JPH0884157A - Data summarizing system - Google Patents

Abstract

PURPOSE: To efficiently summarize data from terminals to a center. CONSTITUTION: The center and respective terminals operate as nodes of a tree and each node receives absolutely minimum information necessary for constructing a partial tree communication network following the node from a parent (hierarchical structure) node as communication network data and transmits information necessary for constructing partial tree communication networks following respective child (hierarchical structure) nodes to all child nodes of the node concerned as communication network data. The communication network data include three data consisting of a parameter for giving feature of the shape of a tree structure of the communication network, a list of terminal address to be the nodes of a partial tree following the node concerned and attribute information to be transmitted from a parent to a child such as 'what number-th child of a parent I am (in terms of a tree structure)' and 'the address of the parent node'. Since the data of all child nodes are simultaneously sent from the parent node directly connected to the center to the center, the center can efficiently summarize the data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for efficiently collecting data held by a large number of specified terminals registered in a center into a center via a public network.

[0002]

2. Description of the Related Art FIG. 2 shows a block diagram of a conventional data aggregation system. (2-1) is a center that aggregates data, and (2-2-1) to (2-2-N) are terminals that hold data to be aggregated.

The center (2-1) registers an address registration section (2-1) for registering a list of addresses to be polled.
1) and a registered address list are divided and polling data is distributed to each window as a polling data generation / distribution unit (2-1-2), and polling is performed from the center window based on the divided address list. The terminal data sequential reception units (2-1-3-1) to (2-1-3-m) that sequentially perform terminal data reception from each terminal and the terminal data collected at each window are collected into one. A data aggregating unit (2-1-4) for aggregating the collected data,
The data transmission unit (2-1-5) transmits the aggregated data to a specific address.

The list of addresses registered in the address registration unit (2-1-1) is divided into the number of windows in the polling data generation / distribution unit (2-1-2), and the terminal data sequential reception unit ( 2-1-3-1)
~ (2-1-3-m). The terminal data sequential reception units (2-1-3-1) to (2-1-3-m) make calls to each terminal based on the received polling data and collect data from each terminal. The collected data is collected by the data collecting unit (2-1-4).
The aggregated data is transmitted to a specific address via the data transmission unit (2-1-5).

The terminals (2-2-1) to (2-2-N) have the same structure. Terminal (2-2-1)
Is a terminal data storage unit (2-2-1-1) that retains and stores the terminal data input to the terminal, and terminal data transmission that transmits the stored terminal data to the center in response to polling from the center. The part (2-2-1-2).

The data stored in the terminal data storage section (2-2-1-1) is held until it is transmitted to the center window. When the terminal data transmission unit receives an incoming call by polling from the center window, the terminal data storage unit (2-2-
The data stored in 1-1) is transmitted to the center window.

[0007]

Each window can collect only one terminal data in one step. Therefore, in the conventional data aggregation system that collects terminal data by polling from the center window, the terminal data collection efficiency is as follows: terminal data m per step when the number of windows is m.
It becomes an individual. Therefore, even if there is a surplus in the data transfer capacity of the network, if the number of center windows is small, it becomes a bottleneck, and efficient data collection cannot be performed.

An object of the present invention is to efficiently collect data from terminals to a center.

[0009]

According to the present invention, a center,
Each terminal operates as a node of a tree, and each node receives from the parent node, as the communication network data, the minimum necessary information for constructing a subtree communication network after that node, and Information sufficient for constructing a subtree communication network after each child node is propagated as communication network data to all child nodes of the node. The contact network data includes parameters characterizing the shape of the tree structure of the contact network,
A list of terminal addresses that will be the nodes of the subtree after that node, and attribute information of the node to be transmitted from the parent to the child, such as "what is the child to the parent" and "address of the parent node". Include three.

