Data storage transfer system and data transfer schedule determining method
Abstract
[Task] Data that does not necessarily need to be sent in real time can be transferred before the transfer completion request time while suppressing the load on the information processing device, storage capacity, and communication network, and all data can be transferred without causing a deadlock. send.
Solution.Data transmission / reception order based on the state of the information processing device and the state of the communication network, provided that the data transfer to the target information processing device is completed before the transfer completion request time predetermined for each data. , The data transfer path, the data storage time and transfer start time of the processing device on the transfer path are determined, and the data transfer schedule is determined. Therefore, virtual frame FkTime of the sum of the specified data sizes + DataiThe sum of the sizes of is the virtual frame FkWhen the capacity of Data is not exceeded (step 106), DataiTo the next virtual frame FiAssign to (step 107).

Term
Term ended
Projected expiry passed 16 March 2019, 7.5 years ago.
- Priority and filed
- Published
- Projected expiry
- Today
5 claims: 2 independent, 3 dependent
- 1【特許請求の範囲】 【請求項1】 データを蓄積して転送する複数の情報処理装置を通信ネットワークを介して接続し、該情報処理装置相互間でデータの送受信を行う蓄積型データ転送方式において、 各情報処理装置には、各データに予め定められている転送完了要求時刻以前に目的の情報処理装置へのデータ転送が完了することを条件として、該情報処理装置の状態および通信ネットワークの状態に基づき、データの送受信順序とデータの転送経路と該転送経路上の情報処理装置のデータ蓄積時間と転送開始時刻とを決定するためのデータ転送スケジュール決定手段を具備したことを特徴とする蓄積型データ転送方式。
- 2【請求項2】 請求項1に記載の蓄積型データ転送方式において、 前記データ転送スケジュール決定手段は、決定したスケジュールと実際の転送の実行との間に時間的ずれが生じた場合、改めてデータの送受信順序とデータの転送経路と該転送経路上の情報処理装置のデータ蓄積時間と転送開始時刻とを決定し直すことを特徴とする蓄積型データ転送方式。
- 3【請求項3】 請求項1または請求項2に記載の蓄積型データ転送方式において、 前記データ転送スケジュール決定手段は、データ転送順序を時系列の転送周期を有する仮想フレームに分割し、分割された仮想フレームの長さと1個の仮想フレームで転送するデータ容量は全て同一量とし、 その場合、測定した情報処理装置および通信ネットワークの状態に基づいて、該情報処理装置および通信ネットワークの現在から将来にわたる輻輳状態を決定ないし推定し、それに応じて将来にわたる各仮想フレームの長さとデータ容量を決定し、 あるデータの転送要求が発生したとき、当該データを転送する仮想フレームを決定し、 情報処理装置間で同一仮想フレーム番号を持つ仮想フレーム内データの転送が終了すると、全ての情報処理装置で同期をとって次の仮想フレームの転送処理に移行することを特徴とする蓄積型データ転送方式。
- 4【請求項4】 請求項3に記載の蓄積型データ転送方式において、 前記データ転送スケジュール決定手段は、転送要求データが発生すると、時間指定されたデータについては、各仮想フレームの転送開始予定時刻以前の仮想フレームに割り当て、 時間指定されていないデータについては、現在転送中の次の仮想フレームから後続する仮想フレームに順に割り当て可能か否かをチェックし、不可能であれば次のフレームに割り当て可能か否かをチェックするという動作を反復して、割り当て可能な仮想フレームに実際に割り当てることを特徴とする蓄積型データ転送方式。
- 5【請求項5】 データを蓄積して転送する複数の情報処理装置を通信ネットワークを介して接続し、該情報処理装置相互間でデータの送受信を行う際のデータ転送スケジュール決定方法において、 該情報処理装置間で同一仮想フレーム番号を持つ仮想フレーム内のデータ集合を転送する場合に、データは全て同一サイズのブロックに分解して、ブロック単位で転送し、先ず全てのブロックに‘unscanned’とラベリングし、収容している全てのブロックのラベルが‘unscanned’である情報処理装置に‘unfinished’とラベリングし、ブロック転送の結果、収容しているブロック全てのラベルが‘scanned’になれば‘finished’に更新し、‘unfinished’とラベルがついている情報転送装置vを1つ選択し、トークンptを情報転送装置vで生成してsearch(v)コマンドを実行し、この動作を全ての情報処理装置のラベルが‘finished’になるまで行い、トークンおよびブロックの転送時に、情報処理装置や通信ネットワークが輻輳状態であれば、転送を一時中断し、輻輳状態から回復した時点で転送を再開するとともに、下記ステップに従って動作することを特徴とするデータ転送スケジュール決定方法。 (Step.1)トークンptが情報処理装置wで生成された場合(w=v)、Step.2.1に進み、(Step.2)送られてきたトークンptの付加情報として、ブロック番号と当該ブロックを収容する情報処理装置名の組が入っている場合、この組を転送リストにリスト構造で記憶し、(Step.2.1)情報処理装置wのラベルが‘unfinished’であれば、情報処理装置wが収容している‘unscanned’ラベルのブロックeを選択し、該ブロックeのラベルを‘scanned’に更新し、その結果、情報処理装置wに収容している‘unscanned’ラベルのブロックがなくなれば、情報処理装置wのラベルを‘finished’に更新し、最後にトークンptにブロック名eとwの情報処理装置名の組を付加して、ブロックeの転送先情報処理装置に送信し、(Step.2.2)情報処理装置wのラベルが‘finished’で、かつ情報処理装置wの空き容量が単位ブロックサイズ以上、さらに転送リストが空集合でなければ、転送リストの最後尾に書き込まれているブロックeを転送させ、実際に転送が終了すると、転送リストから該ブロックeに関する情報を削除し、(Step.2.3)情報処理装置wのラベルが‘finished’で、かつ情報処理装置wの空き容量が0、さらに転送リストが空集合でなければ、転送リストの最後尾に書き込まれているブロックを収容する情報処理装置に対して、リスト再構成コマンドをトークンptに付加して送信し、転送リストから当該ブロックに関する情報を削除し、(Step.2.4)情報処理装置wのラベルが‘finished’で、かつ転送リストが空集合であれば、search(v)を終了し、(Step.3)送られてきたトークンptの付加情報として、ブロックeを転送させるコマンドが付加されている場合には、ブロックeをその転送先情報処理装置に転送し、wのラベルが‘unfinished’ならばStep.2.1へ、wのラベルが‘finished’で、かつ情報処理装置wの空き容量が単位ブロックサイズ以上ならばStep.2.2へ、wのラベルが‘finished’で、かつ情報処理装置wの空き容量が0ならばStep.2.3へ、そうでなければStep.2.4へそれぞれ進み、(Step.4)送られてきたトークンptの付加情報として、リスト再構成コマンドが付加されている場合には、wのラベルが‘unfinished’であればStep.2.1へ、wのコマンドが‘finished’で、かつ情報処理装置wの空き容量が単位ブロックサイズ以上であればStep.2.2へ、どちらでもなければStep.2.3へ、それぞれ進む。
Independent claims5
51 paragraphs in 1 section, as filed
Description: TECHNICAL FIELD [Detailed description of the invention]
【0001】
[Technical field to which the invention belongs]
The present invention relates to a storage type data transfer method and a data transfer schedule determination method that complete a transfer before the transfer completion request time while suppressing the load, storage capacity, and communication network load of the information processing device, and particularly accumulate data. When connecting multiple information processing devices with a transfer function via a network and sending and receiving data between the information processing devices, a method in which a user with a mobile terminal uses data from a fixed terminal at the destination, or VOD The present invention relates to a storage type data transfer method and a data transfer schedule determination method that can be applied as a method of requesting a video by specifying a time in (Video On Demand) or a real-time communication method of transferring by specifying a time in packet units.
【0002】
[Conventional technology]
Conventionally, a method of temporarily accumulating data has been used in order to reduce the load on a communication network or a server and improve the performance. For example, in the "burst server storage-and-forward method" described in JP-A-6-338906, when bandwidth is reserved on a route for data transfer, data is temporarily stored on the data transfer route. , Increasing the probability that the bandwidth required for data transfer can be reserved via all links. However, this method cannot satisfy the specified transfer completion time. In addition, in the "communication network" described in JP-A-6-69921, when the required bandwidth cannot be obtained immediately, the start and end times, transfer routes, etc. are determined for non-real-time data transfer, and the data is temporarily transferred. By accumulating data, the utilization efficiency of communication network resources is improved. However, since the transfer schedule is not determined in consideration of the state of the communication network and the like, and the transfer schedule once determined is not redetermined according to the change in the situation, the transfer completion request time is always set. I can't be satisfied.
