JP2007235681A - Distribution tree configuration apparatus, distribution tree configuration method and program for distribution tree configuration - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide techniques, related to a distribution tree configuration apparatus and the like, capable of reconfiguring a distribution tree according to the situation, even when distribution path performance deterioration occurs or the like. <P>SOLUTION: Until ending multicast distribution, distribution path throughput measurement, performance analysis and distribution tree configuration are performed iteratively. Thus, even without grasping physical network constitution, switch to a distribution tree of higher distribution efficiency according to the situation with respect to distribution path performance deterioration can be made. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a distribution tree configuration, and more particularly to a distribution tree configuration for distributing data by multicast.

  In broadcast distribution that distributes the same data to multiple destinations, IP multicast that replicates data at intermediate routers is widely used due to the necessity of router equipment that supports IP multicast. Is not popular. Therefore, unicast communication is performed between nodes (end hosts), and broadcast distribution is performed by application level multicast (ALM) in which multicast communication is performed by copying and transferring data between nodes.

  A distribution system based on ALM generally has a configuration as shown in FIG. Referring to FIG. 12, the ALM distribution system includes a transmission server 1 for transmitting data and managing a distribution tree, receiving nodes 4 and 6 for transmitting / receiving data and notifying the transmission server 1 of distribution status, The network 5 is configured.

  The transmission server 1 includes, for example, a command input unit 11 that receives an input such as a distribution start command from a user, a data transmission unit 12 that transmits data along a distribution tree, and a throughput measurement unit Z13 that measures data transmission throughput. A tree management device 2 that manages creation of a delivery tree and a delivery status, and a storage device 3 that stores data used for creating the delivery tree and the delivery tree.

  The tree management device 2 includes distribution tree creation means Z21 for creating a distribution tree, and information collection means Z22 for collecting information indicating, for example, the distribution status to each receiving node.

  The storage device 3 creates a delivery tree storage unit 31 for storing the delivery tree created by the delivery tree creation means Z21, and creates a delivery tree, such as delay time and throughput measured in advance for each path. A network information storage unit Z32 for storing data to be referred to.

  The receiving node 4 measures the amount of received data, a data transmitting unit 41 that transmits data along the distribution tree, a data receiving unit 42 that receives data, a throughput measuring unit Z43 that measures data transmission throughput. A received data amount measuring unit Z44 and an information contact unit Z45 for contacting information indicating completion of distribution to the information collecting unit Z22 via the network 5 are provided.

  In the transmission server 1, when an input of distribution start is given to the command input means 11, the distribution tree creation means Z21, for example, information such as delay time and throughput of each path measured in advance, as network information For example, a distribution tree such as Shortest Path Tree is created from the storage unit Z32 and stored in the distribution tree storage unit 31. When the distribution tree is created, the data transmission unit 12 reads the distribution tree from the distribution tree storage unit 31 and transmits data along the distribution tree.

  When the data receiving unit 42 of the receiving node 4 receives the data, the data transmitting unit 41 transmits the data to the next receiving node 6.

  In parallel with the data transmission / reception, the throughput measuring means Z13 and Z43 measure the data transmission throughput, and the received data amount measuring means Z44 measures the data amount received so far by the own node.

  When the reception of data at the receiving node 4 is completed, the information communication unit Z45 transmits a notification of distribution completion to the information collection unit Z22 via the network 5. When the information collecting unit Z22 receives notification of distribution completion from all the receiving nodes, the ALM distribution system ends the multicast distribution.

  ALM does not grasp the physical configuration of actual routers and links, but creates a distribution tree as shown in FIG. 13 using a logical topology such as connections between nodes, and follows the distribution tree. Data transmission.

  Referring to FIG. 13, the distribution tree includes a transmission server 1, reception nodes 4-1 to X, a connection between the transmission server and the reception node, and a connection between the reception node and the reception node (the above connection is referred to as a path). Has been. Each of the reception nodes 4-1 to X has the same configuration and function as the reception server 4 of FIG. In FIG. 13, a three-stage distribution tree is illustrated as an example. However, the number of stages in the distribution tree is not determined, and varies depending on the network situation, the distribution tree creation algorithm, and the like.

  When a distribution tree as shown in FIG. 13 is created, data unicast transmission is performed from the top to the bottom of the distribution tree during data distribution. The transmission server 1 transmits data to the reception nodes 4-1 to 4 -M by unicast communication, and the reception nodes 4-1 to 4 -M are connected to the downstream reception nodes 4 -N to 4-1 connected by a path. Data received for X is transferred by unicast communication.

  However, when creating a distribution tree in ALM, the physical configuration such as routers and links used in actual communication cannot be recognized, so that multiple paths in the distribution tree compete for a specific link, resulting in degraded distribution performance. As a result, there is a problem that it takes time to complete distribution to all nodes.

  On the other hand, conventionally, for example, a distribution tree configuration device, a distribution tree configuration method, and a distribution tree configuration program as shown in Non-Patent Document 1 are used to reduce contention between multiple paths of a distribution tree in a link. ing.

  The distribution tree configuration device, the distribution tree configuration method, and the distribution tree configuration program described in Non-Patent Document 1 investigate a path conflict in a link after creating the distribution tree, and a plurality of conflicting links If there is a path, data is transferred to the destination node of the remaining path via the destination node of one of the paths, thereby reducing path contention in the physical link.

  A mechanism for reducing path contention in the system will be described in detail with reference to the example of FIG. FIG. 14 is a part of a distribution tree as shown in FIG. 13, and is described by a physical configuration. Here, it is assumed that the receiving node 4-1 transmits data, and the receiving nodes 4-2 and 4-3 receive and transfer data. The router 71 is a router that exists at a branch point of a connection that connects each receiving node. The link between each receiving node and the router 71 is a physical configuration link, and the paths 81 to 83 are logical configuration paths. is there.

  When the receiving node participates in the distribution tree, the system checks in advance the router through which the delay time in each link passes by the traceroute command. In this example, it is assumed that when the distribution tree is created, paths 81 and 82 are incorporated in the distribution tree. Then, in the link between the receiving node 4-1 and the router 71, the path 81 and the path 82 are in contention. At this time, the system checks the link or router used only for the path having the same receiving node as the transmission source from the result of the traceroute command, and the path 81 and the path 82 are between the receiving node 4-1 and the router 71. Detects contention on the link. Here, the system calculates the ratio of the delay time in the link between the receiving node 4-1 and the router 71 having a conflicting path to the delay times of the competing paths 81 and 82, and determines the predetermined time. If it is larger than the threshold ratio, the path is changed.

  The path is changed as follows. For example, the proportion of the delay time in the link between the receiving node 4-1 and the router 71 with respect to the path 81 and the path 82 both exceeds the threshold, and the path 82 is related to the delay time of the non-conflicting part. Is shorter, the use of the path 81 is stopped, and instead the data from the receiving node 4-1 is changed to be transmitted to the receiving node 4-3 through the path 83.

  This change can reduce the number of competing paths by one instead of increasing the number of node hops in the distribution tree by one.

In the above system, in order to prevent an excessive increase in the number of node hops in the distribution tree, the above change is performed only once for each competing portion of the distribution tree.
Murakami Keiji, Sato Fumiaki, Application Level Multicast Tree Dynamic Construction Method Considering Link Load and Tree Depth, Information Processing Society of Japan Research Report, November 4, 2004, 2004-DPS-120, Vol. 107, 1-6

  Here, a problem in the conventional data distribution by ALM is that the distribution performance is unexpectedly deteriorated because the physical configuration of the network cannot be grasped. In other words, the purpose of problem solving in ALM is to improve the delivery performance of the delivery tree when contention of multiple delivery paths on a specific link, performance degradation of other delivery paths, etc. occur. It is necessary to reconfigure the distribution tree from the viewpoint of shortening the time until distribution is completed for all nodes.

  However, the reconfiguration of the distribution tree in the prior art taking the technique described in Non-Patent Document 1 as an example is based on the physical configuration of the network in order to avoid an extreme load on the link on a specific physical configuration. The path of the distribution path is detected, and the conflict of the distribution path in the link is detected, and only one of the conflicting distribution paths is locally changed at most once for the link detected as the race condition. Therefore, it cannot be said that the time until distribution to all nodes is completed is improved.

  Therefore, the present invention has been made in view of the above problems, and a distribution tree configuration device that can reconfigure a distribution tree into a distribution tree with high distribution efficiency according to the situation when performance degradation of the distribution path occurs, It is an object of the present invention to provide a technique related to a distribution tree configuration method and a distribution tree configuration program.

