JP6892205B2 - Multipoint communication systems and methods and programs - Google Patents

Description

The present invention relates to a multipoint communication system that performs group communication between a plurality of terminals using a multipoint connection function as an alternating point.

In recent years, along with the development of ultra-high-definition cameras / display devices such as 4K / 8K and AR (Augmented Reality) / VR (Virtual Reality) technology, telework of each company for the purpose of energy saving and life work balance. Demand for video communication services via networks is increasing due to recommended measures.

A video conference is a typical example of a video communication service. A well-known device for configuring a video conferencing system is a multipoint control unit (MCU). The MCU is a device placed on the network as an exchange point for group communication, and the MCU for the video conferencing system receives video / audio packets from all user terminals and then displays the video of each participant on one screen. It is a device that synthesizes and distributes video / audio to all user terminals.

In the multipoint video conference using the MCU, the terminal load is reduced as compared with the P2P (Peer to Peer) type multipoint video conference, but there is a problem that the calculation load and the network load of the MCU server are increased. As an example of a method for reducing the network load of MCUs, a system in which a plurality of MCUs are geographically distributed and deployed is known (see, for example, Patent Document 1). In this system, for example, the MCU in Osaka and the MCU in Tokyo are cascade-connected while separating the MCUs accommodating the conference participation terminals in the Osaka area and the Tokyo area. The video of the conference participants from multiple terminals in the Osaka area is sent to the MCU in Tokyo in a state of being combined on one screen in the MCU in Osaka. Similarly, in the opposite direction, the images of the conference participants of multiple users in the Tokyo area are sent to the MCU in Osaka in a state of being combined on one screen in the MCU in Tokyo. With such a system, it is possible to suppress an increase in the network bandwidth between Tokyo and Osaka.

Japanese Patent No. 3457202 Japanese Unexamined Patent Publication No. 2003-69563

According to the method described in Patent Document 1 described above, since a plurality of MCUs are cascade-connected, the relay network is used even when the number of terminals participating in the conference increases or when an attempt is made to improve the video quality rate. It is expected that it will be possible to prevent an increase in network load at the interconnection part (gateway) with other networks. However, for the one described in Patent Document 1, it is necessary to prepare dedicated hardware in advance, and since the MCU has no choice but to use the fixedly deployed one, the service subscriber situation changes from the initial equipment design. There is a problem that the optimum is impaired in such a case.

Therefore, as described in Patent Document 2, there is a method of dynamically changing the MCU to be used according to the configuration of the participating terminals each time a meeting occurs from the group of MCUs distributed geographically. Proposed. This method allows you to select the optimal MCU (for example, to minimize network traffic caused by the conference) each time the conference is held, so that the participant area configuration is not tied to the initial design. Effective relay network bandwidth can be saved accordingly. Further, even if a certain MCU breaks down, a meeting can be held because another candidate MCU can be selected. Another similar method is to calculate the proximity (the sum of the reciprocals of the available bands of each section) for each participating terminal and select the MCU that minimizes the sum of the proximity at the conference. There is.

However, those described in Patent Document 2 and the like are provided only with a method of selecting a single MCU for each conference, and effectively a plurality of MCUs are effectively selected according to the regional bias of the conference participating terminals. No mention is made of how to elect and cascade MCUs.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a multipoint communication system, a method, and a program in which the load and delay of a network or a server are small.

In order to achieve the above object, the present invention is a multipoint communication system in which group communication is performed between a plurality of terminals accommodated in a network, and a tree structure in which the network topology is lower than the terminal side in the network. A multi-point connection function that operates on a plurality of servers geographically distributed so as to be such as, a network configuration information including at least the position information of the terminal and the position information of the server, and the above, prior to the start of the group communication. Based on the resource information of the network and the server, the weighting of the route section is set for each layer of the tree structure, and the weighting is performed for each arrangement pattern of the multipoint connection function assuming use in the group communication in the network. The usage cost is calculated from the route cost and the operating cost of the server on which the multipoint connection function operates, and the placement pattern for which the minimum usage cost is calculated is set as the optimum placement pattern, and one or more of the optimum placement patterns are used. When there are a plurality of multipoint connection function selection means for selecting the multipoint connection function and the multipoint connection function selected by the multipoint connection function selection means, the multipoint connection functions are linked with each other to form a plurality of points. It is characterized by being equipped with a multipoint connection function control means that controls each multipoint connection function so that the terminal behaves as if a single group communication is being performed while straddling the multipoint connection function. ..

