided to push and delete data.

The stream data storage / delivery system according to the fourth embodiment is the same as that shown in FIG. 1, but the servers included therein are from the first to the third as shown in FIG. It has a configuration different from that of the embodiment.

As shown in FIG. 14, the server in the fourth embodiment includes a storage device 61 for storing content data and a divided storage management device 6 for managing the storage data.
2. Stream division device 6 for dividing stream data
3. In addition to the usage frequency management device 64 that manages the usage frequency of cache data, a push amount management device 65 that manages the amount of data to be pushed to the downstream server is provided inside. Further, the route server 10 is also the cache server 6
It has the same configuration as 0. In the stream data storage / delivery system according to the fourth embodiment,
The stream data S of the route server 10 is distributed (that is, cached) to the downstream cache servers by the same method as in the above embodiment.

<4-2> Operation of Fourth Embodiment In the above-described configuration, the stream data storage / delivery system according to the fourth embodiment uses the same method as that of the second embodiment. The stream data of the route server is distributed (cached) to the downstream cache servers, and the usage frequency of the stream data is sequentially notified to the usage frequency management device of the upstream server by the same method as in the third embodiment. There is.

Here, there is stream data in a certain server, and its usage frequency is represented by Fs. Similar to the third embodiment, this server prepares a certain threshold value Ts, but in the fourth embodiment, in addition to this, a monotonically increasing function f (Fs) with Fs as an argument is prepared. Note that the monotonically increasing function f (Fs) is a function that satisfies 0 = f (0) ≦ f (Fs) ≦ f (Ts) = 1 for all arguments Fs. That is, the monotonically increasing function f (Fs)
Is a monotonically increasing function that takes a value of 0 or more and 1 or less, becomes 0 when the usage frequency Fs is 0, and becomes 1 when the usage frequency Fs reaches the maximum threshold value Ts.

Each server performs the following operation according to the value of the monotonically increasing function f (Fs).

The server has a threshold value having a monotonically increasing function f (Fs) (threshold values are, for example, 0.1, 0.2, 0.3,
..., prepare multiple such as 1.0. ), The value of the monotonically increasing function f (Fs) is notified to the storage management device of the downstream server.

The downstream server, which has received the value of the monotonically increasing function f (Fs), confirms whether or not the corresponding stream data is stored in its own storage device, and when all of the corresponding stream data is not stored. Requests the upstream server to push the data it does not have. However, at this time, the push amount management device 65 of the downstream server does not request all the data of the corresponding stream data, but requests the upstream server for a data amount corresponding to the value of the monotonically increasing function f (Fs). For example, when the total data amount of the corresponding stream data is WS and the data amount already accumulated in the downstream server is WC, the data amount WR that the downstream server requests the upstream server to deliver is WR = WS × f
(Fs) -WC. However, the data amount WR requested to be distributed does not need to be exactly the same as WS × f (Fs) −WC, and may be an amount close to this.

The unit of the requested data is, for example, fragment file data obtained by evenly dividing the corresponding stream data. In this case, it is desirable that the fragment file data distributed downstream have equal intervals. .

A concrete example of the above processing contents is as follows.
Growls F (Fs) = 0.5 is notified from the upstream server
Suppose At this time, the corresponding stream is stored in its own storage device.
If 30% of the system data is accumulated,
F (Fs) -30% = 0.5-0.3 = for the server
Fragment file of 0.2 (= 20%) stream data
Request data (excluding data that you already have).
The fragment file data of the stream data is time-based.
S in column order₁, S_Two,, ...
Stream data fragment File data can be
Be evenly spaced (for example, 5 fragment file data
Fragment file data S to skip data ₁, S₆, S
₁₁,…, S_k, S_{k + 5}, ...)
It

As described above, the upstream server pushes the fragment file data of the requested corresponding stream data to the downstream server by the data amount according to the usage frequency.

The downstream server receiving the fragment file data of the stream data from the upstream server stores (caches) the fragment file data in its own storage device.
To do.

FIG. 15 is a flow chart showing the distributed arrangement operation of stream data in the stream data storage / delivery system according to the fourth embodiment.

As shown in FIG. 15, when the server is waiting to receive a request (step 401) and there is a push request from the downstream server (step 402), the server sends the corresponding stream data to the downstream server. (Step 403).