In order to achieve the above object, the present invention provides a distribution tree configuration device that configures a distribution tree that distributes data by multicast, based on a configuration instruction that configures the distribution tree, and based on a throughput of a distribution path. A distribution tree creating unit that configures a distribution tree; an information analyzing unit that repeatedly gives the configuration instruction to the distribution tree creating unit;
It is characterized by providing. With this configuration, a distribution tree corresponding to the network status is configured.

  According to the present invention, since the distribution tree configuration is repeatedly performed using the distribution path throughput measurement value, the network status can be improved against the performance degradation of the distribution path without knowing the physical configuration of the network. It is possible to configure a distribution tree with higher distribution efficiency according to the above. Accordingly, it is possible to shorten the time until distribution to all nodes is completed.

(First embodiment)
Next, a first embodiment for carrying out the present invention will be described in detail with reference to the drawings.

(Configuration of the first embodiment)
FIG. 1 is a block diagram showing the configuration of the first embodiment for carrying out the present invention.

  Referring to FIG. 1, in the first embodiment of the present invention, a transmission server 1 that performs data transmission and distribution tree management, and a reception node 4 that performs data transmission / reception and notification of measurement results to the transmission server 1 are described. And 6 and the network 5.

  The transmission server 1 includes a command input unit 11 for inputting a multicast distribution command and initial values of parameters, a data transmission unit 12 for transmitting data along the distribution tree, and a data transmission throughput (hereinafter referred to as a throughput measurement value). ) And is stored in the throughput measurement value storage unit 32 of the storage device 3, the tree management device 2 that operates under program control and manages the distribution tree, and the distribution tree and the reconfiguration of the distribution tree And a storage device 3 for storing data.

  The tree management device 2 includes a distribution tree creation unit 21, an information collection unit 22, and an information analysis unit 23.

  When a distribution tree creation command is given from the command input unit 11, the distribution tree creation unit 21 reads an estimated value (hereinafter referred to as a throughput estimation value) of a throughput measurement value of each path from the throughput estimated value storage unit 33. Based on the above, a distribution tree is created and stored in the distribution tree storage unit 31.

  The information collecting unit 22 is notified of the throughput measurement value and the received data amount from the information communication unit 45 of the receiving node 4, stores the throughput measurement value in the throughput measurement value storage unit 32, and has received the received data amount. The data amount is stored in the data amount storage unit 35.

  The information analysis means 23 starts reconfiguring the throughput measurement value and distribution tree of each path used for distribution (hereinafter referred to as distribution path) from the throughput measurement value storage unit 32 and the reconstruction threshold storage unit 34. Threshold values (hereinafter referred to as “reconstruction threshold values”) that are conditions to be read are read, performance degradation of the distribution path is investigated based on these, and it is determined whether or not to reconstruct the distribution tree. If it is determined that the distribution tree is to be reconfigured, a distribution tree reconfiguration command is given to the distribution tree creation means 21.

  In the first embodiment for carrying out the present invention, the information analysis unit 23 compares the throughput measurement value with the size of the reconfiguration threshold to investigate the performance degradation of the distribution path. When the measured value is less than the reconstruction threshold, it is determined that reconstruction is performed. Note that the criterion for determining that the reconstruction is performed is not limited to the case where the measured throughput value is less than the reconstruction threshold value, but may be other cases such as the case where the measured throughput value is less than or equal to the reconstruction threshold value.

  The storage device 3 includes a distribution tree storage unit 31 that stores the distribution tree created by the distribution tree creation unit 21, a throughput measurement value storage unit 32 that stores the throughput measurement value measured by the throughput measurement unit 13, and throughput estimation. A throughput estimation value storage unit 33 for storing values, a reconstruction threshold storage unit 34 for storing reconstruction thresholds, and a received data amount storage unit 35 for storing received data amounts notified from the information collecting means 22 I have.

  FIG. 2 is a block diagram illustrating a hardware configuration of the transmission server 1 according to the first embodiment of this invention.

  Here, the hardware configuration of the transmission server 1 will be described.

  FIG. 2 is a block diagram showing a hardware configuration of transmission server 1 according to the present embodiment.

  Referring to FIG. 2, the transmission server 1 according to the present invention can be realized by a hardware configuration similar to that of a general computer device, and includes a main memory such as a CPU (Central Processing Unit) 91 and a RAM (Random Access Memory). A main storage unit 92 used for a data work area and a temporary data save area, a communication control unit 93 (data transmission unit 12, throughput measurement unit 13) for transmitting and receiving data via the network 5, a liquid crystal display, Presentation unit 94 such as a printer or speaker, input unit 95 (command input unit 11) such as a keyboard or mouse, interface unit 96 connected to peripheral devices to transmit / receive data, ROM (Read Only Memory), magnetic disk, semiconductor Memory etc. Composed of a volatile memory hard disk device in which the auxiliary storage unit 97 (storage device 3), and a system bus 98 for connecting the above components of the information processing apparatus to each other.

  The transmission server 1 according to the present invention is implemented by mounting a circuit component made of a hardware component such as an LSI (Large Scale Integration) in which a program for realizing such a function is incorporated in the transmission server 1. As a matter of course, it can be realized in software by executing a program for providing each function of each component described above by the CPU 91 on the computer processing apparatus.

  That is, the CPU 91 loads the program stored in the auxiliary storage unit 97 to the main storage unit 92 and executes the program, thereby controlling the operation of the transmission server 1 to realize each function described above in software. Is possible.

  The reception node 4 includes a data transmission unit 41 that transmits data along the distribution tree, a data reception unit 42 that receives data, a throughput measurement unit 43 that measures a throughput measurement value that is a data transmission throughput, A received data amount measuring unit 44 that measures the amount of data, and an information contact unit 45 that communicates information indicating completion of distribution to the information collecting unit 22 via the network 5 are provided.

  The information communication unit 45 reads the throughput measurement value measured by the throughput measurement unit 43 and the received data amount measured by the received data amount measurement unit 44, and these are read by the information collection unit of the tree management device 2 of the transmission server 1. 22 is notified.

  In the receiving node 4, when the data receiving unit 42 receives data, the data transmitting unit 41 transmits data to the next receiving node 6.

(Operation of the first embodiment)
Next, the operation of the first embodiment for carrying out the present invention will be described in detail.

  First, the distribution tree creation process and the distribution tree reconfiguration process performed mainly by the distribution tree creation unit 21 of the transmission server 1 will be described with reference to FIGS.

  3 to 5 are flowcharts showing the operation of the transmission server according to the first embodiment of this invention.

(Delivery tree creation processing in the first embodiment)
First, an example (hereinafter referred to as SW method) of the distribution tree creation processing in the first embodiment for carrying out the present invention will be described with reference to FIG.

  The distribution tree creation means 21 adds the transmission server 1 to the distribution tree (step S101). At this point, neither the node nor the path is included in the distribution tree yet.

  Next, the distribution tree creating means 21 determines whether or not there is a node not included in the distribution tree (step S102). If there is a node not included in the distribution tree, the transmission tree included in the distribution tree and The throughput estimation values of all paths that have the node as the transmission source and the node that is not included in the distribution tree as the destination are read from the throughput estimation value storage unit 33 (step S103).

  Then, the distribution tree creating means 21 compares the read throughput estimated values, and adds the path having the largest throughput estimated value and the destination node of the path to the distribution tree (step S104). When there are a plurality of paths having the largest throughput estimation value, for example, the distribution tree creation means 21 selects one path at random from among them, and adds the path and the destination node of the path to the distribution tree. .

  Thereafter, the procedure from step S102 to step S104 is repeated until no nodes are included in the distribution tree (step S102), and a distribution tree is created.

  The distribution tree creation method is not limited to the SW method, and other methods such as a method of randomly adding a path to the distribution tree may be used.

  Also, the outline of the distribution tree reconfiguration process will be described with reference to FIG. 4. When the distribution tree creation means 21 is given a distribution tree reconfiguration command from the information analysis means 23, the distribution tree storage section 31 and the throughput From the measured value storage unit 32, the throughput estimated value storage unit 33, and the received data amount storage unit 35, the distribution tree up to that point, the measured throughput value of each path, the estimated throughput value, and the amount of data received by each node (hereinafter, (Referred to as received data amount) is read (step S201), and the distribution tree is reconstructed based on these data and stored in the distribution tree storage unit 31 (step S202).