According to the present invention, even when the number of terminals participating in group communication increases, it is possible to prevent an increase in the bandwidth of the gateway portion with the network or another network. And since the bandwidth restriction of the gateway part with the network and other networks is relaxed, if the access section between the terminal and the server is a high-speed line such as FTTH (Fiber to the Home) or 5G, wide bandwidth is provided. It enables higher-definition and more realistic video conferencing. Further, even if a failure or a shortage of available resources occurs in a specific server, it is possible to avoid using the server and carry out group communication between multiple points.

The figure explaining the outline of this invention System configuration diagram of multipoint communication system System configuration diagram of a multipoint conference system according to another example Functional block diagram of MCU management function Functional block diagram of the MCU management function according to another example Functional block diagram of the MCU management function according to another example Example of flow from meeting reservation to meeting holding and meeting end in multipoint meeting system Example of flow from meeting reservation to meeting holding and meeting end in a multipoint meeting system (multiple meetings held at the same time) Flow example of MCU optimum placement calculation processing Detailed flow example of MCU optimal placement calculation processing The figure explaining the specific example of the MCU optimum arrangement calculation processing A simple example of cost calculation A simple example of cost calculation A simple example of cost calculation A simple example of cost calculation A simple example of cost calculation The figure explaining the MCU optimum arrangement calculation step which concerns on another example. The figure explaining the MCU optimum arrangement calculation step which concerns on another example. Figure (1/3) Explaining MCU Optimal Placement Calculation Steps Related to Other Examples (Multiple Meetings Held Simultaneously) Figure (2/3) Explaining MCU Optimal Placement Calculation Steps Related to Other Examples (Multiple Meetings Held Simultaneously) Figure (3/3) Explaining MCU Optimal Placement Calculation Steps Related to Other Examples (Multiple Meetings Held Simultaneously)

First, the outline of the present invention will be described with reference to FIG. As shown in FIG. 1B, the present invention is a multipoint communication system that performs group communication between a plurality of terminals 1 housed in a network, and is geographically distributed in the network (for example, for each municipal area). The present invention relates to a multipoint communication system that performs group communication using the multipoint connection function 3 distributed in the above as an exchange point. Here, the "multipoint connection function" means a device that functions as a conventional multipoint connection device (MCU: Multipoint Control Unit), and its mounting form is irrelevant as described later. In FIG. 1 (b), the multipoint connection function 3 is designated as MCU3. Note that FIG. 1 illustrates a multipoint conference system that realizes a multipoint video conference as an application example of a multipoint communication system.

In the prior art shown in FIG. 1A, only one MCU3 is deployed as a whole, and the images of all the conference participating terminals 1 are combined and distributed to each participating terminal 1. Therefore, there is a problem that the video traffic is concentrated on one MCU3. On the other hand, in the present invention, as shown in FIG. 1 (b), the primary summary screen is synthesized for each area in each distributedly deployed MCU3, and the primary summary screen is exchanged between the MCU3s to perform the final synthesis. Deliver to each participating terminal 1. This has the advantage that the relay network can be used in a single screen band.

In the present invention, it is optimal according to network configuration information including the coverage area of the terminal 1 of the group communication participant, network resource information such as the available bandwidth of each network section, and infrastructure information such as server resource information of the edge server 2 ( Select the edge server 2 for (one or more) MCUs (optimal deployment calculation).

Further, in the present invention, a plurality of MCUs 3 are linked to each other, and the MCUs 3 are controlled so as to appear as the same virtual group communication from the user's terminal 1. As an example of control, a parent MCU and a child MCU are selected from the determined MCU3s and temporarily cascade-connected.