Further, as shown in FIG. 15, when the server waits for a request (step 401) and a downstream server gives a usage frequency notification (step 402),
The usage frequency management device 64 of the server that has received the notification updates the usage frequency of the corresponding stream data in its own storage device 61 (when there are multiple downstream servers, the upstream server uses the usage data notified by the multiple downstream servers). The frequency is added and recorded as its own use frequency) (step 404), and a monotonically increasing function f based on the notified use frequency Fs of the server.
Each time (Fs) reaches any of the plurality of thresholds (step 405), the downstream server is notified of the threshold crossing (step 406).

Further, as shown in FIG. 15, when the server is waiting to receive a request (step 401), there is a threshold crossing notification from the upstream server (step 40).
2) The notified downstream server judges whether or not the corresponding stream data has been accumulated (step 407), and if not, a fragment of the relevant stream data for which a push request is made to the upstream server. The column) is selected based on the threshold (step 408), the downstream server acquires the fragment sequence (step 409), and accumulates the fragment sequence of the corresponding stream data (step 410).

When the storage device of the downstream server has no free space for storing the data transmitted from the upstream server, the downstream server needs to delete the stream data in the storage device in advance. As a method of selecting stream data to be deleted, there is a method of deleting stream data fragments in the order of least-used frequency other than the method described in the third embodiment. However, the amount of stream fragments to be deleted is determined according to the frequency of use. There is also a method of deleting stream data fragments that are deleted in order of oldest stored time. However, the amount of stream fragments to be deleted is determined according to the frequency of use. Furthermore, there is also a method of selecting a stream data fragment to be deleted in order of oldest used time. Which method to choose is an operational issue, and other methods can be chosen.

<4-3> Effects of the Fourth Embodiment According to the stream data storage / delivery system and storage / delivery method according to the fourth embodiment described above, the stream data is gradually changed in accordance with the frequency of use. By pushing (and deleting) the stream data, it is possible to reduce the load on the network as compared with the method in which a large amount of data is pushed at one time. Further, even with a cache server having a small capacity, data movement (push or deletion) is performed in fragment units, so it is possible to avoid a situation where deletion and re-caching occur frequently.

The fourth embodiment is the same as the second and third embodiments except for the points described above.

[0083]

As described above, according to claims 1 to 2.
According to the inventions up to 2, since the downstream server stores only the head part data of the stream data,
The storage capacity of the downstream server may be reduced, and the facility cost required for the storage / delivery system can be suppressed.
Also, since the head portion data of the stream data requested by the client is cached in the server closest to the client, the time from the transmission of the delivery request from the client to the start of reproduction can be greatly reduced. Furthermore, since the amount of data of the first part data accumulated in the downstream server is set to the amount of data that can be continuously reproduced at least during the transfer time from the upstream server, the stream data distribution to the client can be performed without interruption. I can do it.

According to the inventions of claims 3 to 11 and claims 14 to 22, special reproduction such as fast-forward reproduction, rewind reproduction, jump reproduction and the like can be provided without waiting time.

Further, according to the inventions of claims 8 to 11 and claims 19 to 22, the stream data having a higher frequency of use is pushed to the server closer to the client, which is very efficient. Good (that is, the cache hit rate on the cache server is high)
Arrangement of stream data. Therefore, the cache hit rate in the server is high, and the amount of communication to the backbone network can be reduced.

Further, according to the inventions of claims 10, 11, 21 and 22, the load on the network can be reduced by pushing (and deleting) the stream data stepwise according to the frequency of use. Also, because data movement (push or delete) is in units of fragments,
It is possible to avoid a situation where deletion and re-caching frequently occur.

[Brief description of drawings]

FIG. 1 is a configuration diagram schematically showing a stream data storage / distribution system according to first to fourth embodiments of the present invention.

FIG. 2 is a configuration diagram schematically showing a server in the stream data storage / delivery system according to the first embodiment.

FIG. 3 is a flowchart church showing an operation of distributed arrangement of stream data in the stream data storage / delivery system according to the first embodiment.

FIG. 4 is an explanatory diagram showing an operation when stream data is distributed and arranged in the stream data storage / distribution system according to the first embodiment.