(Redistribution tree reconfiguration process in the first embodiment)
Next, an example (hereinafter referred to as ECT method) of the distribution tree reconfiguration processing in the first embodiment for carrying out the present invention will be described in detail with reference to FIG.

  The distribution tree creation means 21 adds the transmission server 1 to the distribution tree (step S301). At this point, neither the node nor the path is included in the distribution tree yet.

  Next, the distribution tree creation means 21 determines whether or not there is a node not included in the distribution tree (step S302), and if there is a node not included in the distribution tree, the transmission tree included in the distribution tree and For all paths that have a node as a transmission source and a node that is not included in the distribution tree as a destination, the throughput measurement value of that path is sent from the throughput measurement value storage unit 32, and the throughput estimation value is sent to the throughput estimation value storage unit 33. , And the received data amount of the source and destination nodes is read from the received data amount storage unit 35 (step S303).

  Then, the distribution tree creation means 21 calculates a scheduled time for data transfer to the destination node (hereinafter referred to as a scheduled transfer completion time) from the read throughput measurement value, throughput estimation value, and received data amount. (Step S304). Here, the source candidate node is parent, the destination candidate node is child, and the throughput measurement value (throughput estimated value when there is no parent) of the parent and child is the th_child, which is included in the tree The measured value of the throughput of the path with the parent as the destination among the existing paths is the th_parent, the received data amount of the parent and child is DD_parent and DD_child, and the total data size of the distribution data Is content_size, the transfer completion scheduled time CT is calculated by the following formula, for example.

  The following expression is an expression for calculating the scheduled transfer completion time CT in the example of distribution tree creation processing (ECT method) according to the first embodiment of the present invention.

  Note that the scheduled transfer completion time is not limited to the method using the following formula, but may be calculated by another method.

  The distribution tree creation means 21 compares the calculated transfer completion scheduled times (step S305), and adds the path having the shortest transfer completion scheduled time and the destination node of the path to the distribution tree (step S306). When there are a plurality of paths having the shortest transfer completion scheduled time, for example, the distribution tree creation means 21 selects one path at random from them, and adds the path and the destination node of the path to the distribution tree. To do.

  Thereafter, the procedure from step S302 to step S306 is repeated until no nodes are included in the distribution tree (step S302), and a distribution tree is created.

  The distribution tree reconfiguration process is not limited to the ECT method, and may be performed using another method such as the SW method described above. Also, without reconfiguring the distribution tree from 1, the distribution tree can be reconfigured by selecting a distribution path as described above for a distribution tree partially configured by removing some paths and nodes. May be.

  Next, the distribution processing in the transmission server and the reception server according to the present embodiment will be described in detail with reference to FIGS.

  When the throughput estimation value and the reconstruction threshold value of each path are input to the command input unit 11, the command input unit 11 stores the throughput estimation value in the throughput estimation value storage unit 33 and uses the reconstruction threshold value as the reconstruction threshold value. The data is stored in the storage unit 34 (step A0 (initial setting step)).

  When a multicast distribution start command is input to the command input unit 11, the command input unit 11 gives a distribution tree creation command to the distribution tree creation unit 21. When a distribution tree creation command is given, the distribution tree creating means 21 reads the throughput estimated value of each path from the throughput estimated value storage unit 33, creates a distribution tree based on this, and stores it in the distribution tree storage unit 31. (Step A1 (distribution tree creation step)).

  The distribution tree in the first embodiment for carrying out the present invention is created by, for example, the aforementioned SW method. The distribution tree creation method is not limited to the SW method, and other methods such as a method of randomly adding a path to the distribution tree may be used.

  The data transmitting unit 12 reads the distribution tree from the distribution tree storage unit 31 and starts transmitting data to the data receiving unit 42 of the receiving node 4 along the distribution tree via the network 5. The reception node 4 that has received the data by the data reception means 42 starts data transmission to the reception node 6 that is the next destination by the data transmission means 41 (step A2 (data transmission instruction step)).

  The transmission server 1 and the reception server 4 determine whether or not the distribution is completed for all the nodes based on the information indicating the completion of distribution between the information collecting unit 22 and the information contact unit 45 (step A3 (data distribution) Completion check step)) When distribution is not completed for all nodes, the throughput measurement means 13 and 43 measure the throughput measurement value of the distribution path from the network 5 while distributing the data, and the received data The amount measuring means 44 measures the amount of received data. The information communication unit 45 notifies the information collection unit 22 of the throughput measurement value and the received data amount, and the information collection unit 22 stores the throughput measurement value in the throughput measurement value storage unit 32 and the received data. The amount is stored in the received data amount storage unit 35 (step A4 (throughput measurement instruction step)).

  When the throughput measurement value and the received data amount are measured, the information analysis unit 23 reads the throughput measurement value and the reconstruction threshold value of each distribution path from the throughput measurement value storage unit 32 and the reconstruction threshold storage unit 34 (step A5). (Reconstruction information extraction step)), the performance degradation of the distribution path is investigated based on the read throughput measurement value and the reconfiguration threshold, and it is determined whether or not the distribution tree is reconfigured (step A6 (reconfiguration determination) Step)).

  In the first embodiment for carrying out the present invention, the information analysis means 23 compares the throughput measurement value with the size of the reconstruction threshold. For example, if the throughput measurement value is less than the reconstruction threshold, In some cases, it is determined that reconfiguration is performed. It should be noted that the criterion for determining to perform reconstruction may be other cases as described above.

  If it is determined in step A6 that the distribution tree is to be reconfigured, the information analysis unit 23 gives a distribution tree reconfiguration command to the distribution tree creation unit 21, and the distribution tree creation unit 21 is given a distribution tree reconfiguration command. A distribution tree storage unit 31, a throughput measurement value storage unit 32, a throughput estimation value storage unit 33, and a received data amount storage unit 35, a current distribution tree, a throughput measurement value of each path, a throughput estimation value of each path, and The received data amount of the node is read, and the distribution tree is reconstructed based on these data and stored in the distribution tree storage unit 31 (step A1).

  The distribution tree in the first embodiment for carrying out the present invention is reconfigured until multicast distribution is completed, for example, by the ECT method described above. The distribution tree reconstruction processing is not limited to the ECT method, and may be performed using other methods such as the SW method described above. Also, without reconfiguring the distribution tree from 1, the distribution tree can be reconfigured by selecting a distribution path as described above for a distribution tree partially configured by removing some paths and nodes. May be.

  If it is determined by the information analysis means 23 that the distribution tree is not reconfigured in step A6, the process proceeds to step A2.

  Thereafter, step A1 to step A6 are repeated until distribution is completed for all nodes in step A3.

  On the other hand, when the distribution is completed for all the nodes in step A3, the multicast distribution is terminated.

(Effects of the first embodiment)
Next, the effect of the first embodiment for carrying out the present invention will be described.

  In the first embodiment of the present invention, reconfiguration is performed not only for the participation and withdrawal of a node and the occurrence of a failure, but also for performance degradation of the distribution path, and until the multicast distribution ends, Since the distribution path throughput measurement, performance analysis, and distribution tree reconfiguration are repeated, if the initially created distribution tree competes with a specific physical link, or if the network status changes during distribution, The created distribution tree can be switched to a distribution tree with higher distribution efficiency depending on the situation. For this reason, it is possible to shorten the time until distribution is completed to all nodes.

(Second Embodiment)
Next, a second embodiment for carrying out the present invention will be described in detail with reference to the drawings, focusing on differences from the first embodiment.

(Configuration of Second Embodiment)
FIG. 7 is a block diagram showing the configuration of the second exemplary embodiment for carrying out the present invention.

  Referring to FIG. 7, the storage device 3 according to the second embodiment of the present invention has a distribution path (with a degraded performance in the configuration of the storage device 3 according to the first embodiment shown in FIG. 1). Hereinafter, a performance degradation path storage unit 36 that stores performance degradation paths) and a performance degradation cause path storage unit 37 that stores distribution paths that cause performance degradation (hereinafter referred to as performance degradation cause paths) are newly provided. In addition, the information analysis unit 23 includes a reconstruction threshold value determination unit 231, a performance investigation unit 232, and a performance deterioration cause identification unit 233.