Further, the present invention is also characterized in that the above-mentioned group communication is held based on the reservation information, and the MCU selection calculation is started by calculating back from the time required for the MCU selection calculation and the MCU instance development.

In the present invention, in order to realize the above-mentioned group communication, an MCU management function 4 (or a node / device having the above-mentioned function) for managing an MCU group is provided. As described above, the MCU management function 4 calculates the optimum deployment of the MCU 3 and gives a cooperation instruction to the MCU 3.

In FIG. 1, the network accommodating the plurality of edge servers 2 is connected to another network such as the Internet. Therefore, when viewed from other networks such as the Internet, each edge server 2 appears to be located on the "edge" side of the network, that is, closer to the terminal 1 in the entire network including the other networks. , Each server in the figure shall be called an edge server. The same applies to the following description.

Next, a more specific embodiment of the present invention will be described with reference to FIG. In the present embodiment, the multipoint conference system will be described as in FIG. 1 described above. FIG. 2 is a system configuration diagram of a multipoint conference system.

As shown in FIG. 2, the multipoint conference system includes a plurality of terminals 10, an access network 20, a relay network 30 that spans a plurality of regions, a geographically distributed edge server 40, and the edge server 40. The MCU application 50 deployed above and the MCU management function 60 for controlling the edge server 40 and the MCU application 50 are included.

The access network 20 is a network that accommodates the terminal 10 in the relay network 30. FIG. 2 does not mean that a plurality of terminals 10 are accommodated in the relay network 30 by one access network 20, but is a logical network showing the relative deployment positions of the access networks 20 in the network topology. Please note that.

The relay network 30 is a logical network configured by interconnecting a plurality of regional networks constructed in a geographically distributed manner, and includes a plurality of network devices 35 such as routers distributed in a geographically distributed manner. As shown in FIG. 2, the network topology of the relay network 30 in the present embodiment is assumed to have a tree structure in which the network device 35 is a node. The relay network 30 is connected to another network 70 such as the Internet via the uppermost network device 35 having a tree structure. In the following description, each node in the access network 20 and the relay network 30 will be referred to as a "network section".

The edge server 40 is connected to the network device 35 in the vicinity of the network device 35. That is, the edge servers 40 are also geographically distributed in the relay network 30. The edge server 40 is connected to an arbitrary network device 35 in the relay network 30. FIG. 2 illustrates only the network device 35 to which the edge server 40 is connected. Further, a plurality of edge servers 40 may be connected to one network device 35. FIG. 2 shows a case where the number of connected edge servers 40 is one for all network devices 35.

The MCU management function 60 includes resource information (network topology, location information of terminal 10, location information of edge server 40, route cost, etc.) of one or more of the access network 20, the relay network 30, and the edge server 40. It is provided with means for acquiring the available bandwidth of each network section, the available resources of the edge server, the operating cost of the edge server, etc.). In the example of FIG. 2, among the resource information, the information related to the network is acquired via the network management node 81, and the information related to the server resource is acquired via the server infrastructure management node 82.

The MCU management function 60 includes means for obtaining conference reservation information (meeting start / end time, maximum number of participants, list of participating terminals, etc.) from the user. In the example of FIG. 2, the case where the conference reservation information is acquired via the conference reservation system 90 is shown.

The MCU management function 60 has a function of selecting one or a plurality of optimum edge servers 40 for MCU according to the area of the conference participants based on the acquired resource information and conference reservation information.

The MCU management function 60 includes means for issuing a command such as holding / ending a conference to the MCU 50 on each edge server 40 based on the result calculated by the selection function. If the MCU 50 is software that runs on a VM (Virtual Machine) on the virtual hypervisor, the VM is created / deleted / started / terminated for the virtual hypervisor and VM, and the MCU application 50 is started / terminated. It is equipped with means for issuing commands such as. In the example of FIG. 2, the various commands are described by using an example in which the various commands are executed via the server infrastructure management node 82.