【0003】
[Problems to be Solved by the Invention]
In this way, with the conventional technology, it was possible to reduce the load on the communication network and server and improve the performance, but there is no reliable proof that the transfer can always be performed before the specified transfer completion time. Therefore, the transfer completion request time could not always be satisfied. Therefore, to satisfy the transfer completion request time, to enable the transfer completion request to be satisfied by rescheduling even if the transfer completion request time cannot be satisfied with the once determined schedule, and to accumulate data. It is an object of the present invention to enable all data to be transferred without causing a deadlock that cannot be transferred because the free space of the data transfer destination is insufficient even when the free space of the data is small.
【0004】
Therefore, an object of the present invention is to solve these conventional problems and transfer data that does not necessarily need to be transmitted in real time while suppressing the load of the information processing device, the storage capacity, and the load of the communication network. It is an object of the present invention to provide a storage type data transfer method and a data transfer schedule determination method that can be transferred before the completion request time. Another object of the present invention is to transfer all data without causing a dead lock that cannot be transferred because the free space of the data transfer destination is insufficient even when the free space for data storage is small. It is to provide a storage type data transfer method and a data transfer schedule determination method which can be performed.
【0005】
[Means for solving problems]
In order to achieve the above object, the storage-type data transfer method of the present invention provides the information processing apparatus with the condition that the data is transferred to the destination information processing apparatus before the transfer completion request time predetermined for each data. A data transfer schedule determining means for determining a data transmission / reception order, a data transfer path, and a data storage time and a transfer start time in an information processing apparatus on the transfer path based on a state and a state of a communication line, and the data transfer schedule. A means for providing a measurement amount for expressing the state of a communication network and a measurement amount for expressing the state of an information processing device to a determination means, and a means for accumulating and transferring data according to a schedule determined by a data transfer schedule determination means. Is equipped with.
【0006】
In the present invention, for example, when a user having a mobile terminal uses data that is normally used by connecting to a server on a LAN from a fixed terminal at a mobile destination, the data is close to the mobile terminal from the viewpoint of communication delay and the like. It is preferable that the data is on the server, and in that case, the data normally used from the fixed terminal may be transferred to the destination by the time of use at the destination. At this time, if the transfer is completed before the transfer completion request time while suppressing the load of the information processing device, the storage capacity, and the load on the network, the network operator can effectively use the equipment, and the user can use the equipment effectively. Can be expected to improve convenience. Further, in VOD (Video On Demand), it is expected that a user requests a certain video by specifying a time, but in that case, it is sufficient that the video data is transferred by the transfer completion request time. Further, by including information on the transfer completion request time of the packet in the packet, it can be applied to a real-time communication system in which the time is specified for each packet and transferred.
【0007】
The function of the data transfer schedule determining means of the present invention makes it possible to end the data before the transfer completion request time. In addition, since the data transfer schedule determination means reschedules according to changes in the state of the communication network and the information transfer device, it is affected by real-time communication and high-priority data transfer requests. It is possible to meet the transfer completion request time without any problems. Here, as the measurement amount expressing the state of the communication network, for example, there are the number of hops, the round trip time, the packet loss rate, the minimum link bandwidth on the route, the throughput (information transfer amount ÷ transfer time), and the like. It is also possible to use a function value with these values as parameters as a network state. Further, the measured quantity expressing the state of the information processing device includes the CPU load, the number of processes, and the like. These can be measured using various tools already in practical use.
【0008】
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, examples of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing a configuration example of a server and a control device to which the present invention is applied via a communication network. Each information processing device (here, servers 11 to 14) measures the load and storage capacity of the information processing device and the load of the communication network, and notifies the control device 10. Further, each information processing device accumulates data according to an instruction from the control device 10 and transfers the data to another information processing device. On the other hand, the control device 10 determines a data transfer schedule based on the information notified from the information processing devices 11 to 14, and notifies the information processing devices 11 to 14 of the schedule. Hereinafter, the information processing device is referred to as a server. The broken line 20 in FIG. 1 is the communication line between the control device 10 and the servers 11 to 14, and the solid line 30 is the communication line between the servers 11 to 14. The servers 11 to 14 and the control device 10 are arranged on the communication network. First, each virtual frame length and virtual frame capacity are determined based on the measured values of the states of the servers 11 to 14 and the communication network and the predicted values in the future.