  The reconstruction threshold value determination unit 231 reads the throughput measurement value of each distribution path from the throughput measurement value storage unit 32 of the storage device 3, and when there is no throughput measurement value in the distribution path, the throughput of the storage device 3 is instead used. The throughput estimation value of the delivery path is read from the estimated value storage unit 33, the reconstruction threshold value of the delivery path is calculated based on the read throughput measurement value or throughput estimate value, and the calculated delivery path reconstruction threshold value is re-established. The configuration threshold value is stored in the configuration threshold value storage unit 34.

  In the second embodiment for carrying out the present invention, the reconfiguration threshold value of all distribution paths is set to the minimum value of the throughput measurement values of all distribution paths (throughput estimation values for distribution paths having no throughput measurement values). Set consistently.

  Note that the method for setting the reconfiguration threshold is not limited to the method of setting the reconfiguration threshold for all distribution paths to the minimum value of the throughput measurement values of all distribution paths, and the reconfiguration threshold for each distribution path is Other methods such as a method of individually setting N% (N is a positive real number) of the measured throughput value of the distribution path may be used.

  The performance investigation unit 232 reads the throughput measurement value and the reconstruction threshold of each distribution path from the throughput measurement value storage unit 32 and the reconstruction threshold storage unit 34, and the distribution path is determined based on the read throughput measurement value and the reconstruction threshold. It is determined whether or not the performance degradation path.

  If the performance investigation unit 232 determines that the path is a performance degradation path, the performance investigation unit 232 stores the distribution path in the performance degradation path storage unit 36 and gives a distribution tree reconfiguration command to the distribution tree creation unit 21 ′.

  In the second embodiment for carrying out the present invention, the performance investigation means 232 compares the throughput measurement value with the size of the reconstruction threshold. For example, if the throughput measurement value is less than the reconstruction threshold, In some cases, it is determined that the delivery path is a performance degradation path.

  Note that the criterion for determining a performance degradation path is not limited to the case where the measured throughput value is less than the reconstruction threshold value, but may be other cases such as when the measured throughput value is less than or equal to the reconstruction threshold value. .

  The performance degradation cause identification unit 233 reads all performance degradation paths from the performance degradation path storage unit 36, reads the throughput measurement values of all read performance degradation paths from the throughput measurement value storage unit 32, and converts them into the read throughput measurement values. The performance degradation cause path is identified based on the result, and the identified performance degradation cause path is stored in the performance degradation cause path storage unit 37.

  In the second embodiment for carrying out the present invention, the performance degradation cause path is identified by, for example, the smallest throughput measurement value in the performance degradation path, and the most number of node hops from the source among them. The performance degradation path having a small is used as the performance degradation cause path.

  Note that the method for identifying the performance degradation cause path is not limited to the above method, and other methods such as a method in which all performance degradation paths having the minimum throughput measurement values are used as the performance degradation cause paths may be used.

  Also, the distribution tree creation means 21 ′ in the second exemplary embodiment of the present invention, when a distribution tree reconfiguration command is given from the information analysis means 23, the distribution tree storage section 31, the throughput measurement value storage section 32, and the throughput estimation. In addition to reading the distribution tree so far, the throughput measurement value of each path, the estimated throughput value of each path, and the received data amount of each node from the value storage unit 33 and the received data amount storage unit 35, performance degradation It differs from the distribution tree creation means 21 in the first embodiment in that the performance degradation cause path is read from the cause path storage unit 37, the distribution tree is reconstructed based on these, and stored in the distribution tree storage unit 31. .

  Since other points are the same as those of the first embodiment, description thereof is omitted.

(Operation of Second Embodiment)
Next, with reference to FIG. 8 and FIG. 7, the distribution processing in the transmission server and the reception server of the present embodiment will be described in detail focusing on differences from the first embodiment.

  When the throughput estimation value of each path is input to the command input unit 11, the command input unit 11 stores the throughput estimation value in the throughput estimation value storage unit 33 (step A0).

  The operations of the command input means 11, the distribution tree storage unit 31, and the throughput estimated value storage unit 33 in the second embodiment shown in step A1 are the same as those operations in the first embodiment. Description is omitted.

  In addition, the operation of the distribution tree creating means 21 ′ in the second embodiment shown in step A1 is the same as that of the distribution tree creating means 21 ′ in the first embodiment in the first distribution tree creating process. The description is omitted because it is the same as the creation process.

  After step A1, the reconstruction threshold value determining unit 231 reads the throughput measurement value from the throughput measurement value storage unit 32 for each distribution path. The distribution path throughput estimated value is read from the throughput estimated value storage unit 33, the distribution path reconfiguration threshold is calculated based on the read throughput measurement value or throughput estimated value, and the calculated distribution path reconfiguration threshold is recalculated. The information is stored in the configuration threshold storage unit 34 (step A51 (reconfiguration threshold calculation step)).

  In the second embodiment for carrying out the present invention, the reconfiguration threshold value of all distribution paths is unified to the minimum value of the throughput measurement values of all distribution paths (throughput estimation values for distribution paths without throughput measurement values). And set it. As described above, another method may be used for setting the reconstruction threshold.

  Data transmitting means 12 and 41, data receiving means 42, throughput measuring means 13 and 43, received data amount measuring means 44, and information communication means 45 in the second embodiment shown in steps A2 to A4, Since the operations of the information collection unit 22, the distribution tree storage unit 31, the throughput measurement value storage unit 32, and the received data amount storage unit 35 are the same as those in the first embodiment, the description will be omitted. Omitted.

  When the throughput measurement value and the amount of received data are measured in step A4, the performance investigation unit 232 determines whether or not the above measurement has been made for all distribution paths (step A52 (distribution path investigation completion check step)). When the above measurement is not completed for all distribution paths, the throughput measurement value and the reconstruction threshold of each distribution path are read from the throughput measurement value storage unit 32 and the reconstruction threshold storage unit 34 (step A53 (distribution path information acquisition). Step)), comparing the throughput measurement value of each read distribution path with the reconfiguration threshold to determine whether the distribution path that has not been measured is a performance degradation path (step A54 (throughput measurement value evaluation step)) ).

  In the second embodiment for carrying out the present invention, the criterion for determining that the delivery path is a performance degradation path by the performance investigation unit 232 is, for example, when the throughput measurement value is less than the reconfiguration threshold. It is. Note that the criterion for determining that the path is a performance degradation path may be other cases as described above.

  When the performance investigation unit 232 determines in step A54 that the delivery path is not a performance degradation path (for example, when the throughput measurement value is equal to or greater than the reconfiguration threshold value), the performance investigation unit 232 performs no processing, Return to Step A52.

  On the other hand, in step A54, for example, when the throughput measurement value is less than the reconfiguration threshold value, the performance investigation unit 232 determines that the distribution path is a performance degradation path, and determines the distribution path as a performance degradation path storage unit. 36 (step A55 (path performance evaluation step)), and the step returns to step A52.

  Thereafter, step A52 to step A55 are repeated until all the distribution paths are examined in step A52 (until it is determined that the above measurement has been performed for all the distribution paths).

  On the other hand, when all the distribution paths have been checked in step A52 (when it is determined that the above measurement has been made for all distribution paths), the performance deterioration cause identifying unit 233 stores the performance deterioration paths in the performance deterioration path storage unit 36. It is determined whether there is a path with degraded performance by checking whether it is stored (step A57 (performance degradation path check step)). If no performance degradation path is stored, no processing is performed. Instead, the process proceeds to step A2.

  When the performance degradation path is stored in step A57, the performance degradation cause identification unit 233 reads all the performance degradation paths from the performance degradation path storage unit 36, and obtains the throughput measurement values of all the performance degradation paths read. The performance degradation cause path is read from the throughput measurement value storage unit 32, the performance degradation cause path is identified based on the read values, and the identified performance degradation cause path is stored in the performance degradation cause path storage unit 37 (step A58 (performance degradation cause identification step)).

  In the second embodiment for carrying out the present invention, in the performance degradation cause path identification, the performance degradation cause identification means 233 has the smallest throughput measurement value in the performance degradation path, and most of them. A performance degradation path with a small number of node hops from the source is identified as a performance degradation cause path. As described above, another method may be used to identify the performance degradation cause path.