Various forms of the edge server 40 and the MCU 50 can be considered. The edge server 40 may be a virtual hypervisor (or virtual host), and the MCU 50 may be software that operates on a virtual machine (VM). Further, the edge server 40 may be a container host, and the MCU application 50 may be provided on the container. Further, the edge server 40 can be a general-purpose server, and the MCU 50 can be an application software. Further, it can be in the form of an appliance product in which the edge server 40 and the MCU 50 are integrated.

The MCU 50 executes the process related to the group communication according to the instruction from the MCU management function 60. When group communication is carried out by a plurality of MCUs 50, each MCU 50 cooperates with another MCU 50, and the group communication is performed so that the terminal 10 behaves as if a single group communication is carried out from the terminal 10 while straddling the plurality of MCUs 50. Carry out such processing. In the case of a conference system as in the present embodiment, the MCU 50 firstly synthesizes data from each terminal 10 under its own control, and transmits the synthesized data to another MCU 50. In addition, the MCU 50 creates final composite data by synthesizing the composite data synthesized by itself and the composite data received from another MCU 50, and transmits the final composite data to each terminal 10 under its own control. The MCU 50 controls the holding / deletion / termination of group communication, the midway participation / midway exit of the terminal 10, and the like in cooperation with other MCU50s.

The network management node 81 includes infrastructure resource information (network topology, location information of the terminal 10, location information of the edge server 40, route cost, and each of the access network 20 and the relay network 30) related to the network. A means for acquiring the usable band of the network section, etc.) is provided. The infrastructure resource information is acquired from various network devices and network management devices constituting the access network 20 and the relay network 30. In the example of FIG. 2, at least a part of the infrastructure resource information is acquired from each network device 35. Further, the network management node 81 includes means for notifying the MCU management function 60 of the acquired information.

The server infrastructure management node 82 includes means for acquiring resource information related to the edge server 40 (usable resources (CPU, memory, storage, GPU, etc.) of the edge server 40, operating costs when the server is used, etc.). The server resource information is acquired directly from the edge server 40 or from the management device if there is a management device that manages the edge server 40. Further, the server infrastructure management node 82 includes means for notifying the MCU management function 60 of the resource information related to the acquired edge server 40.

Further, the server infrastructure management node 82 includes means for performing various operations on the edge server 40 and the MCU 50 based on instructions from the MCU management function 60. For example, when the edge server 40 is a virtual hypervisor (virtual host server) and the MCU 50 is a software application running on a VM, the server infrastructure management node 82 transmits a VM image to each edge server 40. , A means for creating a new VM, deleting an existing VM, and starting / stopping a VM. Further, the MCU application 50 is provided with means for creating / deleting a conference room, instructing a terminal to make a call, linking a plurality of MCU applications, and executing the conference room as the same virtual conference room.

It should be noted that the above-mentioned instructions from the MCU management function 60 such as cooperation, conference room creation / deletion, and calling to the MCU 50 are directly sent from the MCU management function 60 without going through the server infrastructure management node 82. Is also good.

Further, as shown in FIG. 3, the above-mentioned instructions from the MCU management function 60 such as cooperation, conference room creation / deletion, and calling to the MCU 50 are given from the MCU management function 60 via a separately provided MCU management node 83. It may be in the form of sending.

Next, the MCU management function 60 will be described in detail with reference to FIG. FIG. 4 is a functional block unit of the MCU management function 60. As shown in FIG. 4, the MCU management function 60 includes a conference reservation information receiving unit 61, an infrastructure information receiving unit 62, an MCU selection calculation unit 63, and an edge server / application control transmitting unit 64.

The conference reservation information receiving unit 61 receives conference reservation information from the user. The conference holding reservation information includes information such as a conference start / end time, a maximum number of participants, and a list of participating terminals.

The infrastructure information receiving unit 62 provides resource information such as a network topology, information for specifying the position of each terminal 10, information for specifying the position of each server infrastructure, and available bandwidth of each network section to the network management node 81. Receive from. In addition, the server infrastructure management node 82 receives information on resource information (usable CPU / memory / GPU / storage, etc.) of each regionally distributed edge server 40.