[0010]

[Operation] By propagating the communication network data, a communication network having an endless tree structure with each terminal as a node is constructed. In the contact network data, the parameters that characterize the shape of the tree structure, the terminal address list, and the node attribute information are used as information for building the contact network.
In addition, the attribute information of the node is also used to determine which side of the parent and the child should call when collecting the terminal data through the communication network after the construction of the communication network, and to allow the parent and child to cooperate. When data is aggregated through this widespread communication network, the terminal data can be aggregated not only between the center and terminals but also between the terminals, and in one step, the terminal data that exceeds the number of center contact points Can be sucked up by a node.

[0011]

[Embodiment 1] FIG. 1 is a system configuration diagram for explaining an embodiment 1 of the present invention, in which (1-1) is a center for collecting data from each terminal through a public network, and (1-2-1) to (1-
2-N) is a group of terminals connected to the public network.

The center (1-1) has an address registration unit (1-1-1) for registering addresses and a tree structure parameter registration unit (1-1-for registering tree structure parameters that characterize the shape of the tree structure). 2), a contact network data generation and distribution unit that generates and distributes information necessary for configuring a contact network having a tree structure having each contact as a root from the registered address list and the tree structure parameter ( 1-1-3), a window node information management unit (1-1-4) that manages the attribute information of the node such as the address of the child node and the root of the tree as the node information, the address list and the tree structure. Communication network data sequential transmission unit (1-1-5) for sequentially transmitting communication network data including three parameters and node information to child nodes.
1) to (1-1-5-m) and a terminal data successive reception unit (1-1-6-) that sequentially receives terminal data from child nodes.
1) to (1-1-6-m) and a data aggregation unit (1-1-) that aggregates the terminal data received at each window in the center.
7) and an aggregated data transmission unit (1-1-8) that transmits aggregated data to a specific address.

The terminals (1-2-1) to (1-2-N) are
Are the same. The terminal (1-2-1) has a contact network data receiving unit (1-2-1-1) for receiving contact network data from the center window or terminal of the parent node, and a child node of the own node. A lower partial connection network data generation unit (1-2-1-2) that generates, as connection network data, information necessary for constructing a subtree having the child node as a root.
And a contact network data sequential transmission unit (1-2-1-3) for sequentially transmitting contact network data to each child node, an address of one's own child node, an address of one's own parent node, and what number is for the parent. Window node information management unit (1-2.
1-4), a terminal data sequential reception unit (1-2-1-5) that sequentially receives terminal data from child nodes, and an own terminal data storage unit (1-2 that stores data input from the outside.
-1-6), a data aggregating unit (1-2-1-7) that aggregates the terminal data received from the child node and the terminal data of its own terminal as one aggregated data, and the aggregated data to the parent node. Terminal data transmission unit (1-2
1-8).

A list of addresses registered in the address registration unit (1-1-1) and a tree structure parameter registration unit (1-
The tree structure parameter registered in 1-2) is transmitted to the communication network data generation / distribution unit (1-1-3). The communication network data generation / distribution unit (1-1-3) knows the number of child nodes that each window should have from the structure parameter, divides the address list into the total number of child nodes of all the windows, and the number of child nodes of each window of each window. Communication network data sequential transmission unit (1-1-5-1) to
Send to (1-1-5-m). The communication network data generation / distribution unit (1-1-3) also sends the tree structure parameters to the transmission units (1-1-5-1) to (1-1-5-m) in sequence. The sequential transmission unit (1-1-5-1) selects an appropriate one from the sent address list as a child node candidate and tries to connect to that address. (It is effective to select an address that is physically as close as possible as a child node candidate in order to reduce the communication cost.) If the connection is successful, send the address list of the nodes that make up the subtree rooted at that child node. . It also propagates the tree structure parameter to its child nodes. Further, the parent node transmits its own address which is the parent address, the number of the child the child is, the number of steps from the root of the tree, and the node information to be transmitted from the parent to the child node. A group of data sent to each of these child nodes will be called contact network data. Sequential transmission unit (1-1-5-1)
When the data transmission to the child node is completed, the address of the successfully connected child node is sent to the contact node information management unit (1-1-4) together with the information on the child node of the child node. Notice. If the attempt to connect to the child node candidate fails, the address is removed from the address list, another candidate is selected from the address list as the child node candidate, and the connection is tried again. This makes it possible to flexibly construct a tree-structured communication network by excluding terminals that are registered in the address list but cannot respond due to a failure or the like.