FIG. 5 is an explanatory diagram showing an operation during delivery of stream data in the stream data storage / delivery system according to the first embodiment.

FIG. 6 is an explanatory diagram showing a distribution operation of stream data in a stream data storage / distribution system (a system in which a cache server downstream of a root server is a single stage) according to the first embodiment.

FIG. 7 is an explanatory diagram showing a distribution operation of stream data in the stream data storage / distribution system (system in which the cache server downstream of the root server has two stages) according to the first embodiment.

FIG. 8 is a configuration diagram schematically showing a server in a stream data storage / delivery system according to a second embodiment of the present invention.

FIG. 9 is a flowchart church showing an operation of distributed arrangement of stream data in the stream data storage / delivery system according to the second embodiment.

FIG. 10 is an explanatory diagram showing an operation when stream data is distributed and arranged in the stream data storage / delivery system according to the second embodiment.

FIG. 11 is a flowchart church showing operations during trick play in the stream data storage / distribution system according to the second embodiment.

FIG. 12 is a configuration diagram schematically showing a server in a stream data storage / delivery system according to a third embodiment of the present invention.

FIG. 13 is a flowchart church showing an operation of changing the arrangement of stream data according to the usage frequency in the stream data storage / delivery system according to the third embodiment.

FIG. 14 is a configuration diagram schematically showing a server in a stream data storage / delivery system according to a fourth embodiment of the present invention.

FIG. 15 is a flowchart church showing an operation of changing the arrangement of stream data according to the frequency of use in the stream data storage / delivery system according to the fourth embodiment.

[Explanation of symbols]

10 route server 20, 40, 50, 60 cache server 21, 41, 51, 61 storage device 22, 42, 52, 62 storage management device 30 client 43, 63 stream division device 54, 64 usage frequency management device 65 push amount management Device S Stream data S _H Stream data beginning part data S _T Stream data remaining part data S _HH Data S _H beginning part data S _HT data S _H remaining part data S ₁ , S ₂ , ..., S _n stream data Fragment file data S _1H , S _2H , ..., _{Sn Hn} fragment file data beginning part data

Claims (22)