The MCU selection calculation unit 63 optimally balances network cost, server infrastructure resources, and conference quality when holding a conference based on a part or all of the conference reservation information and infrastructure resource information. One or more MCU50s such as are calculated. The details of the processing in the MCU selection calculation unit 63 will be described later.

In the edge server application control transmission unit 64, a single virtual conference room is held from the user terminal 10 by the selected MCU 50 in cooperation with another MCU 50 based on the optimum calculation result calculated by the MCU selection calculation unit 63. A control instruction is sent to the edge server 40 corresponding to the selected MCU 50 and the MCU 50 so as to behave as if it were performed. The control instruction is transmitted to the edge server 40 and the MCU 50 via the server infrastructure management node 82. When the MCU 50 operates as an application on the edge server 40 as a virtual server infrastructure, the edge server application control transmission unit 64 deploys an instance of the MCU application for each virtual server infrastructure, and each deployed MCU is Instruct the cooperation between application instances, create a virtual conference room, call the conference participation terminal as soon as the conference is held, disconnect at the end of the conference, end the virtual conference room, delete unnecessary MCU application instances, etc. Is sent via the server infrastructure management node 82.

As described above, in the case where the instruction to the MCU 50 among the control instructions is directly transmitted from the MCU management function 60 without going through the server infrastructure management node 82, as shown in FIG. 5, the edge The server / application control transmission unit 64 includes an edge server control transmission unit 64a and an application control transmission unit 64b. Further, as shown in FIG. 6, the instruction to the MCU 50 among the control instructions is the same in the case of the mode in which the instruction to the MCU 50 is transmitted via the MCU management node 83.

Next, an example of the flow from the conference reservation to the conference holding and the conference end in the multipoint conference system according to the present embodiment will be described with reference to FIG. 7. Here, for the sake of simplicity, the procedure for holding a single conference will be described.

The MCU management function 60 first accepts a reservation for a conference to be held (step S1). Next, the MCU management function 60 calculates the optimum arrangement of the MCU 50 (step S2). Here, the optimum placement calculation process of the MCU 50 is calculated back from the time required for the calculation and the MCU instance deployment immediately before the start of the conference, more specifically, so that the virtual conference room can be used at the conference start time. Start at the right time. The details of the optimum placement calculation process will be described later.

Next, the MCU management function 60 sends a control instruction to the edge server 40 and the MCU 50 to deploy the MCU instance to the MCU position selected in the previous step and to link each MCU (cascade connection, etc.) (step S3). ..

A meeting is held by the above processing (step S4). At the start of the conference, the MCU 50 calls the participating terminals 10 as necessary (callout). When the conference is completed, the MCU 50 disconnects the connected participating terminal 10 and performs the conference termination process (step S5). In addition, the MCU management function 60 sends a control instruction to the edge server 40 to delete the MCU instance of the MCU 50 that held the conference.

Next, an example of the flow from the conference reservation to the conference holding and the end of the conference when there are a plurality of conference reservations scheduled to be held at the same time will be described with reference to FIG. Here, a case where n conferences are reserved at the same time will be described.

The MCU management function 60 first accepts a reservation for a conference to be held (step S11). Next, the MCU management function 60 calculates the optimum arrangement of the MCU 50 for each of the meetings 1 to n, and stores "cost (described later)" as an evaluation index of the optimum arrangement (step S12).

Next, the MCU management function 60 performs the same cost calculation based on the information of all the terminals participating in any of the conferences 1 to n (step S13). However, the candidate location for MCU deployment is limited to the location selected in the previous step. Then, the MCU management function 60 compares the total cost calculated in the previous step with the cost calculated in this step, and determines the final arrangement. In addition, as another embodiment of this step, a method of subtracting the resources occupied by holding the conference 1 to calculate the optimum deployment after the conference 2 and once disregarding the resources of other conferences at the same time at this stage. The method of calculating is conceivable.