The contact network data receiving units (1-2-1-1) of the terminals (1-2-1) to (1-2-N) that receive an incoming call from the center window which is the upper node are the upper node. And receives the contact network data from the lower part contact network data generator (1-2-1-2). The same applies when an incoming call is received from a terminal as an upper node. The lower-order partial contact network data generation unit (1-2-1-2) knows the address of its own parent node and the number of the child of its own parent from the contact network data, and uses this information to determine its own terminal node. Notify the information management unit (1-2-1-4). Also, from the tree structure parameters in the connection network data and the number of steps from the root (window) of the own node, the number of child nodes that the own node should have,
After removing its own address from the address list, it is divided into the number of child nodes and the communication network data sequential transmission unit (1-2-1
Send to 3). Communication network data sequential transmission unit (1-2-1
3) selects one address in the received address list as a child node candidate and tries to connect. If the connection is successful, the contact network data is transmitted. If the connection fails, the address is removed from the address list, another address is selected again as a child node candidate, and the connection is retried. When the transmission is completed, the address of the node confirmed as the child node is notified to the own terminal node information management unit (1-2-1-4) together with the information indicating the child number of the child node.

By propagating the communication network data to the lower nodes, each stage of the tree structure is constructed and the communication network is constructed. When the communication network is completed, preparations for collecting terminal data are completed.

In each of the terminals (1-2-1) to (1-2-N), data is input from a keyboard or the like and stored in its own terminal data storage section (1-2-1-6). The terminal data sequential reception unit (1-2-1-5) receives a call or an incoming call from itself in order to all child nodes owned by itself, establishes a data link, and thereafter The data obtained by aggregating the terminal data of the nodes forming the subtree of is received as the terminal data of the child node. For the address of one's own child, the own terminal node information management unit (1-2
It can be obtained by inquiring to 1-4). Whether you want to call yourself or wait for an incoming call,
The determination of is also determined by the node information such as what number the child is for the child. (Details will be described with reference to FIG. 5.) The data aggregating unit (1-2-1-7) stores the own terminal data stored in the own terminal data storage unit (1-2-1-6), Terminal data sequential reception unit (1-2-1
5) receives the terminal information of all the child nodes that are successively received by 1
The data is collected as one group of terminal data and sent to the terminal data transmission unit (1-2-1-8). The terminal data transmission unit (1-2-1-8) receives an incoming call from the parent node or calls itself to establish a data link with the parent node and transmit the aggregated terminal data. To do. The address of its own parent is the own terminal node information management unit (1-2
It can be obtained by inquiring to 1-4). Whether you want to call yourself or wait for an incoming call,
The determination of is also determined by the node information such as the number of children of the parent. (Details are explained in FIG. 5.)

FIG. 3 shows the contents of the tree structure parameters used in the first embodiment and the tree-structured communication network constructed with each center window as the root. As a tree structure, it takes the form of a complete binary tree at a specific depth, and constructs a connecting network that is a straight line at subsequent depths. The reason why the shape is linear after a certain depth is to prevent the tree structure of the connecting network from spreading indefinitely and to control the load on the network to an appropriate amount. In FIG. 3, as tree structure parameters, in particular, a branch limit start stage value 2 that indicates the number of stages at which the number of branch stages begins to be limited, and a branch branch number 2 that indicates the number of child nodes that each node has when the branch is not limited.
Shows the shape of the contact network when is given. In particular, if it is fixed as a binary tree, it is not necessary to provide the number of branching branches as a tree structure parameter, but here, this parameter is provided considering expansion to an n-ary tree.

FIG. 4 shows communication network data propagated from an upper node to a lower node when a complete binary tree portion is constructed in the communication network of the first embodiment. It should be noted that when constructing a connection network for a linear portion, the value of the sibling identifier of the node information transmitted from the parent to the child is always "brother".