  Next, when the information analysis means 23 determines whether or not the distribution tree is to be reconfigured by the performance investigation means 232 (step A6), and if it is determined that the distribution tree needs to be reconfigured, the performance investigation means 232 Then, a distribution tree restructuring instruction is given to the distribution tree creating means 21 ′, and when the distribution tree restructuring instruction is given, the distribution tree creating means 21 ′ delivers the distribution tree storage unit 31, the throughput measurement value storage unit 32, and the throughput estimated value. From the storage unit 33, the received data amount storage unit 35, and the performance degradation cause path storage unit 37, the current distribution tree, the measured throughput value of each path, the estimated throughput value of each path, the received data amount of each node, and the performance degradation The cause path is read, and the distribution tree is reconstructed based on the cause path and stored in the distribution tree storage unit 31 (step A1).

  The distribution tree according to the second embodiment for carrying out the present invention is reconfigured by selecting a performance degradation cause path by the ECT method described above, for example. The distribution tree reconfiguration method is not limited to the ECT method, and may be another method such as the SW method described above. Also, without reconfiguring the distribution tree from 1, the distribution tree can be reconfigured by selecting a distribution path as described above for a distribution tree partially configured by removing some paths and nodes. May be.

  On the other hand, if the information analysis means 23 determines that the performance investigation means 232 does not reconstruct the distribution tree in step A6, the process goes to step A2.

  Thereafter, step A1, step A51, step A2 to step A4, step A52 to step A58, and step A6 are repeated until distribution to all nodes is completed in step A3.

  On the other hand, when the distribution is completed for all the nodes in step A3, the multicast distribution is terminated.

(Effect of the second embodiment)
Next, the effect of the second embodiment for carrying out the present invention will be described.

  In the second embodiment of the present invention, a reconfiguration threshold is clearly provided for each distribution path, and a reconfiguration is performed by selecting a distribution path that causes distribution performance. Therefore, even in a network where paths with different usable bandwidths are mixed, it is possible to avoid degradation of distribution performance and unnecessary reconfiguration, and even reconfiguration to extra distribution paths. Can be avoided. As a result, it is possible to switch to a distribution tree with high distribution efficiency even in a small number of reconfiguration locations while reducing the reconstruction overhead.

(Third embodiment)
Next, a third embodiment for carrying out the present invention will be described in detail with reference to the drawings, focusing on differences from the first embodiment.

(Configuration of the third embodiment)
FIG. 9 is a block diagram showing the configuration of the third exemplary embodiment for carrying out the present invention.

  Referring to FIG. 9, the tree management apparatus 2 according to the third embodiment of the present invention is configured to estimate the throughput from the measured throughput value in the configuration of the tree management apparatus 2 according to the first embodiment shown in FIG. Is provided with throughput estimation means 24 for estimating.

  The throughput estimation unit 24 includes an initial value setting unit 241, an estimated value update unit 242, and an update condition evaluation unit 243.

  The initial value setting unit 241 sets initial values of throughput estimation values for all paths and stores them in the throughput estimation value storage unit 33.

  When the update condition evaluation unit 243 receives the distribution path throughput estimated value update command, the estimated value update unit 242 reads the distribution path throughput measurement value from the throughput measurement value storage unit 32 and based on the read throughput measurement value. Thus, the throughput estimated value of the delivery path is obtained and stored in the throughput estimated value storage unit 33.

  The update condition evaluation unit 243 reads the throughput measurement value of each distribution path from the throughput measurement value storage unit 32, reads the throughput estimation value of each distribution path from the throughput estimation value storage unit 33, and receives the received data amount of each node. Whether or not to update the estimated throughput value is read from the stored data amount storage unit 35.

  When the update condition evaluation unit 243 determines to update the throughput estimation value, the update condition evaluation unit 243 gives a throughput estimation value update command to the estimation value update unit 242.

  Further, the distribution tree creation means 21 "in the third exemplary embodiment of the present invention measures the throughput of each path from the throughput measurement value storage unit 32 when a distribution tree reconfiguration command is given from the information analysis means 23". It differs from the distribution tree creation means 21 in the first embodiment in that instead of reading a value, the throughput estimated value of each path is read from the throughput estimated value storage unit 33.

  Further, the information analysis means 23 "in the third embodiment of the present invention replaces reading the throughput measurement value of each distribution path from the throughput measurement value storage unit 32 when investigating the performance degradation of the distribution path. The difference from the information analysis unit 23 in the first embodiment is that the throughput estimation value of each distribution path is read from the throughput estimation value storage unit 33.

  In the third embodiment for carrying out the present invention, the information analyzing means 23 "compares the throughput estimated value with the size of the reconfiguration threshold and investigates the performance degradation of the distribution path. It is determined that the reconstruction is performed when the throughput estimation value is less than the reconstruction threshold, and the criterion for determining the reconstruction is not limited to the case where the throughput estimation value is less than the reconstruction threshold. Other cases such as a case where the measured throughput value is equal to or less than the reconstruction threshold value may be used.

  Since other points are the same as those of the first embodiment, description thereof is omitted.

(Operation of the third embodiment)
Next, the operation of the third embodiment for carrying out the present invention will be described in detail.

  First, the distribution tree reconfiguration process performed mainly by the distribution tree creation means 21 of the transmission server 1 will be described with reference to FIG.

  FIG. 10 is a flowchart illustrating the operation of the transmission server according to the third embodiment of this invention.

(Distribution Tree Reconfiguration Processing in the Third Embodiment)
An example (hereinafter referred to as ECT ′ method) of the distribution tree reconfiguration processing in the third embodiment for carrying out the present invention will be described in detail.

  First, the distribution tree creating means 21 "adds the transmission server 1 to the distribution tree (step S401). At this time, neither the node nor the path is included in the distribution tree.

  Next, the distribution tree creating means 21 "determines whether or not there is a node not included in the distribution tree (step S402), and if there is a node not included in the distribution tree, the transmission server included in the distribution tree. For all paths whose destination is a node not included in the distribution tree with the node as the transmission source, the throughput estimation value of the path is read from the throughput estimation value storage unit 33, and the source and destination node The received data amount is read from the received data amount storage unit 35 (step S403).

  Then, the distribution tree creation means 21 "calculates a scheduled time (hereinafter referred to as a scheduled transfer completion time) for completing the transfer of data to the destination node from the read throughput estimation value and the amount of received data (Step 5). S404) Here, the transmission source candidate node is parent, the destination candidate node is child, the throughput estimation value of the path having the parent as the transmission source and the child as the destination is ET_child, and the path is included in the path included in the tree. If the estimated throughput of the path with the parent as the destination is ET_parent, the received data amount of the parent and child is DD_parent and DD_child, and the total data size of the distribution data is content_size, the transfer completion scheduled time CT is As shown It is calculated with the following formula.

  The following expression is an expression for calculating the scheduled transfer completion time CT in the example of distribution tree creation processing (ECT 'method) according to the third embodiment of the present invention.

  Note that the scheduled transfer completion time is not limited to the method using the following formula, but may be calculated by another method.

  The distribution tree creating means 21 "compares the calculated transfer completion scheduled times (step S405), and adds the path having the shortest transfer completion scheduled time and the destination node of the path to the distribution tree (step S406). When there are a plurality of paths having the shortest transfer completion scheduled time, for example, the distribution tree creating means 21 "selects one path at random from the paths, and sets the path and the destination node of the path in the distribution tree. to add.

  Thereafter, the procedure from step S402 to step S406 is repeated until no nodes are included in the distribution tree (step S402), and a distribution tree is created.

  The distribution tree reconstruction process is not limited to the ECT method, and may be performed using another method such as the SW method described above. Also, without reconfiguring the distribution tree from 1, the distribution tree can be reconfigured by selecting a distribution path as described above for a distribution tree partially configured by removing some paths and nodes. May be.

  Next, with reference to FIG. 11 and FIG. 9, the distribution processing in the transmission server and the reception server of the present embodiment will be described in detail.

  When the reconfiguration threshold value of each path is input to the command input unit 11, the command input unit 11 stores the reconfiguration threshold value in the reconfiguration threshold storage unit 34 (step A0).

  When a multicast distribution start command is given to the command input unit 11, the initial value setting unit 241 sets initial values of throughput estimation values of all paths, and sets the initial values of the throughput estimation values set as the throughput estimation value storage unit 33. (Step A61 (throughput estimated value initial value setting step)).

  In the third embodiment for carrying out the present invention, the initial value setting means 241 uses the smaller of the access line upstream band of the transmission source node and the access line downstream band of the destination node for each path as the throughput estimated value. Set as the initial value of.