[Claims] 1. A stream data to be stored in a server from a first stage to a m-th stage (m is an integer of 2 or more) hierarchically arranged on a network and distributed to a client via the network. In the storage / delivery method, there is a storage step of storing a part of the first stream data held by the first stage server in a downstream server of the second stage and below, wherein the storage step is i = 1, 2, ..., [1A] performed for each of m−1, when the (i + 1) th stage server exists downstream of the i-th stage server, the i-th stage server retains the i-th stream data held by itself as the leading partial data A process of dividing into partial data, [1B] a process in which the i-th stage server transmits the head partial data of the i-th stream data to the (i + 1) -th stage server,
And [1C] the (i + 1) th stage server includes a process of holding the head partial data of the received i-th stream data as the (i + 1) th stream data, and in the process [1A], the head of the i-th stream data is included. When the playback time of the partial data by the client is P _iH and the time required to transfer the remaining partial data of the i-th stream data from the i-th stage server to the (i + 1) _-th stage server is T _iT , P _iH ≧ A method for accumulating / delivering stream data, characterized by satisfying _TiT . 2. A reproducing step of distributing the first stream data to the requesting client when the client makes a distribution request for reproducing the first stream data to the j-th (j is an integer of 2 or more and m or less) stage server. [2A] the j-th stage server distributes the j-th stream data held by itself to the request source client, [2B] the j-th stage server 1) A process of requesting and receiving the remaining part data of each of the (j-1) th to the first stream data to the upstream server of the stage or more,
And [2C] Subsequent to the processing [2A], the j-th stage server sets the remaining partial data of each of the received (j-1) to 1st stream data in time series of the first stream data. The method for storing / delivering stream data according to claim 1, further comprising a process of delivering the request to the request source client in a compliant order. 3. A stream data is stored in a server from a first stage to an m-th stage (m is an integer of 2 or more) hierarchically arranged on a network and is distributed to a client via the network. In the storage / delivery method, there is a storage step of storing a part of the first stream data held by the first stage server in a downstream server of the second stage and below, wherein the storage step is i = 1, 2, ..., [3A] When the (i + 1) th stage server exists downstream of the i-th stage server, the i-th stream data is divided into a plurality of pieces to obtain a plurality of fragment file data. Processing, [3B] processing for dividing each of the plurality of fragment file data divided in the processing [3A] into head partial data and remaining partial data, [3C] the processing [3B ], The head part data of a plurality of fragment file data divided in
And (3D) the (i + 1) th stage server stores the head part data of the plurality of received fragment file data as a plurality of (i + 1) th stream data, In the process [3B], the playback time by the client of the head part data of the plurality of fragment file data obtained by dividing the i-th stream data is P respectively.
_iH1 , P _iH2 , ..., P _iHn, and it is necessary to transfer the remaining partial data of the plurality of fragment file data obtained by dividing the i-th stream data from the i-th stage server to the (i + 1) _-th stage server. The time is _{Ti T1} ,
T _iT2, ..., in the case of the _{T iTn, P iH1} ≧ _T
iT1 , P _iH2 ≧ T _iT2 , ..., P _iHn ≧ T _iTn
A method for accumulating and delivering stream data characterized by satisfying the following conditions. 4. When the client makes a distribution request for normal reproduction of the first stream data to the j-th (j is an integer of 2 or more and m or less) stage server, the requesting client makes the first request.
A normal reproduction step of delivering stream data, wherein the normal reproduction step is [4A] in which the j-th stage server is temporally earliest among the head part data of a plurality of fragment file data held by itself. Delivering the first partial data of the fragment file data to the requesting client, [4B] The j-th stage server sends the remaining partial data of the plurality of fragment file data to the upstream servers of the (j-1) th stage and above. Process of requesting and receiving, and [4C] Subsequent to the process [4A], the j-th stage server stores the remaining part data of a plurality of fragment file data,
The j-th stage server includes a process of delivering, to the requesting client, the head part data of the plurality of fragment file data held by the j-th stage server in the order of the time series of the first stream data. A method for storing and delivering stream data according to item 3. 5. The request source client when the client issues a jump reproduction distribution request, which is a reproduction start request from the specified position of the first stream data, to the j-th (j is an integer of 2 or more and m or less) stage server. Has a first trick play process for delivering the first stream data to the first trick play process, and the first trick play process includes: [5A] The jth stage server starts a plurality of fragment file data held by itself. A process of selecting fragment file data that is temporally closest to the specified position among the partial data and delivering the leading partial data of the selected fragment file data to the request source client. [5B] The j-th stage server, The remaining part data of the selected fragment file data and the fragment file that is later than the selected fragment file data in the upstream servers of the (j-1) th stage and above. Processing for requesting and receiving the remaining portion data of the data, and [5C] Subsequent to the processing [5A], the j-th stage server receives the remaining portion data of the fragment file data received and the j-th stage server itself. 