FIG. 5 illustrates how terminal data is taken up from a lower node to an upper node in the communication network according to the first embodiment. In the figure, thin arrows indicate the directions of calling and receiving, and thick arrows indicate the direction in which terminal data flows. The amount of data transfer represented by a thick arrow differs between each step, but is the same between those that occur simultaneously in each step.

The procedure taken by each node in order to download terminal data using the communication network in the first embodiment is as follows.

[When the own node is a leaf node] 1-a. If the parent is the eldest son (including the only son), the parent node is called and the data is transmitted to the parent node. If the parent node is the second son of the grandfather node, it may be busy, but in that case, the call is re-established after waiting an appropriate time. 1-b. If it is the second son of the parent node, it waits for an incoming call from the parent node and transmits data to the parent node.

[When the own node is a node having only one child node] 1. Wait for an incoming call from a child node and receive data from the child node. 2-a. If the parent is the eldest son (including the only son), the parent node is called and the data is transmitted to the parent node. If the parent node is the second son of the grandfather node, it may be busy, but in that case, the call is re-established after waiting an appropriate time. 3-b. If it is the second son of the parent node, it waits for an incoming call from the parent node and transmits data to the parent node.

[When the own node is a node having two child nodes (an elder brother node and a younger brother node)] Wait for an incoming call from the older brother node and receive data from the older brother node. 2. Call to the younger brother node and receive data from the younger brother node. However, if the younger brother node has not prepared the data to be sent yet, it immediately disconnects the line, waits for a return call from the younger brother node, and receives the data from the younger brother node. If the younger brother node is not a leaf node, it may be busy, but in that case, it waits for an appropriate time and calls again. 3-a. When the self is the eldest son of the parent node, a call is made to the parent node and data is transmitted to the parent node. 3-b. When it is the second son of the parent node, it waits for an incoming call from the parent node and transmits data to the parent node. However, when an incoming call from the parent node is received before an incoming call from the younger brother node, the parent node is notified that the data is not ready yet, the line is immediately disconnected, and the incoming call from the younger brother node is waited for.
Then, after receiving the data from the younger brother node, the call is returned to the parent node and the data is transmitted to the parent node.

Both the brother node and the brother node can be estimated to have the same time required to collect the terminal data of the node of the subtree rooted at the node. Therefore, when both the older brother node and the younger brother node make calls to the own node, there is a high possibility that the connection requests collide with each other and one of them becomes busy. When you are busy, you have to wait an appropriate amount of time to call again, but depending on how you decide the waiting time, you will lose time. On the other hand, in the first embodiment, when the own node receives an incoming call from the elder brother node and siphons the data from the elder brother node, immediately this time, the self node makes a call to the younger brother node, and the younger brother node sends the data Suck up. In the younger brother node, the possibility of being busy remains as before, but as an estimate, the time required for the own node to request data from the younger brother node and the subtree after the younger brother node There is a difference in the arrival time of the data of the nodes of, so the possibility of being busy is low.

Further, when the data request from the own node occurs before the data of the node of the subtree of the younger brother node arrives, as soon as the data of the subtree reaches the younger brother node, the data is directed from the younger brother node side to the parent. I try to call back.
When the data arrives at the younger brother, the data is immediately transferred to the parent. This also reduces the loss of time.

In the example of the first embodiment, a binary tree is used, but it is easy to expand to an n-ary tree by setting the number of branch branches of the tree structure parameter to n. Each node determines whether to make a call or wait for an incoming call, depending on whether it is the eldest son or the eldest son, or the child of the communication partner. The procedure of can be realized almost in the same manner as described above.

FIG. 6 compares the state of data collection in the first embodiment with the state of data collection by conventional polling. In the figure, circles represent terminal data. The numbers of circles are the same between (a) and (b). Since the data is sucked up using a divergent communication network, the data collection efficiency is higher than that of the conventional polling, and if the same number of data is collected, the data can be collected in a shorter collection time.

