JPH0226429A - Optimum hierarchy selecting system for inter-multipoint communication processor - Google Patents

Abstract

PURPOSE:To effectively use a line in inter-multipoint communication by selecting optimumly a inter-multipoint communication processor where the necessary interstational line capacity between the inter-multipoint communication processor and a terminal to make the inter-multipoint communication is minimum in the sequence of a high-order network and a least significant network even in a hierarchical network where star networks, a mesh network, etc., are combined. CONSTITUTION:A high-order hierarchical station (TS).G and least significant hierarchical stations (LS).A and LS.B form the star network, in the same way, TS.H and LS.C and TS.I, LS.D, LS.E and LS.F also form the star-like networks. High-order TS.G, TS.H and TS.I form the mesh network with one another. Even in the hierarchical network where the star networks and the mesh network are combined, the inter- multipoint communication processor where the necessary interstational line capacity between the inter-multipoint communication processor and the terminal to make the inter-multipoint communication is minimum is optimumly selected in the sequence of the high-order network and the least significant network. Thus, the line in the inter-multipoint communication can be used effectively.

Description

[Detailed Description of the Invention] [Summary] The present invention provides one long point-to-point communication process for audio or video signals from multiple terminals in a video conference or remote monitoring using multiple terminals in a hierarchical network. Optimal selection method for long point-to-point communication processing device on hierarchical network, especially in connection with multi-point communication in which signal processing such as screen reduction and synthesis is performed by connecting to a device and sent to multiple terminals. The purpose of this study is to realize an optimal selection method for point-to-point communication processing equipment that can be applied to hierarchical networks. ,
In addition, if all the terminals are not accommodated in the lowest hierarchy station in a star network centered on one upper station in the second hierarchy, among the upper stations in the second hierarchy or higher, the A route is set in which upper stations accommodating communication-related terminals are connected in series, and the total number of terminals accommodated by the upper stations on one side of the route with a certain upper station as the center and the route on the other side are set. Detecting a specific upper station where the total number of terminals accommodated in the upper station is less than half of the total number of terminals, and forming a star network centered on the specific upper station,
If there is a lowest-level station accommodating terminals related to the multi-point-to-point communication that is more than half of the total number of terminals, a long-point point-to-point communication processing device installed in the lowest layer station is selected; The system is configured to select a long point-to-point communication processing device installed in the specific upper station.

In addition, in a state in which all terminals involved in point-to-point communication are not accommodated in one station in the lowest layer, and all the terminals are in the highest layer in a star network centered on one upper station in the second layer, In the case of a state where the terminals are accommodated in a lower hierarchy station, there is a lowest hierarchy station that accommodates more than half of the terminals involved in the point-to-point communication in a star network centered on the upper station. Then, select the long point-to-point communication processing device installed at the lowest layer station,
If it does not exist, the configuration is such that a long point-to-point communication processing device installed in the upper station is selected.

[Industrial application field]

The present invention connects audio or video signals from multiple terminals to one point-to-point communication processing device in video conferencing or remote monitoring using multiple terminals in a hierarchical network, and performs functions such as reducing and synthesizing screens. The present invention relates to point-to-point communication in which signals are processed and sent to a plurality of terminals, and particularly relates to an optimal selection method for a long-point point-to-point communication processing device on a hierarchical network in such a case.

[Conventional technology]

As a representative service of broadband 15DN (integrated service digital integration 'IP), video conferencing has been expected to be in high demand since its inception.

In video conferencing, images sent from terminals at multiple locations (conference rooms) are reduced to form a composite video of all locations and audio is synthesized, or video and audio of a specific location where a speaker is present is created.
Select audio. To this end, at a specific exchange, the input video and audio signals from each terminal (image/audio input device, etc.) at multiple points are combined via the exchange, and the signals are then sent back to each point via the exchange. A long-point point-to-point communication processing device (MPCS, hereinafter the same) that sends out signals to terminals is essential for ease of use for users and effective use of lines. Note that since most of the band used in video conferencing is video, the following explanation will focus on the flow of video signals.

As an invention related to MPC3 as described above, the present applicants have proposed a "video processing device" (Japanese Patent Application No. 320136/1983).
has been applied for.

Such an MPC3 is generally attached to a large number of stations (exchanges) scattered within a network, but depending on which station's long point-to-point communication processing device is selected when a call is made, the long-point to long-point communication processing device The inter-office line capacity required for each station differs greatly.