5. The method according to claim 3, further comprising a process of delivering the head portion data of the plurality of fragment file data held therein to the requesting client in the order of the time series of the first stream data. The method for storing and delivering stream data described in. 6. A second special reproduction step for distributing the first stream data to the requesting client when the client makes a fast forward reproduction distribution request to the j-th (j is an integer of 2 or more and m or less) stage server. In the second special reproduction step, the j-th stage server distributes the leading part data of the plurality of fragment file data held by itself to the request source client in order from the earliest one in time. The method for accumulating / delivering stream data according to any one of claims 3 to 5, further comprising processing. 7. A third special reproduction for distributing the first stream data to the requesting client when the client requests the j-th (j is an integer of 2 or more and m or less) stage server for the fast-reverse reproduction. In the third trick play step, the j-th stage server sequentially reverses the temporal flow of the first partial data of the plurality of fragment file data held by itself from the later one. 7. The method of storing / delivering stream data according to claim 3, further comprising a process of delivering to the requesting client. [8A] When the server delivers the stream data to the requesting client in response to the delivery request from the client, the delivering server sequentially notifies the upstream server of the usage frequency of the stream data. [8B] The upstream server updates the usage frequency of the stream data of its own based on the notified usage frequency, [8C] the updated usage frequency is further notified to the upstream server, [8D] The processes [8B] and [8C] are repeated until there are no upstream servers, and [8E] When the usage frequency of any server reaches its own threshold value, for the corresponding stream data that has reached the threshold value. The data stored in the upstream server but not in the downstream server is stored in the downstream server. The method for storing / delivering stream data described in any of the above. 9. The stream data storage / storage according to claim 8, wherein the stream data is stored in a server other than the one downstream server in parallel with the processing [8E]. Delivery method. [10A] When a server delivers stream data to a requesting client in response to a delivery request from a client, the delivering server sequentially notifies the upstream server of the usage frequency of the stream data. [10B] This upstream server updates the usage frequency of the stream data of its own based on the notified usage frequency, [10C] the updated usage frequency is further notified to the upstream server, [10D] The processes [10B] and [10C] are repeated until there are no more upstream servers, and [10E] each server has a plurality of thresholds and a plurality of fragment file data information corresponding to each of the plurality of thresholds. Each time the usage frequency of the upstream server reaches any of a plurality of thresholds that the upstream server has in advance 8. The fragment file data which is in the upstream server but is not in the downstream server is accumulated in the downstream server in an amount determined by the reached threshold value, according to any one of claims 3 to 7. How to store and distribute stream data. 11. The stream data storage / storage of claim 10, wherein the stream data is stored in a server other than the one downstream server in parallel with the processing [10E]. Delivery method. 12. The server has first to m-th (m is an integer of 2 or more) stages hierarchically arranged on a network, and stream data is sent to the first-stage to m-th servers. In a stream data storage / delivery system that stores and delivers to a client via a network, each of the servers from the first stage to the m-th stage stores a part of the first stream data held by the first stage server. When accumulating in the downstream servers after the second tier: i = 1, 2, ..., M−1, respectively, [12A] When the (i + 1) th tier server exists downstream of the i th tier server , The i-th stage server divides the i-th stream data held by itself into the head part data and the remaining part data, and [12B] the i-th stage server converts the i-th stream data to the (i +) th part of the head part data. [12C] The (i + 1) th stage server holds the head part data of the received i-th stream data as the (i + 1) th stream data, and stores the data in the servers from the 1st stage to the m-th stage. In each of the above processes [12A], the playback time by the client of the head part data of the i-th stream data is set to P _iH, and the remaining part data of the i-th stream data is transferred from the i-th stage server to the (i + 1) _-th stage server. A stream data storage / distribution system configured to satisfy P _iH ≧ T _iT , where T _iT is the time required to _perform . 13. The j-th (j is 2 or more m from the client)
[13A] The j-th stage server that the j-th stage server holds when the j-th stage server makes a reproduction distribution request for reproduction of the first stream data.
[13B] The j-th stage server distributes the stream data to the request source client, and the j-th stage or higher upstream servers deliver the remaining partial data of each of the (j-1) to the first stream data. [13C] Subsequent to the processing [13A], the j-th stage server receives the remaining partial data of each of the received (j-1) to 1st stream data as the first stream data. 13. The stream data storage / delivery system according to claim 12, wherein the stream data is delivered to the request source client in order according to the time series. 14. A server having first to m-th (m is an integer of 2 or more) stages hierarchically arranged on a network, and stream data is transmitted to the servers from the first-stage to the m-th stage. In a stream data storage / delivery system that stores and delivers to a client via a network, each of the servers from the first stage to the m-th stage stores a part of the first stream data held by the first stage server. When accumulating in the downstream servers of two stages or less, i = 1, 2, ..., M−1, respectively, [14A] When the (i + 1) th stage server exists downstream of the i-th stage server. , The i-th stream data is divided into a plurality of pieces of fragment file data, and [14B] each of the plurality of pieces of fragment file data divided in the process [14A] is a head partial data. Divided into data and remainder data, first the head portion data of a plurality of fragmented files data divided in [14C] the process [14B] (i +