Next, the MCU management function 60 sends a control instruction to the edge server 40 and the MCU 50 to deploy the MCU instance to the MCU position selected in the previous step and to link each MCU (cascade connection, etc.) (step S14). ..

A meeting is held by the above processing (step S15). At the start of the conference, the MCU 50 calls the participating terminals 10 as necessary (callout). When the conference is completed, the MCU 50 disconnects the connected participating terminal 10 and performs the conference termination process (step S16). In addition, the MCU management function 60 sends a control instruction to the edge server 40 to delete the MCU instance of the MCU 50 that held the conference.

Even if the meeting holding time is the same, the ending time differs depending on the meeting. Therefore, as a variation of the embodiment, a method of performing cost comparison in consideration of time can be considered. In addition, exception handling such as not performing the optimum placement calculation in a meeting of less than 15 minutes can be considered.

Next, the details of the MCU optimum arrangement calculation process in step S2 and step S12 will be described with reference to FIG. In this example, it is assumed that the information that is a part of the information necessary for the optimum arrangement calculation process and is not frequently changed is held in advance by the MCU management function 60. Such information may be acquired periodically or in response to an information update trigger at another timing, instead of requesting each time the optimum placement calculation is performed. Examples of such information include the network topology, the location of the edge server 40, the route cost of each network section, the usage cost of each edge server 40, and the like.

As shown in FIG. 9, the MCU management function 60 inquires the network management node 81 about the position of the conference participation terminal 10 (step S21). Here, the network management node 81 is an identifier (private IP address or E) on the VPN (Virtual Private Network) of the conference participating terminal 10 obtained from the conference reservation information and the conference service component (for example, Gatekeeper in the case of H.323). .164 number, etc.) and the infrastructure side identifier (NGN (Next Generation Network), in-network IPv6 address, line number, etc., mobile network, terminal identifier, GTP (GPRS Tunneling Protocol) tunnel ID, etc.) The position of the conference participating terminal 10 on the network topology is returned to the MCU management function 60 with reference to the correspondence of the above.

Next, the MCU management function 60 requests the network management node 81 for available bandwidth information for each network section at that time point (step S22).

Next, the MCU management function 60 requests the server infrastructure management node 82 for available resource information of the edge server 40 at that time point (step S23).

The MCU management function 60 executes the optimum placement calculation process based on the information acquired in the above steps and the information acquired / held in advance (step S24).

Next, an example of the detailed processing in step S24 will be described with reference to FIG. Here, as shown in FIG. 2, it is assumed that the network topology of the relay network 30 has a tree structure in which the other network 70 side is the upper layer and the terminal 10 side is the lower layer.

The MCU management function 60 performs the following processing for each layer in order from the highest layer (steps S31 to S37). In the processing of each layer (layer i), first, the maximum value J of the number of edge servers (j) in the layer i is checked from the network topology information acquired in advance as described above (step S32). Next, the MCU management function 60 lists (j = 0 to J) the edge server list of the layer i for which resources are available (step S33). Here, if there is no usable edge server in the layer i, the process is moved to the next lower layer (steps S34, S36, S37).

Next, assuming that the MCU management function 60 arranges the MCUs in all the locations of _{Node i-0 to Node i-J (which can be used),}
(A) Sum of route costs from each terminal (Endpoint ₀ to N) to the nearest MCU (the shortest reachable when tracing the network topology in the upstream direction) (b) Sum of edge server operating costs (however, however) Edge servers without conference traffic distribution and edge servers with 1 input and 1 output are not counted)
Is calculated in consideration of the weight and added up (step S35). Then, the MCU management function 60 stores the total value as Cost [i]. Strictly speaking, it is possible to calculate the cost of a plurality of patterns even in the same layer, such as which MCU is the parent MCU (or one MCU is newly arranged in the upper layer to be the parent MCU). In that case, the one with the lowest cost may be substituted into Cost [i] as the representative cost value of the relevant layer. Here, the parent MCU is an MCU that is the center of control in a system (for example, a cascade-connected MCU group) in which one virtual conference is held by coordinating two or more MCUs. Means. When three or more MCUs are cascaded in a tree-like topology, the MCU located at the root of the tree generally operates as the parent MCU. (However, in this embodiment, the MCU that performs tree-like cascade connection is used as an example, but it is possible to consider a system in which a plurality of MCU groups that cooperate with each other cooperate on an equal footing without the concept of a parent MCU / child MCU. Note that) The above processing is performed in each layer toward the lower level (steps S36 and S37).