For example, as shown in FIG. 8, the lowest layer stations A-F (or LS layer stations, hereinafter LS-
A-LS-F) and an upper layer station G that relays between other stations.
, H1■ (or TS ladder station, hereinafter referred to as TS-G-TS・■)
In a hierarchical network consisting of LS-A-LS-E, terminals a-e for multipoint communication are connected to LS-A-LS-E, and terminals f-k are connected to LS and F. Consider the case. In this case, a case will be considered in which a reduced composite video of all terminals a to e is created in MPCS, and switching and distribution connections are performed at a station (exchange) to which MPCS is connected.

In this case, for example, as shown in FIG. 8(a), L
Assuming that MPCSI of S-F is selected, MPC
SI needs to set up individual lines with all terminals a to f for multipoint communication, but in this case, the required inter-station line capacity from each terminal a to f to MPCSI is calculated as follows. Ru. In FIG. 8, numbers in parentheses indicate line capacity.

First, between stations A and C and between stations B and G, terminals a and b
One line is required for each. Between stations G and H,
Two lines are required, one between A and G and one between B and 0. One line corresponding to terminal C is required between stations C and H. Between stations H and I, between G and H and between C and H,
A total of 3 lines are required between stations 0 and DT and between stations E and R.
One line each corresponding to terminals d and e is required between the two terminals. Between stations 1 and F, between H and I, between D and H, and between E and
5 lines with 1 line are required.

Therefore, the sum of the above-mentioned line capacities between the stations, 15 lines, becomes the required inter-office line capacity from each terminal a to f to the MPCSI.

On the other hand, the required inter-station line capacity from MPCSI to each terminal a-f can be calculated by distributing the signal (video signal) output from MPCSI via LS-F to each station I, I(, G, and E-A). Therefore, a total of eight lines are required, one line between each station.

From the above, the total required inter-office line capacity between the MPCSI and each terminal a to f is 23 lines.

On the other hand, for example, as shown in FIG. 8(b), the MP of LS-A
The required inter-office channel capacity from each terminal a to f to MPCSI when CSI is selected is calculated as follows.

First, between stations F-1, six lines are required for each of the terminals f to k. One line each is required between stations D-1 and E-1, corresponding to terminals d and e. Station 1-
Between H, eight lines are required including the above F-1, EI and D-1. One line corresponding to terminal C is required between stations C and H. Between stations HG, the above 1-H
A total of 9 lines are required, including between the two lines and between C and H. Station B-
Between G and G, one line corresponding to each terminal is required.

Between stations G and A, 10 lines are required, including the lines between H and G and between B and 0.

Therefore, the sum of the above-mentioned line capacities between the stations, 37 lines, becomes the required inter-station line capacity from each terminal az f to the MPCSI.

On the other hand, the required inter-station line capacity from MPCS l to each terminal a y f can be achieved by distributing the signal (video signal) output from MPCSI via LS-A to each station G, H1 and B-F. Therefore, a total of eight lines are required, one line between each station.

From the above, the total required inter-office line capacity between MPCSI and each terminal azf is 45 lines.

Similarly to FIGS. 8(a) and (b), the MPC of each station A-F
The required inter-office line capacity when S is selected is shown in the table of FIG. The required line capacity from MPCS to each terminal is as follows:
No matter which station selects MPCS, there will be 8 lines in the same way as in the case of FIG. As can be seen from the figure, there is a case where the MPCS of the optimal station LS-F is used and a case where the MPCS of the worst station LS-F is used.
It can be seen that the required inter-office line capacity differs by about twice when using A or LS-B. Therefore, in order to utilize the line efficiently, an algorithm is required to select the optimal MPCS of the station.

As a conventional example of an algorithm for selecting the optimal MPCS of a station, the present applicants have disclosed the following algorithm (IEICE Office System Study Group 0
387-24).

Now, in a simplified network model consisting of LS-A-LS-D and terminals a to e as shown in Fig. 10,
First, we focus on the station (exchange) on the far left (right), and if the number of connection requesting terminals accommodated in the station on the right (left) of that station is greater than the number of other connection requesting terminals, then Shift to one station to the right (left) and check the number of connection requesting terminals accommodated in the station to the right (left) of that station. Repeat this action to the right (
MPC3 of the station at the last stationary position after shifting to the station on the left)
Select.