  Command input means 11, data transmission means 12 and 41, data reception means 42, throughput measurement means 13 and 43, and received data amount measurement means 44 in the third embodiment shown in steps A1 to A4, The operation of the information communication unit 45, the information collection unit 22, the distribution tree storage unit 31, the throughput measurement value storage unit 32, the throughput estimated value storage unit 33, and the received data amount storage unit 35 is the first operation. Since these operations are the same as those in the embodiment, description thereof is omitted.

  Further, the operation of the distribution tree creating means 21 "in the third embodiment shown in step A1 is the same as the operation of the distribution tree creating means 21 in the first embodiment when creating the first tree. Therefore, explanation is omitted.

  When the throughput is measured in step A4, the throughput estimation unit 24 starts updating the throughput estimation value of the distribution path at regular time intervals (step A62 (throughput estimation value update start step)).

  Next, the update condition evaluation unit 243 determines whether or not the measurement has been performed for all the distribution paths (step A63 (distribution path survey completion check step)), and the measurement has not been completed for all the distribution paths. The throughput measurement value of each distribution path is read from the throughput measurement value storage unit 32, the throughput estimation value of each distribution path is read from the throughput estimation value storage unit 33, and the received data amount of each node is received. And using these, it is determined whether or not to update the throughput estimation value (step A64 (distribution path information acquisition step)).

  When it is determined not to update the throughput estimation value for the delivery path (step A65 (throughput estimation value update condition evaluation step)), the update condition evaluation unit 243 does not perform any processing and proceeds to step A63. Return.

  In the third embodiment for carrying out the present invention, the condition for the throughput estimation means 24 to update the throughput estimation value is that the throughput measurement value is greater than the throughput estimation value or the throughput measurement value in each distribution path. Is smaller than the estimated throughput value and the received data amount of the transmission source node of the distribution path is larger than the received data amount of the destination node, or the measured throughput value is smaller than the estimated throughput value and the transmission path source This is a case where the received data amount of the node is equal to the received data amount of the destination node, and the throughput measurement value of the path immediately upstream of the distribution path is larger than the throughput measurement value of the distribution path.

  When it is determined to update the throughput estimation value for the delivery path (step A65), the update condition evaluation unit 243 gives a throughput estimation value update command to the estimated value update unit 242, and the estimated value update unit When a throughput estimated value update command is given, 242 reads the throughput measurement value of the delivery path from the throughput measurement value storage unit 32, obtains the throughput estimate value of the delivery path based on the read throughput measurement value, and estimates the throughput. The value is stored in the value storage unit 33 (step A66 (throughput estimated value calculation step)). Thereafter, the step returns to step A63.

  In the third embodiment for carrying out the present invention, the estimated value updating means 242 rewrites and updates the throughput estimated value to the value of the throughput measured value.

  Thereafter, steps A63 to A66 are repeated until all distribution paths have been checked in step A63 (until it is determined that the above measurement has been made for all distribution paths).

  On the other hand, when all the distribution paths have been checked in step A63 (when it is determined that the above measurement has been performed for all distribution paths), the information analysis unit 23 "uses the throughput estimated value storage unit 33 and the reconstruction threshold storage unit. 34 is used to read the throughput estimation value and the reconstruction threshold of each distribution path from step 34 (step A5), investigate performance degradation of the distribution path based on the read throughput estimation value and the reconstruction threshold, and determine whether to reconstruct the distribution tree Is determined (step A6).

  In the third embodiment for carrying out the present invention, the information analyzing means 23 "compares the throughput estimated value with the size of the reconstruction threshold. For example, the throughput estimated value is less than the reconstruction threshold. In this case, it is determined that the reconstruction is performed, and the criterion for determining that the reconstruction is performed may be other cases as described above.

  When it is determined in step A6 that the distribution tree is to be reconfigured, the information analysis unit 23 "gives a distribution tree reconfiguration command to the distribution tree generation unit 21", and the distribution tree generation unit 21 " When an instruction is given, the current distribution tree, the throughput estimation value of each path, and the received data amount of each node are read from the distribution tree storage unit 31, the throughput estimated value storage unit 33, and the received data amount storage unit 35. Based on the above, the distribution tree is reconstructed and stored in the distribution tree storage unit 31 (step A1).

  The distribution tree in the third embodiment for carrying out the present invention is reconfigured by, for example, the ECT ′ method described above. The distribution tree reconfiguration method is not limited to the ECT ′ method, and may be another method such as the SW method described above. Also, without reconfiguring the distribution tree from 1, the distribution tree can be reconfigured by selecting a distribution path as described above for a distribution tree partially configured by removing some paths and nodes. May be.

  On the other hand, if it is determined by the information analysis means 23 "that the distribution tree is not reconfigured in step A6, the process proceeds to step A2.

  Thereafter, step A1 to step A4, step A62 to step A66, and step A5 to step A6 are repeated until distribution to all nodes is completed in step A3.

  On the other hand, when the distribution is completed for all the nodes in step A3, the multicast distribution is terminated.

(Effect of the third embodiment)
Next, the effect of the third embodiment for carrying out the present invention will be described.

  In the third embodiment of the present invention, since the throughput estimation value is first updated in consideration of the influence of the upstream path, the throughput of each distribution path itself can be estimated in the network situation at that time. .

  In addition, the throughput estimation process is performed many times in conjunction with the loop procedure in the distribution tree reconfiguration process, so there is no need to scrutinize the network when creating the first distribution tree. Without being used, the throughput of various paths in the network can be estimated following changes in the network state, and the accuracy of the throughput estimation value in the distribution path can be improved according to the distribution of the distribution path.

  And by reconfiguring the distribution tree in conjunction with the estimation information in the throughput estimation process, distribution paths with high distribution efficiency can be distributed according to the network performance and status, and the loop procedure is repeated. Thus, these synergistic effects can be further improved, and the delivery completion time can be further shortened.

  The present invention provides a device for simultaneously distributing contents such as data files, video streaming, and security patches to one or a plurality of nodes via the Internet, and a program for realizing the functions of the device by a computer Can be applied. In addition, the function to configure and reconfigure the distribution tree provides multiple distribution trees with each node as the distribution source, and uses them to distribute participants' comments from multiple locations to other locations. It can also be applied to uses such as meetings.

It is a block diagram which shows the structure of the 1st Embodiment of this invention. It is a block diagram which shows the hardware constitutions of the transmission server of the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the transmission server of the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the transmission server of the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the transmission server of the 1st Embodiment of this invention. It is a flowchart which shows the operation | movement of the 1st Embodiment of this invention. It is a block diagram which shows the structure of the 2nd Embodiment of this invention. It is a flowchart which shows the operation | movement of the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the 3rd Embodiment of this invention. It is a flowchart which shows operation | movement of the transmission server of the 3rd Embodiment of this invention. It is a flowchart which shows the operation | movement of the 3rd Embodiment of this invention. It is a block diagram which shows the structure of a general application level multicast. It is a block diagram which shows the structure of the logical tree of a general application level multicast. It is a figure for demonstrating the background art of this invention.

Explanation of symbols

1: Transmission server 11: Command input means 12: Data transmission means 13, Z13: Throughput measurement means 2: Tree management device 21, 21 ', 21 ", Z21: Distribution tree creation means 22, Z22: Information collection means 23, 23 ": Information analysis means 231: Reconfiguration threshold determination means 232: Performance investigation means 233: Performance degradation cause identification means 24: Throughput estimation means 241: Initial value setting means 242: Estimated value update means 243: Update condition evaluation means 3: Storage Apparatus 31: Distribution tree storage unit 32: Throughput measurement value storage unit Z32: Network information storage unit 33: Throughput estimated value storage unit 34: Reconfiguration threshold storage unit 35: Received data amount storage unit 36: Performance degradation path storage unit 37 : Performance degradation cause path storage unit 4, 4-1 to 4-X: Receiving node 41: Data transmission means 2: data receiving means 43, Z43: Throughput measurement unit 44, Z44: received data amount measuring means 45, Z45: us means 5: Network 6: the receiving node 71: Router 81 to 83: Path 91: CPU
92: Main storage unit 93: Communication control unit 94: Presentation unit 95: Input unit 96: Interface unit 97: Auxiliary storage unit 98: System bus

Claims (34)