【0009】
FIG. 5 is a diagram showing a configuration in one server in FIG. The server 11 has a data transfer schedule determining means 16, a means 17 for providing a measurement amount expressing the state of the communication network and the information processing device, and a means for accumulating and transferring data according to the schedule determined by the schedule determining means 16. 18 and are provided. The data transfer schedule determining means 16 determines the data transmission / reception order and the data transfer route based on the state of the information processing device and the state of the communication line on condition that the data is transferred to the target server before the transfer completion request time. Determine the data accumulation time and transfer start time at the server on the transfer path. The state expression measurement quantity providing means 17 provides a measurement quantity that expresses the state of the communication network and a measurement quantity that expresses the state of the server. The data storage / transfer means 18 accumulates data according to a schedule, reads the data, and transfers the data to the communication line 20.
【0010】
FIG. 2 is a processing flowchart in the server when a data transfer request is generated, which shows an embodiment of the present invention. Data on server p<sub>i</sub>When a transfer request to server q occurs (step 101), the virtual frame number k currently being transferred gradually increases by 1 from k = 0 (steps 102, 103), and the virtual including the data transfer completion request time. If frame number B is transferred and there is no such virtual frame (step 104), then the data Data<sub>i</sub>Returns an error that the transfer cannot be made before the transfer request time (step 105). On the other hand, data Data<sub>i</sub>If is specified, the virtual frame F before the specified time<sub>k</sub>And virtual frame F<sub>k</sub>The sum of the size of the data already allocated to and timed and the data Data<sub>i</sub>The sum of the sizes of is the virtual frame F<sub>k</sub>If there is something that does not exceed the capacity of (step 106), virtual frame F<sub>k</sub>Data to Data<sub>i</sub>(Step 107).
【0011】
FIG. 3 is a flowchart of a virtual frame transfer process showing an embodiment of the present invention. Server p is virtual frame F<sub>0</sub>Before starting the transfer of the virtual frame F, the data transfer request for which the time is not specified is sent in order from the one with the largest data size within the range where the free space is not exceeded in the virtual frame capacity.<sub>0</sub>Assign to (step 208). Next, the server p is the virtual frame F<sub>0</sub>With the start of the transfer (step 201), the virtual frame start time is synchronized with all other servers (step 202). In addition, the starting server is elected from all the servers (step 203). Then, according to the schedule control procedure shown below, the virtual frame F<sub>0</sub>The transfer of the internal data is started (step 204). Virtual frame F on each server<sub>0</sub>When the transfer of internal data is completed, the next virtual frame F<sub>1</sub>If it is before the scheduled start time of, jump to step 208 (step 205). Meanwhile, the next virtual frame F<sub>1</sub>If it is after the scheduled start time of, the virtual frame F will be renewed in order from the one with the earliest transfer completion request time.<sub>1</sub>Reassign to the one with the earliest scheduled start time among the subsequent allocatable virtual frames (step 206). As a result, if there is something that does not satisfy the transfer completion request time, an error is returned in response to the transfer request (step 207). Otherwise, proceed to step 208.
【0012】
FIG. 4 is a flowchart of a schedule control procedure showing an embodiment of the present invention. All data shall be decomposed into blocks of the same size and transferred in block units. First, label all blocks as'unscanned'. Label the information transfer device with the label'unscanned' on all contained blocks as'unfinished'. If the label of all the contained blocks becomes'scanned'as a result of block transfer, it is updated to'finished'. Select one information processing device v labeled'unfinished', generate a token pt with the information transfer device v, and execute search (v). Do this until the labels on all information processors are'finished'. If the information processing device or communication network is in a congested state when transferring tokens and blocks, the transfer shall be suspended and the transfer shall be resumed when the congested state is recovered. Hereinafter, the information processing device holding the token pt is referred to as w.
【0013】
Execute the search (v) instructions in the following order. When w = v, that is, when the token pt is generated by the information processing device w, the process jumps to step 305 (step 301). Next, if it corresponds to Case 1 and the block number and block accommodating server name are included as additional information of pt (step 302), proceed in the direction of the arrow of Case 1 and store this set in the transfer list in a list structure. (Step 304). If a command to transfer block e is added as additional information of the sent token pt corresponding to Case 2 (step 302), the process proceeds in the direction of the arrow in Case 2 and block e is transferred to the transfer destination. Transfer to the information processing device (step 303). If a list reconstruction command is added as additional information of the sent token pt corresponding to Case 3, the process proceeds in the direction of the arrow in Case 3 and the process proceeds to step 305.