By adopting the above algorithm, station C
When focusing on , the number of connection requesting terminals accommodated in the station to the right of that station is one corresponding to terminal e,
The number of connection requesting terminals corresponding to the other terminals a and b is smaller than 2. Therefore, it can be seen that connecting MPC3I to station C is optimal.

[Problem to be solved by the invention] However, the above conventional algorithm is based on the premise that it is applied to a hierarchyless network topology (form) as shown in Fig. 10, and is the most The problem is that it cannot be applied to hierarchical network topologies such as star networks, rope networks, or cables that combine these, which are often adopted, and the range of application is extremely narrow.

An object of the present invention is to realize an optimal selection method for multipoint communication processing devices applicable to hierarchical networks.

[Means to solve the problem]

FIG. 1 is a diagram showing an operation flowchart of the present invention.

The method of the present invention provides long-distance point-to-point communication performed by connecting multiple terminals of a hierarchical network and one multipoint-to-point communication processing device.
This is applied when all terminals involved in point-to-point communication are not accommodated within one station in the lowest hierarchy. Detection of this state can be accomplished by, for example, sending a message from the originating terminal related to the point-to-long point communication to the lowest layer station on the originating side in which it is accommodated, specifying all terminating terminal numbers and long point-to-point communication related to the long point-to-point communication. By transmitting a call signal through a common line, the lowest layer station on the originating side performs the above detection.

First, if all the terminals are not accommodated in the lowest floor open station in the star network centered on one upper station in the second layer, the following process is executed. Note that this state can be detected, for example, by sending a calling signal specifying all the terminating terminal numbers and point-to-point communication from the lowest layer station on the originating side to the upper station on the originating side that accommodates the station via a common line. By sending the information, the originating upper station performs the detection described above.

In the above state, first, as shown in FIG.
A route is set that connects in series upper stations that accommodate terminals related to the point-to-point communication among the upper stations of the second or higher hierarchy. This setting is realized, for example, by each of the upper stations exchanging control information via a common line starting from the originating upper station.

Next, as shown in S2 in FIG. 1, the total number of terminals accommodated in the upper stations on the route on one side and the number of terminals accommodated in the upper stations on the route on the other side with a certain upper station as the center are calculated. A specific upper station is detected for which the total number of terminals is less than half of the total number of terminals. This detection of the specific higher-level station is realized, for example, by each higher-level station on the route exchanging information regarding the number of terminals involved in the point-to-point communication with other higher-level stations via a common line.

If there is an open station on the lowest floor accommodating more than half of the total number of terminals related to the point-to-point communication in the star network centered on the specified upper station, then
4, select the multipoint communication processing device installed at the lowest floor opening, and if it does not exist, select
As shown in FIG. 2, a multipoint communication processing device installed at the specific upper station is selected. This selection may be made, for example, between the specific higher-level station and each lowest-floor station that accommodates terminals related to long-point communication in a star network centered on the specified higher-level station, between the long-point points via a common line. This is achieved by exchanging information regarding the number of terminals involved in communication.

On the other hand, in the case where all the terminals are accommodated in the lowest hierarchy station in the star network centered on one upper station in the second hierarchy, the processes of 31 and S2 in FIG. 1 are executed. There is no need to do so, and it is sufficient to regard that one higher-level station as the specific higher-level station and directly execute the processes 53 and below.

[For production]

By using the above means, even in a hierarchical network such as a combination of a star network and a reticular network, communication between a multipoint communication processing device and a terminal that performs multipoint communication can be performed in the order of the upper network and the lowest network. Optimum selection of a multipoint communication processing device that minimizes the required inter-office line capacity between stations can be achieved, thereby making it possible to efficiently utilize lines in multipoint-to-point communication.

In addition, in multipoint-to-point communication performed by connecting multiple terminals in a hierarchical network and one multipoint-to-point communication processing device, all terminals involved in multipoint-to-point communication are accommodated in one station in the lowest hierarchy. In such a case, the multi-point communication processing unit 5 of that station may be selected to perform the multi-point communication processing.

〔Example] Hereinafter, embodiments of the present invention will be described in detail.

Figure 2 shows the network topology (form) targeted by this example.
FIG. As shown in the figure, this embodiment performs relaying between the lowest layer stations A to F (or LS layer stations, hereinafter referred to as LS-A-LS/F) that accommodate terminals a to k, and other stations. G, H5■ (or TS ladder station, hereinafter TS
-G-TS-1) is applied to a hierarchical network consisting of T
S-G, S-A, and LS-B form a star-like network, and similarly, TS-II and LS-C, and TS-■, LS-D, and LS
-E.