A distribution tree configuration device that configures a distribution tree that distributes data by multicast,
A distribution tree creating unit configured to configure the distribution tree based on a performance of a distribution path based on a configuration command configuring the distribution tree;
Information analysis means for repeatedly giving the configuration command to the distribution tree creation means;
A distribution tree construction device comprising: A throughput measurement value storage unit that stores throughput measurement values as the performance of the distribution path is provided.
2. The distribution tree according to claim 1, wherein the distribution tree creation unit reads the throughput measurement value from the throughput measurement value storage unit when the configuration instruction is given, and configures the distribution tree. Configuration equipment. A distribution tree storage unit for storing the distribution tree configured by the distribution tree creating means;
3. The distribution tree configuration device according to claim 1, wherein the distribution tree creation unit configures the distribution tree based on the throughput and stores the distribution tree in the distribution tree storage unit. It further comprises a configuration threshold value storage unit for storing a predetermined configuration threshold value,
The information analysis unit reads the configuration threshold value from the configuration threshold value storage unit, and determines whether to give the configuration command based on the throughput and the read configuration threshold value. Item 4. The distribution tree configuration device according to any one of items 3 to 4. The distribution tree creation means reads the first distribution tree from the distribution tree storage unit when the configuration command is given, and receives the received data amount, the throughput, and the first distribution received from each receiving node. 5. The distribution tree configuration device according to claim 3, wherein a second distribution tree is configured based on the tree, and the configured second distribution tree is stored in the distribution tree storage unit. A distribution tree configuration device that configures a distribution tree that distributes data by multicast,
A throughput measurement value storage unit for storing the measured throughput of the delivery path as a throughput measurement value;
A distribution tree storage unit for storing the configured distribution tree;
A delivery tree creating means constituting the delivery tree;
Information analysis means for repeatedly giving the distribution tree creating means a configuration command that constitutes the distribution tree;
A performance deterioration cause path storage unit for storing a performance deterioration cause path determined by the information analysis means,
The distribution tree creation means includes:
When the configuration instruction is given, the throughput measurement value is read from the throughput measurement value storage unit, the first distribution tree is read from the distribution tree storage unit, and the performance deterioration cause is read from the performance deterioration cause path storage unit A distribution tree configured to read a path, configure a second distribution tree based on the read throughput measurement value, the first distribution tree, and the performance degradation cause path, and store the second distribution tree in the distribution tree storage unit; Configuration equipment. A configuration threshold storage unit that stores the configuration threshold;
The information analysis means includes
For each distribution path, the throughput measurement value is read from the throughput measurement value storage unit, the configuration threshold value is read from the configuration threshold value storage unit, and the distribution path is compared by comparing the read throughput measurement value and the configuration threshold value. A performance investigation means for determining whether or not the performance degradation path, and storing the performance degradation path in the performance degradation path storage unit;
Read a set of performance degradation paths stored in the performance degradation path storage unit, read a throughput measurement value for each performance degradation path from the throughput measurement value storage unit, and based on the read throughput measurement value for each performance degradation path The performance degradation cause identifying means for determining a performance degradation cause path from the set of performance degradation paths and storing the determined performance degradation cause path in the performance degradation cause path storage unit. Distribution tree configuration device. Configuration threshold value determining means for reading the throughput measurement value from the throughput measurement value storage unit, calculating a configuration threshold value based on the read throughput measurement value, and storing the configuration threshold value in the configuration threshold value storage unit The distribution tree construction device according to claim 7. A received data amount storage unit that stores the amount of data received by each receiving node as a received data amount;
The distribution tree creation means reads the throughput measurement value from the throughput measurement value storage unit, reads the received data amount from the received data amount storage unit when the configuration instruction is given, and Read the first distribution tree from the storage unit, read the performance deterioration cause path from the performance deterioration cause path storage unit, read the throughput measurement value, the received data amount, the first distribution tree and the performance deterioration 9. The system according to claim 6, wherein a second distribution tree is configured based on a cause path, and the configured second distribution tree is stored in the distribution tree storage unit. Distribution tree configuration device. A distribution tree configuration device that configures a distribution tree that distributes data by multicast,
A throughput estimated value storage unit for storing the estimated throughput of the delivery path as a throughput estimated value;
A distribution tree storage unit for storing the configured distribution tree;
A delivery tree creating means constituting the delivery tree;
Information analysis means for repeatedly giving the distribution tree creating means a configuration command that constitutes the distribution tree;
Throughput estimation means for estimating the throughput of the distribution path at regular intervals and storing it in the throughput estimated value storage unit,
The distribution tree creation means includes:
When the configuration instruction is given, the throughput estimation value is read from the throughput estimation value storage unit, the first distribution tree is read from the distribution tree storage unit, and the read throughput estimation value and the first distribution are read 2. A distribution tree construction device comprising a second distribution tree based on a tree and storing the constructed second distribution tree in the distribution tree storage unit. A throughput measurement value storage unit that stores the measured throughput of the delivery path as a throughput measurement value;
The throughput estimation means includes
When a throughput estimated value update command is given, the throughput measured value is read from the throughput measured value storage unit, the throughput estimated value is obtained based on the read throughput measured value, and the obtained throughput estimated value is Estimated value updating means for storing in the throughput estimated value storage unit;
Reading the throughput measurement value from the throughput measurement value storage unit, reading the throughput estimation value from the throughput estimation value storage unit, the received data amount received by each receiving node, the read throughput measurement value and the throughput estimation value The distribution tree construction device according to claim 10, further comprising: an update condition evaluation unit that determines whether or not to give the throughput estimated value update command to the estimated value update unit on the basis of the above. It further comprises a configuration threshold value storage unit for storing a predetermined configuration threshold value,
The information analysis means reads the throughput estimation value from the throughput estimation value storage unit, reads the configuration threshold value from the configuration threshold value storage unit, and gives the configuration command based on the read throughput estimation value and the configuration threshold value The distribution tree construction device according to claim 10 or 11, wherein it is determined whether or not. A received data amount storage unit that stores the amount of data received by each receiving node as a received data amount;
The distribution tree creation means reads the throughput estimated value from the throughput estimated value storage unit, reads the received data amount from the received data amount storage unit when the configuration command is given, and The first distribution tree is read from the storage unit, the second distribution tree is configured by configuring the second distribution tree by the read throughput estimation value, the received data amount, and the first distribution tree. The distribution tree configuration device according to claim 10, wherein the distribution tree storage unit stores the distribution tree in the distribution tree storage unit. 14. The distribution tree configuration device according to claim 1, wherein the information analysis unit repeats the operation until distribution of data by multicast is completed. A distribution tree configuration method for configuring a distribution tree for distributing data by multicast,
A distribution tree creation step of configuring the distribution tree based on performance of a distribution path based on a configuration instruction configuring the distribution tree;
An information analysis step of repeatedly giving the configuration command to the distribution tree creation step;
A distribution tree construction method characterized by comprising: A throughput measurement value storing step for storing a throughput measurement value as the performance of the delivery path is provided.
16. The distribution tree creation step of configuring the distribution tree by reading the throughput measurement value stored in the throughput measurement value storage step when the configuration command is given in the distribution tree creation step. Delivery tree construction method. In the distribution tree creation step, a second distribution tree is configured based on the first distribution tree stored in the distribution tree storage unit and the throughput, and the configured second distribution tree is used as the first distribution tree. 17. The distribution tree construction method according to claim 15, wherein the distribution tree storage unit is repeatedly stored instead of the distribution tree storage unit. In the distribution tree creation step, when the configuration command is given, the first distribution tree is read from the distribution tree storage unit, the received data amount received by each receiving node, the throughput, and the read first A second distribution tree is configured on the basis of the distribution tree, and the configured second distribution tree is repeatedly stored in the distribution tree storage unit instead of the first distribution tree. The delivery tree construction method according to claim 15 or claim 16. A distribution tree configuration method for configuring a distribution tree for distributing data by multicast,
An information analysis step of determining whether to configure the distribution tree based on a distribution path throughput and a predetermined configuration threshold;
A distribution tree creation step of configuring a second distribution tree based on the first distribution tree and the throughput stored in the distribution tree storage unit when it is determined to perform the configuration in the information analysis step;
A distribution tree storing step of storing the second distribution tree in the distribution tree storage unit instead of the first distribution tree when the second distribution tree is configured; Tree construction method. A distribution tree configuration method for configuring a distribution tree for distributing data by multicast,
Determining a configuration threshold for each delivery path;
A performance degradation path storage step for storing a delivery path whose throughput is less than the configuration threshold as a performance degradation path;