Next, when all of Cost [0] to Cost [I] are empty values, the MCU management function 60 gives a service provision impossible error flag and ends the process (steps S38 and S40). On the other hand, in the MCU management function 60, when there is a value in any of Cost [0] to Cost [I], the MCU arrangement pattern is the smallest of Cost [0] to Cost [I] which is not an empty value. Is determined as the optimum arrangement (steps S38 and S39).

Next, a specific example of the optimum arrangement calculation process described in detail with reference to FIG. 10 will be described. The following are examples of information used in the calculation.

-Network topology-Participant location information-Edge MCU operation cost (values that are balanced with the route cost are set in advance. Weighting may be different for each layer, and "edge of areas where maintenance and operation such as remote islands are required" There may be forms such as "the cost of using the server is high" and "when a special event is being held, the cost of using the edge server in the vicinity of the area is temporarily set high".)
-Available bandwidth of each network section-Available resource information of the edge server, estimated usage cost when the MCU is operated on a certain edge server (used to judge whether the edge server can be used)
-Conference terminal bandwidth-Bandwidth between MCUs-Upper limit of the number of MCU cascade connection stages-Upper limit of the number of MCUs-Weighting coefficient of the route cost of each network section (the higher the rank, the larger the coefficient (higher cost), etc.)
-Minimum number of users (for example, "2" or more because there is no point in operating an MCU that can accommodate only one terminal (1 input / 1 output))

In this specific example, the following conditions are assumed for the sake of simplicity.

-Network topology: Tree configuration, 4 hierarchies (i = 0-3) (see Fig. 11)
-Participant location information: As shown in Fig. 11. For convenience, the names of the terminals participating in the conference are set to terminals 1 to 8. ・ Edge MCU operating cost: For simplicity, it is not weighted and is uniformly set to “4” ・ Maximum allowable band of a specific route: The bottleneck in this example It is assumed that it does not exist. ・ Available resource information of the edge server: It is assumed that all edge servers have sufficient available resources. ・ Terminal bandwidth: Uniformly "1 Mbps". In addition, the route cost when passing through a route without a weighting coefficient is set to a cost of "1" per 1 Mbps.-MCU band: After the received video data is synthesized in the MCU, the band is uniformly set to one screen (1 Mbps). Suppose that it is compressed and delivered to other MCUs and subordinate terminals ・ MCU cascade connection stage number upper limit: “2” ・ MCU number upper limit: no upper limit ・ Route cost weighting coefficient of each network section: (hierarchy) 0 (re-upstream): 4, layer 1: 3, layer 2: 2, layer 3: 1 (see FIG. 11)
・ Minimum number of users: “2”

Based on the above assumptions, the optimum placement calculation process described in detail with reference to FIG. 10 is executed. First, the cost calculation in layer 0 will be described with reference to FIG. Since there is only one node in layer 0, Node _0-0 , the total cost when the MCU is deployed at this position is calculated. As a result, as shown in FIG. 12, Cost [0] becomes 84.

Next, the cost calculation in the layer 1 will be described with reference to FIG. In layer 1, the total cost when the MCU is deployed at the positions of _{Node 1-0} and Node _{1-1 is calculated.} As a result, as shown in FIG. 13, Cost [1] becomes 64.