LS-F also forms a star-like network, and the upper TS-G and TS-H
and TS-I form a reticular network with each other.

In this case, between each station, each link (for example, CCITY SS No.
, 7).

Further, a multipoint communication processing device (MPCS, hereinafter the same), which is not particularly shown, is connected to each LS and TS.

Then, the MPC according to this embodiment shown in FIGS. 3 to 5
The optimal selection algorithm No. 3 is executed by a control device (not shown) in each LS or TS while exchanging information with other LSs or TSs via the common line 3.

The operation of this embodiment in the above network topology will be explained below.

Here, terminals a to e, which should perform point-to-point communication, are connected to each of LS-A to LS-E, and LS-F
Similarly, consider the case where terminal f=k is connected. Also, M at any one station among each LS or TS
PC3 is selected by the optimum selection algorithm described later so that the required inter-office line capacity during multi-point communication is the smallest, and a reduced composite video of all terminals a to e is created in the Mpcs, and that station (exchange ), consider the case where an exchange/distribution connection is performed.

Under the above conditions, for example, if the caller operates terminal a, the operation flowchart shown in FIG. 3 is activated. FIG. 3 is a diagram showing an overall operation flowchart for selecting an optimal MPC3 during multipoint communication.

First, the caller at terminal a sends a number for identifying the terminals of all called parties and a calling signal indicating multipoint-to-point communication to the calling side LS, that is, LS-A (see Fig. 336, hereinafter Fig. 3). ).

As a result, the LS-A first identifies the numbers of all the called parties that have been sent, and all the called parties are connected to the originating side LS, that is, the LS-A.
It is determined whether it is included in A (S7).

As a result, if all called parties are included in LS-A, naturally the originating LS, that is, the MPC connected to LS-A,
3 (313), and after performing call control for communication between multiple points (312), communication between multiple points can be started.

However, in the connection state shown in Figure 2, all called parties are using LS/A.
Since it is not included in the above process, the calling party L
S, that is, LS-A, and the upper TS, that is, T
A control signal, that is, a number identifying the terminals of all called parties and a calling signal indicating point-to-point communication is transferred to the SG via the common line 3 (S8).

As a result, the TS that accommodates the originating LS determines whether all called parties are included in the constellation network centered on the originating LS (S9).

As a result, if all called parties are included, the MP for the star network is
The process proceeds to process 2 (Sll), which is the optimal selection algorithm of C3. Note that this will be described later.

However, in the connection state shown in Figure 2, all called parties are
TS that accommodates S, that is, TS-G that accommodates LS-A
Since it is not included in the above process 2, processing 1 (S
Proceed to IO).

FIG. 4 shows an operation flowchart of process 1 in S10 of FIG. 3 above.

First, the lower L related to the terminal that requests connection for long point-to-point communication.
Regarding the upper TSs accommodating S, a route for connecting them in series is determined (314 in FIG. 4, see FIG. 4 below). That is, in the example of FIG.
The lower LSs related to this are LS.A-LS-F, and the upper TSs that accommodate them are TS-G, TS-H, and TS.■. Therefore, the route for connecting these in series is the route shown at 4 in FIG. Note that this route determination operation is performed on the TS including the originating LS, that is, the LS in the example of FIG.
This is done by starting from the TS-G containing A and sequentially transferring the called party number and calling signal to other related stations via the common line 3, but such route determination operation is generally This is the basic operation of the exchange network, and is flexibly determined depending on the traffic situation of the mesh body, so its details will be omitted.

Next, the TS including the originating LS is selected as the initial station (315). In the example of FIG. 2, TS-C including LS-A
; is selected.

Next, count the total number N of terminals that perform point-to-point communication under the upper TS on the left side of the TS that includes the own station, that is, the originating LS, and similarly count the number of terminals that perform long point-to-point communication under the upper TS on the right side. The total NR of the number of terminals performing point-to-point communication is counted (S16).