In the first embodiment, in order to transmit the contact network data, when the connection to the child node candidate fails, the node is immediately removed from the address list. However, it is possible to easily analogize a method of adding not only an address but also the number of connection failures to that address as ancillary information as an item of the address list and transmitting it to the lower node. In this case, the address that failed to connect because it happens to be busy is
It can be incorporated into the communication network by the lower node. However, if the number of connection failures accumulates through successive nodes and exceeds a certain threshold, it is considered not to be busy but to the failure of the terminal, so it must be removed from the address list. Also, if addresses that have failed to connect many times are removed from the address list, there will be vacancies in the nodes that make up the contact network.To compensate for this, a substitute address list is provided and the center will send the normal address list. It is also possible to easily consider a method of propagating together with the prosthesis and moving it into the communication network.

Further, in the example of the first embodiment, after each node receives terminal data from its own child node and sends it together with its own terminal data to the parent node, there is no work to be done. The rest of the work is done at a node above the own node. This means that each stage of the tree can be operated in a pipeline manner. In other words, the terminal data that occurs regularly, before the previous terminal data has not yet been collected at the center window,
Can be sucked up to the contact network. Also, at this time,
In order to synchronize the operation of each node, if a signal instructing the timing of making a call to each terminal is broadcast from the center, the pipeline processing can be aligned, which is effective.

[0031]

[Embodiment 2] FIG. 7 shows the form of a contact network formed with each center window as a root in Embodiment 2 and the contents of tree structure parameters. Since the system configuration is the same as that of the first embodiment, it will be omitted. In the second embodiment, as a tree structure, a window node has n child nodes C1, C2, ..., Cn, and each subtree rooted at each Ci (i = 1 to n) has a different straight line. Generate a tree structure that is a morphology.
As the tree structure parameter, the number n of branch branches representing the number of branches from the contact node is given. Unlike the first embodiment, no branching is performed after the window, so the branching limit starting stage numerical value is not set. The number Li of nodes forming a linear subtree rooted at each Ci depends on the total number N of addresses registered in the center, the number m of windows, and i, and c = ((N / m) −
When (1 + 2 + ... + n)) / n, it can be determined by Li = i + c. FIG. 7 shows an example in which the number of branch branches n is 6, the number of windows m is 2, and the total number N of addresses registered in the center is 54.

FIG. 8 shows the flow of communication network data in the communication network of the second embodiment. The address registered in the center is
The value of the number n of branch branches of the tree structure parameter and the number m of windows,
Further, based on the total number N of registered addresses, the address list is divided into m address lists each having a size Li (i = 1 ... N). Then, n address lists each having a different size are transmitted from each communication network data sequential transmission unit (1-1-5-1) to.
Allotted to (1-1-5-m). Then, each address list is transmitted as communication network data from each window to each Ci in order from the largest size. Note that, since the communication network after branching at the counter is in a linear form, the only nodes that require sibling identifiers as node information are the child nodes immediately below the counter node.

FIG. 9 illustrates how terminal data is taken up through the communication network in the second embodiment.
The amount of data transfer represented by a thick arrow differs between each step, but is the same between those that occur simultaneously in each step.

Each node constituting the communication network collects terminal data from the child node according to the following procedure, and sends it to the parent node together with its own terminal data.

[When the own node is the contact node] 1. Wait for an incoming call from C1 and receive data from C1. 2. Thereafter, calls are sequentially made to Ci (i = 2 to n) and data is received from Ci. However, if data is not yet prepared in Ci, immediately disconnect the line, wait for a return call from Ci, and if i <n after receiving data from Ci, call to the next Ci + 1. Move to. If Ci is busy, wait an appropriate amount of time before calling again.

[In the case where the number of stages of the local node is 1] Wait for an incoming call from a child node and receive data from the child node. 2-a. When the own node is the eldest son to the parent node, the call is made to the parent node and the data is transmitted to the parent node. 2-b. If the own node is not the eldest son of the parent node, it waits for an incoming call from the parent node and sends data to the parent node. However, if an incoming call from the parent node is received before an incoming call from the child node, inform the parent node that the data is not ready yet, disconnect the line immediately, and wait for the incoming call from the child node. . After receiving the data from the child node, the call is returned to the parent node and the data is transmitted to the parent node.