Identifying a performance degradation cause path based on the performance degradation path, and storing a performance degradation cause path storing the identified performance degradation cause path;
An information analysis step of checking whether there is a performance degradation path stored by the performance degradation path storage step and determining whether to give a configuration instruction that constitutes the distribution tree;
When it is determined in the information analysis step that the configuration instruction is given, a second distribution tree is determined based on the throughput, the first distribution tree stored in the distribution tree storage unit, and the identified performance degradation cause path. A distribution tree creation step to configure;
A distribution tree storing step of storing the second distribution tree in the distribution tree storage unit instead of the first distribution tree when the second distribution tree is configured; Tree construction method. A distribution tree configuration method for configuring a distribution tree for distributing data by multicast,
Determining a configuration threshold for each delivery path;
A performance degradation path storage step for storing a delivery path whose throughput is less than the configuration threshold as a performance degradation path;
Identifying a performance degradation cause path based on the performance degradation path, and storing a performance degradation cause path storing the identified performance degradation cause path;
An information analysis step of examining whether there is a performance degradation path stored by the performance degradation path storage step and determining whether to give a configuration command that constitutes the distribution tree;
When it is determined in the information analysis step that the configuration instruction is given, a second distribution tree is determined based on the throughput, the first distribution tree stored in the distribution tree storage unit, and the identified performance degradation cause path. A distribution tree creation step to configure;
In the distribution tree creation step, when the configuration instruction is given, the first distribution tree stored in the distribution tree storage unit is read, and the performance deterioration cause path stored in the performance deterioration cause path storage step is read out. The second distribution tree configured by configuring the second distribution tree based on the read, the received data amount received by each receiving node, the throughput, the read first distribution tree, and the performance degradation cause path A distribution tree storing step of storing the distribution tree storage step in the distribution tree storage unit instead of the first distribution tree. A distribution tree configuration method for configuring a distribution tree for distributing data by multicast,
A throughput estimation step of storing the throughput estimation value generated by estimating the throughput of the distribution path in the throughput estimation value storage unit;
An information analysis step of determining whether to give a configuration command to configure the distribution tree using the throughput estimation value and a predetermined configuration threshold;
When it is determined that the configuration instruction is given in the information analysis step, a second distribution tree is configured by the first distribution tree stored in the distribution tree storage unit and the throughput estimation value stored in the throughput estimation step. A distribution tree creation step;
A distribution tree storing step of storing the second distribution tree in the distribution tree storage unit instead of the first distribution tree when the second distribution tree is configured; Tree construction method. A distribution tree configuration method for configuring a distribution tree for distributing data by multicast,
A throughput estimation step of storing the throughput estimation value generated by estimating the throughput of the distribution path in the throughput estimation value storage unit;
An information analysis step of determining whether to give a configuration command to configure the distribution tree using the throughput estimation value and a predetermined configuration threshold;
When it is determined that the configuration instruction is given in the information analysis step, a second distribution tree is configured by the first distribution tree stored in the distribution tree storage unit and the throughput estimation value stored in the throughput estimation step. A distribution tree creation step;
In the distribution tree creation means step, when the configuration command is given, the throughput estimation value is read from the throughput estimation value storage unit, the first distribution tree is read from the distribution tree storage unit, and each receiving node A second distribution tree is configured by the received received data amount, the read throughput estimation value, and the first distribution tree, and the configured second distribution tree is used instead of the first distribution tree. And a distribution tree storage step of storing in the distribution tree storage unit. In the throughput estimation step,
An update condition evaluation step for determining whether or not to update the throughput estimate based on the throughput estimate and the throughput of the distribution path;
When it is determined that the update is performed in the update condition evaluation step, a throughput estimation value is determined based on the throughput, and a throughput estimation value update step of storing the determined throughput estimation value in the throughput estimation value storage unit is repeated. 24. The distribution tree construction method according to claim 22 or 23. The distribution tree construction method according to any one of claims 15 to 24, wherein, in the information analysis step, the operation is repeated until distribution of data by multicast is completed. A distribution tree configuration program for configuring a distribution tree for distributing data by multicast,
A distribution tree creating process for configuring the distribution tree based on performance of a distribution path based on a configuration instruction for configuring the distribution tree;
An information analysis process that repeatedly gives the configuration command to the distribution tree creation process;
A distribution tree configuration program characterized by causing a computer to execute. Let the computer execute a throughput measurement value storage process that stores the throughput measurement value as the performance of the delivery path,
27. In the distribution tree creation process, when the configuration instruction is given, the throughput measurement value stored in the throughput measurement value storage process is read to configure the distribution tree. Distribution tree configuration program. In the distribution tree creation process, when the configuration command is given, the first distribution tree is read from the distribution tree storage unit, the received data amount received by each receiving node, the throughput, and the read first first A second distribution tree is configured based on the distribution tree, and distribution tree storage processing for storing the configured second distribution tree in the distribution tree storage unit instead of the first distribution tree is repeated. The distribution tree configuration program according to claim 26 or 27. 29. The information analysis process according to any one of claims 26 to 28, wherein in the information analysis process, it is determined whether or not the distribution tree is configured based on the throughput and a configuration threshold value, and the configuration command is given. The distribution tree configuration program described. Causing the computer to execute a configuration threshold determination process for calculating a configuration threshold after the configuration of the distribution tree based on the throughput before the configuration of the distribution tree;
In the information analysis process,
Based on the throughput and the configuration threshold, a performance investigation process for determining whether or not the performance of each delivery path has deteriorated;
27. The computer according to claim 26, further comprising: causing a computer to execute a performance deterioration cause identifying process for identifying a distribution path that causes performance deterioration based on the distribution path determined to have deteriorated performance and the throughput. 30. The distribution tree configuration program according to any one of items 29. An update condition evaluation process for determining whether or not to update the throughput estimation value based on the throughput estimation value generated by estimating the throughput and the throughput of the distribution path and the received data amount of each node;
Causing the computer to execute throughput estimation value update processing for updating the throughput estimation value based on the throughput;
The information analysis process includes
30. The distribution tree configuration program according to claim 26, wherein it is determined whether to configure a distribution tree based on the throughput estimation value and a configuration threshold, and the configuration command is given. . A distribution tree configuration program for configuring a distribution tree for distributing data by multicast,
Processing for determining a configuration threshold for each delivery path;
A performance degradation path storage process for storing a delivery path whose throughput is less than the configuration threshold as a performance degradation path;
A process for identifying a performance degradation cause path based on the performance degradation path; a performance degradation cause path storage process for storing the identified performance degradation cause path;
An information analysis process for examining whether there is a performance degradation path stored by the performance degradation path storage process and determining whether to give a configuration instruction that constitutes the distribution tree;
When it is determined that the configuration instruction is given in the information analysis process, a second distribution tree is determined based on the throughput, the first distribution tree stored in the distribution tree storage unit, and the identified performance degradation cause path. The distribution tree creation process to configure;
In the distribution tree creation process, when the configuration instruction is given, the first distribution tree stored in the distribution tree storage unit is read, and the performance deterioration cause path stored in the performance deterioration cause path storage process is read out. The second distribution tree configured by configuring the second distribution tree based on the read, the received data amount received by each receiving node, the throughput, the read first distribution tree, and the performance degradation cause path A distribution tree storage program for storing a distribution tree in the distribution tree storage unit instead of the first distribution tree. A distribution tree configuration program for configuring a distribution tree for distributing data by multicast,
Throughput estimation value storage processing for storing the throughput estimation value generated by estimating the throughput of the distribution path in the throughput estimation value storage unit;
An information analysis process for determining whether to give a configuration command to configure the distribution tree using the throughput estimation value and a predetermined configuration threshold;
When it is determined that the configuration instruction is given in the information analysis step, a second distribution tree is determined based on the first distribution tree stored in the distribution tree storage unit and the throughput estimation value stored in the throughput estimation step. The distribution tree creation process to configure;
In the distribution tree creation process, when the configuration command is given, the throughput estimated value is read from the throughput estimated value storage unit, the first distribution tree stored in the distribution tree storage unit is read, and each receiving node The second distribution tree is configured based on the received received data amount, the read throughput estimation value, and the first distribution tree, and the configured second distribution tree is the first distribution tree. A distribution tree configuration program that causes a computer to repeatedly execute a distribution tree storage process stored in the distribution tree storage unit instead of the above. The distribution tree configuration program according to any one of claims 26 to 33, wherein in the information analysis process, the operation is repeated until distribution of data by multicast is completed.

Images