【0014】
Next, in step 305, if Case 1 is applicable and the label of the information processing device w is'unfinished', the process proceeds to the arrow of Case 1 and the block e of the'unscanned' label contained in the information processing device w is selected. And update the label of e to'scanned' (step 306). Then, the pair of the information processing device names of the block names e and w is added to the token pt and transmitted to the transfer destination information processing device of e (step 307). As a result, when the block of the'unscanned'label contained in the information processing apparatus w disappears (step 308), the label of the information processing apparatus w is updated to'finished' (step 309). If the label of the information processing device w is'finished', the free space of the information processing device w is equal to or larger than the unit block size, and the transfer list is not an empty set, proceed to the arrow of Case 2 and transfer. Transfer the block e written at the end of the list. That is, a command for transferring e to the information processing device z accommodating e is added to the token pt and transmitted (step 310). When the transfer is actually completed, the information about e is deleted from the transfer list.
【0015】
Further, in step 305, if it corresponds to Case 3, the label of the information processing device w is'finished', the free space of the information processing device w is 0, and the transfer list is not an empty set, the process proceeds to the arrow of Case 3. A list reconstruction command is added to the token pt and transmitted to the information processing device that accommodates the block written at the end of the transfer list (step 311), and the information about the block is deleted from the transfer list. If it corresponds to Case 4, the label of the information processing device w is'finished', and the transfer list is an empty set, the process proceeds to the arrow of Case 4 and search (v) is terminated.
【0016】
[Effect of the invention]
As described above, according to the present invention, data that does not necessarily need to be transmitted in real time is transferred before the transfer completion request time while suppressing the load of the information processing device, the storage capacity, and the load of the communication network. be able to. Further, even when the free space for data storage is small, all the data can be transferred without causing a deadlock caused by insufficient free space at the data transfer destination.
[Simple explanation of drawings]
[Figure 1]
It is a figure which shows the configuration example through the communication network of the server and the control device to which this invention is applied.
[Figure 2]
It is a flowchart of the process of the server at the time of the occurrence of the data transfer request which shows one Example of this invention.
[Fig. 3]
It is a flowchart of the virtual frame transfer process which shows one Example of this invention.
[Fig. 4]
It is a flowchart of the schedule control procedure which shows one Example of this invention.
[Fig. 5]
It is a figure which shows the component of the server in FIG.
[Explanation of symbols]
10 ... Control device, 11 ~ 14 ... Server (information processing device), 20 ... Communication line between control device and server, 30 ... Communication line between servers, 16 ... Data transfer Schedule determination means, 17 ... state expression measurement amount providing means, 18 ... data storage / transfer means.
6 sheets
Sheet 1 Sheet 2 Sheet 3 Sheet 4 Sheet 5 Sheet 6
Every citation, both ways
| Document | Relation | Office | Cited during |
|---|---|---|---|
| KR20120019679A | Cited by | Republic of Korea | Search report |
| WO2013034798A1 | Cited by | World Intellectual Property Organization (WIPO) | International search |
| US9274898B2 | Cited by | United States of America | Applicant |
| KR20120018988A | Cited by | Republic of Korea | Search report |
| US10292181B2 | Cited by | United States of America | Applicant |
| JP2005020235A | Cited by | Japan | Search report |
| US9112726B2 | Cited by | United States of America | Applicant |
2 priority claims, no other members on record
Priority claims2
| Document | Office | Kind | Date |
|---|---|---|---|
| 6940899 | Japan | A | |
| JP19990069408 | – | – | – |
2 legal events, as the office reported them to INPADOC
Over the term
Point at a mark for the eventEvents
| Event | Code | |
|---|---|---|
| Decision of refusalJAPANESE INTERMEDIATE CODE: A02A02 | A02 | |
| Notification of reasons for refusalJAPANESE INTERMEDIATE CODE: A131A131 | A131 |
Numbers
- Publication
- 2000-270010
- Publication, DOCDB
- 2000270010
- Publication, EPODOC
- JP2000270010
- Application
- 11069408
- Application, DOCDB
- 6940899
- Application, EPODOC
- JP19990069408
Titles2
- Japanese
- 蓄積型データ転送方式およびデータ転送スケジュール決定方法
- English
- INDUSTRIAL APPLICABILITY: Storage type data transfer method and data transfer schedule determination method
Classification
- IPC, 6
- G06F13 00
- G06F15 177
- H04L12 54
- H04L12 58
- H04L29 08
- H04M3 00