[In the case of a node having two or more stages of its own node and not a leaf] 1. Wait for an incoming call from a child node and receive data from the child node. 2. Make a call to the parent node and send data. When the parent node is Ci (i = 2 to n), it may be busy, but in that case, the call is reissued after waiting an appropriate time.

[When the own node is a leaf node] 1. Call is made to the parent node and data is sent to the parent node.

In the second embodiment, a node that may be busy is Ci (i = 2 ... n). After receiving an incoming call from C1 and receiving data, the window node
i (i = 2 ... n) is sequentially polled. As an estimate, the time required for the data of subtrees after that node to be aggregated in Cm is Ci (i = 1 ...
It is the same as the time required to collect the terminal data of the child node of m-1). If all the straight line portions of the communication network have the same length (this corresponds to the communication network when the branch restriction start stage value is 1 and the number of branch branches is n in the first embodiment). The terminal data of the subtree arrives at each child node at approximately the same time, and then waits for polling at the center window. The youngest child node is idle until polling arrives. The second embodiment is trying to improve on this point, and the data of the subtree after the child node of the window is collected in the child node so that it is completed one step before polling from the window comes. There is. In other words, while waiting for the polling visit from the window, the younger brother node collects data from as many nodes as possible.

If the aggregation of the data of the subtrees after Ci is not sufficient to complete one step before the polling from the window is completed, the length of the straight line portion after C2 ... Cn remains unchanged. , C1 and the subsequent straight line portion can be lengthened by one node, so that the correction can be easily completed two steps before. However, even if the busy state causes a large loss of time,
The loss becomes an additional margin time for the data collection of the subtree for the sibling node younger than the node. In other words, the loss of time caused by the busy state of the older sibling node makes the younger sibling node less likely to be busy.

FIG. 10 compares the state of data collection in the second embodiment with the state of data collection by conventional polling. In the figure, circles represent terminal data, and the numbers of circles are the same between (a) and (b).

A communication network in which the first and second embodiments are combined can be easily analogized. FIG. 11 shows an example of the contact network and the direction of calling when collecting terminal data in the contact network.

[0043]

[Effects of the Invention] By constructing a divergent tree-structured communication network rooted at each center window and collecting data through this communication network, terminal data can be collected with efficiency exceeding the data collection capacity by polling the center window. can do.

Further, since each node can systematically determine from which side the parent and the child make the call based on the information about the own node in the communication network tree, the timing of the incoming call concentrated on one node can be determined. , It is possible to minimize the loss of data collection work due to busy by shifting as much as possible.

[Brief description of drawings]

FIG. 1 is a system configuration diagram illustrating a first embodiment of the present invention.

FIG. 2 shows a configuration diagram of a conventional data aggregation system.

FIG. 3 shows the contents of tree-structured parameters to be used and a tree-structured communication network configured with each center window as a root in the first embodiment.

FIG. 4 shows communication network data propagated from an upper node to a lower node when a complete binary tree portion is constructed in the communication network of the first embodiment.

FIG. 5 illustrates a procedure in which terminal data is downloaded from a lower node to an upper node in the communication network according to the first embodiment.

FIG. 6 compares the state of data collection in the first embodiment with the state of data collection by conventional polling.

FIG. 7 shows a form of a contact network formed with each center window as a root and contents of tree structure parameters in the second embodiment.

FIG. 8 shows a flow of communication network data in the second embodiment.

FIG. 9 illustrates how terminal data is taken up through a communication network in the second embodiment.

FIG. 10 compares the state of data collection in Example 2 with the state of data collection by conventional polling.

FIG. 11 shows an example of a contact network in which the first embodiment and the second embodiment are combined, and the direction of calling when collecting terminal data in the contact network.