1) It is transmitted to the stage server, and [14D] The (i + 1) th stage server accumulates the head part data of the plurality of received fragment file data as a plurality of (i + 1) th stream data, and from the first stage Each of the servers up to the m-th stage sets the playback time by the client of the first partial data of the plurality of fragment file data obtained by dividing the i-th stream data in the process [14B], respectively.
_iH1 , ..., P _iHn, and the time required to transfer the remaining partial data of the plurality of fragment file data obtained by dividing the i-th stream data from the i-th stage server to the (i + 1) _-th stage server respectively. T _iT1 , ..., T
_In the case of _iTn , P _iH1 ≧ T _iT1 and P _iH2 ≧
A stream data storage / distribution system configured to satisfy T _iT2 , ..., P _iHn ≧ T _iTn . 15. The j-th (j is 2 or more m from the client)
(The following integer) When a normal-stream delivery request for the first stream data is made to the stage server, [15A] The j-th stage server holds the head part data of the plurality of fragment file data held by itself. The head part data of the fragment file data which is the earliest in time is delivered to the requesting client, and [15B] the j-th stage server transfers the plurality of fragment file data to the (j-1) -th and higher upstream servers. Requesting and receiving the remaining part data, [15C] Subsequent to the processing [15A], the j-th stage server holds the remaining part data of a plurality of fragment file data and the j-th stage server itself. 15. The stream according to claim 14, wherein the head part data of the plurality of fragment file data is distributed to the requesting client in the order of the time series of the first stream data. Over data storage and delivery system. 16. The j-th (j is 2 or more m from the client)
[16A] When the j-th server has a jump reproduction distribution request, which is a reproduction start request from the designated position of the first stream data, [16A] Of the head part data of the fragment file data, the fragment file data that is closest in time to the specified position is selected, and the head part data of the selected fragment file data is delivered to the requesting client, [16B] j. The stage server requests the remaining part data of the selected fragment file data and the remaining part data of the fragment file data temporally later than the selected fragment file data to the upstream servers of the (j-1) th stage and above. [16C] Subsequent to the processing [16A], the j-th stage server receives the remaining partial data of the received fragment file data and the j-th stage. 16. The server distributes the head part data of the plurality of fragment file data held by itself to the requesting client in the order of the time series of the first stream data. A stream data storage / distribution system described in Kani. 17. The j-th (j is 2 or more m from the client)
(The following integer) When a fast-forward playback distribution request is made to the stage server, the j-th stage server sequentially requests the first partial data of the plurality of fragment file data held by itself from the temporally earlier one. The stream data storage / delivery system according to any one of claims 14 to 16, wherein the stream data is delivered to the original client. 18. The client receives the j-th (j is 2 or more m
(Integral number below) When the junior server makes a request for delivery for fast-reverse reproduction, the j-th server indicates that the head part data of the plurality of fragment file data held by the j-th server is the time after the first one. 18. The stream data storage / delivery system according to any one of claims 14 to 17, wherein the stream data is delivered to the request source client in the reverse order of the flow. [19A] When the server delivers the stream data to the requesting client in response to the delivery request from the client, the delivering server sequentially notifies the upstream server of the usage frequency of the stream data. [19B] This upstream server updates the usage frequency of the stream data of its own based on the notified usage frequency, [19C] the updated usage frequency is further notified to the upstream server, [19D] The processes [19B] and [19C] are repeated until there are no more upstream servers, and [19E] When the usage frequency of one of the servers reaches its own threshold value, regarding the corresponding stream data that has reached the threshold value. , It is characterized by accumulating data in the downstream server that is in the upstream server but not in the downstream server A storage / delivery system for stream data according to any one of claims 12 to 18. 20. The stream data storage / storage of claim 19, wherein the stream data is stored in a server other than the one downstream server in parallel with the processing [19E]. Delivery system. [21A] When a server delivers stream data to a requesting client in response to a delivery request from a client, the delivering server sequentially notifies the upstream server of the usage frequency of the stream data. [21B] This upstream server updates the usage frequency of the stream data of its own based on the notified usage frequency, [21C] the updated usage frequency is further notified to the upstream server, [21D] The processes [21B] and [21C] are repeated until there are no more upstream servers, and [21E] each server has a plurality of thresholds and a plurality of fragment file data information corresponding to each of the plurality of thresholds. Each time the usage frequency of the upstream server reaches any of a plurality of thresholds that the upstream server has in advance 19. The fragment file data which is in the upstream server but is not in the downstream server is accumulated in the downstream server in an amount determined by the reached threshold value. Stream data storage / distribution system. 22. The stream data storage / storage of claim 21, wherein the stream data is stored in a server other than the one downstream server in parallel with the processing [21E]. Delivery system.