Now, in the example of FIG. 2, first, the TS including the originating LS selected in S15 is TS-G, and in this case, TS-G
Since there is no upper TS further to the left, Nt is 0.
becomes. Also, the upper TS on the right side of TS-G is TS
-H and TS·■, and the total NR of the number of terminals under them that perform point-to-long point communication is 9 in total for terminals c - k. Note that the operation of counting the number of terminals is performed while the own station exchanges information regarding terminal numbers related to point-to-point communication with other stations via the common line 3.

Once NL and NR are determined by the above operation, the total number of terminals N, NL and NR that perform point-to-point communication in the upper station
The magnitude relationship is compared (S17).

As a result, if NL>N/2 and NR<N/2, select the upper TS one position to the left of your own station in the route determined in step 314, and transfer control to it via the common line ( 31B).

On the other hand, if N, <N/2?NR>N/2, then the S
In the route determined in step 14, select the upper TS one position to the right of your own station and transfer control to it via the common line (3
19).

Note that the case of NL≦N/2 and NR≦N/2 will be described later.

In the example of FIG. 2, the total number of terminals N is 11 terminals a - k, and as mentioned above, NL=O1NR=9, so
S19 is executed and control is transferred to TS-H. The situation is shown in FIG.

Subsequently, for the new higher-level station TS selected by the above process, NL and NR are calculated again and their magnitudes are compared (S16.517).

In the example of Fig. 2, for TS-H, the upper TS on the left side is TS-G, and the total number of terminals N, which perform long point-to-point communication under it, is the number of terminals a and b. Total 2
becomes. Also, the upper TS on the right side of TS-H is TS
-I, and the total NR of the number of terminals under it that perform point-to-point communication is 8 in total for terminal d-. Therefore, S
Based on the comparison result of step 17, S19 is executed again, and control is transferred to TS.■ as shown in FIG.

Then, the new upper station TS- selected by the above process is
Regarding 1, the calculation of NI and NR and the comparison of their magnitude are performed again (816.517). Regarding TS-I, the upper TSs on the left side are TS-G and T.
SH, and the total number NL of terminals that perform point-to-point communication under these terminals is 3 (terminals a''-c).

Further, since there is no upper TS on the right side of TS-1, NR is O. Therefore, according to the comparison result in S17, NI=3≦N/2 and Ntr=0≦N/2,
Unlike the above, S20 is executed, and it is determined that any MPC3 of the stations in the network centered on the upper 'FS is the optimal MPC3 for performing multi-point communication from now on. In the example of FIG. 2, the stations in the star network shown in FIG.
MPC3 of one of the stations in LS-F is optimal.

Through the above operations, once a specific upper TS is determined, the fourth
Process 1 in the figure is completed, and the process moves to process 2 from SIO to Sll in FIG. 3.

FIG. 5 shows an operation flowchart of process 2 in FIG. 3 311 above. Here, among the MPC3 of stations in a star network centered on a specific upper TS determined by the SIO in FIG.
An operation is performed to select the optimal one and determine the final MPC3. In addition, in 39 of FIG. 3, if it is determined that all called parties are included under the upper TS including the originating LS, it is only necessary to consider the star network centered on the upper TS, and the third Since there is no need to execute process 1 of 310 in the figure or FIG. 4, process 2 of Sll is directly executed.

In FIG. 5, first, one lower LS accommodating a terminal that performs point-to-point communication under the upper TS of interest is selected, and control is transferred to the lower LS (321). In the example in Figure 2, the lower LS-D-L under the upper TS・■
SF is the target, and for example, LS-D is selected.

Next, in the lower LS selected in the above operation, the number of terminals performing point-to-point communication is counted. In the example of FIG. 2, the number n of terminals that perform point-to-point communication in the LS-D is one terminal d.

When n is determined by the above operation, the total number N of terminals that perform point-to-point communication in the lower station is compared with the magnitude of n (S23).

As a result, if n>N/2, the process of S24 is executed (described later), and if n≦N/2, the process of S25 is executed.

In S25, it is first determined whether all lower LSs that accommodate terminals that perform point-to-point communication have been checked, and if all have been checked, the S
27 will be described later), selects another lower LS if there is another lower LS, and transfers control to that lower LS (
S25-S26). In the example of FIG. 2, since LS-E and LS-F still exist in addition to LS-D, S26 is executed and control is transferred to, for example, LS-E. The situation is shown in FIG.

Next, for the new lower LS selected by the above process, n is calculated again and the magnitude relationship is compared (S22.523). In the example of FIG. 2, for LS/W, the number n of terminals that perform point-to-point communication is one terminal e, so n < N / 2, and according to the comparison result in S23, it is 325 again, and then in S26 is executed, and control is transferred to LS/F as shown in FIG.