[Explanation of symbols]

 1-1 Center 1-1-1 Address registration section 1-1-2 Tree structure parameter registration section 1-1-3 Communication network data generation and distribution section 1-1-4 Window node information management section 1-1-5 Communication network Data sequential transmission unit 1-1-6 Communication network data sequential reception unit 1-1-7 Data aggregation unit 1-1-8 Aggregated data transmission unit 1-2-1 to 1-2-N Terminal 1-2-1- 1 Communication Network Data Receiving Unit 1-2-1-2 Lower Part Communication Network Data Generating Unit 1-2-1-3 Communication Network Data Sequential Transmitting Unit 1-2-1-4 Own Terminal Node Information Management Unit 1-2 1-5 Terminal data sequential reception unit 1-2-1-6 Own terminal data storage unit 1-2-1-7 Data aggregation unit 1-2-1-8 Terminal data transmission unit

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Harushi Tsukada 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nihon Telegraph and Telephone Corporation

Claims (5)

[Claims] 1. A data aggregating system for aggregating data from a specified number of terminals pre-registered in a center to a center via a public network, the tree structure consisting of a specified number of terminals and having a center window as a root. Data aggregation system having means for configuring the communication network of the above, a means for collecting the data of each terminal to the center window through the communication network, and a means for collecting the data delivered to each center window as one data in the center . 2. The center registers means for registering an address of a terminal which is a node of a communication network, means for registering a parameter characterizing a divergent shape of a tree structure of the communication network, and a center window for each center. Means for generating and distributing the data necessary for generating the tree-structured contact network as contact network data for each contact, means for sequentially transmitting the contact network data from each contact to the terminal of the child node, and the center A means for sequentially receiving terminal data from child nodes of the counter, a means for aggregating the terminal data within the center, a means for transmitting the aggregated data to a specific address, and attribute information as a node of the center contact tree are provided. Each of the terminals has means for managing it as node information, each means for receiving the communication network data from the terminal of the parent node or the center window, and a subtree rooted at the own terminal. Means for generating contact network data, means for sequentially transmitting the generated contact network data to child nodes, means for sequentially receiving terminal data from child nodes, and data input to the own terminal as own terminal data. Means for storing, means for collecting the received terminal data and own terminal data as one data, means for transmitting the collected data to the parent node, and attribute information as a node of the own terminal tree as node information A data aggregation system having means for managing. 3. As one of the parameters for characterizing the shape of the divergent tree structure of the communication network, a parameter for designating to have at most one child after a certain number of stages is provided, whereby the communication network is provided. The data aggregation system according to claim 2, wherein the tree structure of the tree can be prevented from spreading indefinitely and the load on the network can be suppressed to an appropriate amount. 4. The means for sequentially transmitting the contact network data of the center and the terminal to the terminal of the child node determines the number of child nodes that it should have from the tree structure parameter and the node information, and sets the address list to the child node. , Select one address from the divided list as a child node candidate, make a call to that child node candidate, and if the connection is successful, officially adopt that node as a child node. Then, the node information of the child node such as "your own address as a parent node" and "how many children of you", the divided address list including the child node, and the tree structure parameter as the connection network data. If the call to the node candidate fails to connect, on the other hand, the child node candidate is excluded from the split address list, another address is selected instead, and the call is made again. The data aggregation system according to claim 2, wherein a communication network having a tree structure is dynamically configured by excluding a terminal that is registered in an address by attempting a call but cannot function as a node due to a failure or the like. 5. The means for sequentially receiving terminal data from child nodes of a center and a terminal, and the means for transmitting aggregated data to a parent node are “what number of children are parents themselves?” Based on the attribute information as a tree node of its own terminal including "whether it is a leaf node", based on the attribute information, the parent and child nodes decide which side to call and which "which side to call" As such, each node forming the tree in the communication network waits for one incoming call as much as possible, and after completion of data transmission / reception, sequentially calls to some nodes to transmit / receive data, Also, if there is no choice but to receive multiple incoming calls, try to shift the predicted incoming call timing as much as possible, and send the aggregated data before the incoming call from the child node to receive the data. For parents If an incoming call from a node is received first, immediately disconnect the line, wait for an incoming call from the child node, receive data from the child node, and then call back to the parent node. To send the aggregated data, and when the call is made and the other party is busy,
The data aggregation system according to claim 2, wherein a procedure for re-calling after waiting for an appropriate time is used.