Then, for LS-F, the calculation of n and the comparison of their magnitudes are performed again (S22. For LS-F, the number n of terminals that perform point-to-point communication is 6 for terminal f-. Therefore, unlike the above, n>N/2 and the process of S24 is executed, as shown in FIG.
The MPC3 in the lower LS at that time is finally selected as the optimal one.

On the other hand, there is no lower LS that satisfies the condition of 324, and S2
When all lower LSs have been investigated in step 5, step S27
It is determined that the optimal MPC3 does not exist in the lower LS, and the MPCS in the higher TS at that time is finally selected as the optimal one.

When the optimal MPCS is selected through the above operations, processing 2 in FIG. 5 is finished, and the process moves from Sll to 312 in FIG. 3, where call control regarding multipoint communication is performed based on the selected MPCS. Start multipoint communication. That is, in the example of FIG. 2, MPCS is activated in LS-F, call control of multipoint communication is performed centering on that station, terminals a to k are linked, and multipoint communication is started. . Note that call control at this time is performed via the common line 3.

According to the algorithms in Figures 3 to 5 above, the required inter-station link capacity l when MPCS of LS/F in Figure 2 is selected is the required inter-station link capacity l when another LS or TS is selected. This makes it possible to minimize the capacity.

〔Effect of the invention〕

According to the present invention, even in a hierarchical network such as a combination of a star network, a mesh network, etc., terminals that perform multipoint communication with a multipoint communication processing device are connected in the order of the upper network and the lowest network. It becomes possible to provide an algorithm for optimally selecting a multipoint communication processing device that minimizes the required inter-station line capacity between
This enables efficient use of lines in multipoint communication.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an operation flowchart of the present invention. FIG. 2 is a diagram showing a network topology targeted by this embodiment. FIG. 3 is a diagram showing an overall operation flowchart of this embodiment. , FIG. 4 is a diagram showing an operation flowchart of process 1, FIG. 5 is a diagram showing an operation flowchart of process 2, FIG. 6 is an explanatory diagram of the operation of selecting the optimal MPC 3 in the upper network, 8(a) and 8(b) are explanatory model diagrams of the influence of the required inter-station channel capacity depending on the position of the multipoint communication processing device. The figure is a diagram showing the relationship between stations from which MPCS is selected and required inter-station channel capacity. 1...Multipoint communication processing device (MPCS), LS...
・Between the lowest hierarchy, TS...between the upper hierarchy.

Claims (1)

[Claims] 1) In multipoint communication performed by connecting multiple terminals of a hierarchical network and one multipoint communication processing device, all terminals involved in the multipoint communication are connected to one multipoint communication processing device in the lowest hierarchy. If the terminal is not accommodated in one station, and all the terminals are not accommodated in the lowest hierarchy station in a constellation network centered on one upper station in the second hierarchy, then Among the stations, a route is set to serially connect higher-level stations that accommodate terminals related to the multipoint communication (S1), and the upper-level stations accommodated in the higher-level stations on one side of the route centering on a certain higher-level station are set. detecting a specific upper station where both the total number of terminals and the total number of terminals accommodated in the upper station on the route on the other side are less than half of the total number of terminals (S2); If there is a lowest layer station that accommodates more than half of the total number of terminals related to multipoint communication in a star network centered on a specific upper layer station, the multipoints installed at the lowest layer station A multipoint communication processing device hierarchy characterized in that a multipoint communication processing device is selected (S3→S4), and if the multipoint communication processing device does not exist, a multipoint communication processing device installed in the specific upper station is selected (S3→S5). Shape optimal selection method. 2) In multipoint communication performed by connecting multiple terminals in a hierarchical network and one multipoint communication processing device, a state in which all terminals involved in multipoint communication are not accommodated in one station in the lowest hierarchy. , and all the terminals are accommodated in the lowest hierarchy station in a constellation network centered on one upper station, which is the second hierarchy, for the constellation network centered on the upper station If there is a lowest hierarchy station that accommodates more than half of the total number of terminals involved in the multipoint communication, select a multipoint communication processing device installed in the lowest hierarchy station (S3→S4). , If the multipoint communication processing device does not exist, the multipoint communication processing device installed in the upper station is selected (S3→S5).