Next, the cost calculation in the layer 2 will be described with reference to FIG. In layer 2, the total cost when the MCU is deployed at the positions of _{Node 2-0 to Node 2-3 is calculated.} However, since _{only one conference participation terminal is accommodated under Node 2-1 and the} node has one input and one output, there is no point in operating the MCU on the node. Therefore, Node _2-1 is not used. Then, the participating terminals under _{Node 2-1} are other nodes of the same layer that can be reached in the shortest time, in other words, other nodes of the same layer that can be accommodated at the lowest cost, instead of _{Node 2-1.} It will be accommodated in Node _2-0 (see * 1 in FIG. 14).

Further, FIG. 14 describes a case where the core MCU (referred to as “parent MCU”; the other MCUs are referred to as “child MCU”) among the plurality of MCUs is Node _{2-0 (FIG. 14).} See 14 * 2). Although the description is omitted in this example, strictly speaking, _{the route cost between MCUs when each of Node 2-2} and Node _2-3 is a parent MCU is also calculated and compared in size. As a result, as shown in FIG. 14, Cost [2-1] becomes 70.

Next, in FIG. 15, in layer 2, the participating terminals under _{Node 2-1} that are not used are the nodes of the other upper layer that can be reached at the shortest, in other words, the smallest, instead of _{Node 2-1. It will be accommodated in Node 1-0} , which is another node of the same layer that can be accommodated at cost (see * 1 in FIG. 15). Further, the parent MCU is not selected from the MCUs in the same layer, but _{is separately arranged in Node 1-0} one layer above (see * 2 in FIG. 15). As a result, as shown in FIG. 15, Cost [2-2] becomes 68.

Next, in FIG. 16, in layer 2, the participating terminals under _{Node 2-1} that are not used are the other nodes of the same layer that can be reached at the shortest, in other words, the smallest, instead of _{Node 2-1. It will be accommodated in Node 2-0} , which is another node of the same layer that can be accommodated at cost (see * 1 in FIG. 16). Further, the parent MCU is not selected from the MCUs in the same layer, but _{is separately arranged in Node 1-1} one layer above (see * 2 in FIG. 16). As a result, as shown in FIG. 16, Cost [2-3] becomes 63.

By the above processing, Cost [2-3] is the smallest among each Cost [2-1] to Cost [2-3] for layer 2, so 63 is assigned to Cost [2] as the representative cost of layer 2. To do.

In this example, the calculation process example for the layer 3 is omitted. Since Cost [2] is the smallest cost among the costs Cost [0] to Cost [2] of the layers 0 to 2, the MCU arrangement configuration when the cost is calculated is stored in the memory and the calculation is completed.

Another example of the MCU optimal placement calculation step will be described with reference to FIG. For example, if the MCU expansion cost is set high or the conference participating terminals have a large regional bias, it is possible to deal with variations in the best MCU placement hierarchy according to the regional dispersion of the participating terminals. It is preferable to do so. That is, it is preferable to look at not only the total cost but also the partial cost increase / decrease. In the example of FIG. 17, the case where the conference participating terminals (terminals with hatches in the figure) are regionally biased is shown.

in this case,
・ Calculation of layer 0 ・ Calculation _{under layer 1: Node 1-1} , calculation under layer 1-2, ..., calculation under layer 1-k ・ Layer 2: Node _1-1-1 to _{1-1 The} calculation is performed in the step of confirming whether the route cost between −l + Node _1-1 to Node _{1-1-x has increased.} It is preferable that it is judged that the previous layer is advantageous depending on the MCU expansion cost, and that only some branches are allowed to return to the layer.

Another example of the MCU optimal placement calculation step will be described with reference to FIG. In this example, it is assumed that the MCU is first deployed at all the confluence points, the MCU operating points are gradually excluded, and the arrangement where the sum of costs is more optimal is searched for. The optimum solution differs depending on conditions such as cost weighting, the upper limit of the number of MCUs, and the upper limit of the number of multiple stages of MCUs.

In this example, there may be MCU arrangement pattern for ² 11 = ways 2048 to explore in brute (strictly Furthermore, also the parent of the same MCU arrangement - also necessary to consider the pattern of the child MCU relation). Therefore,
・ Preferentially exclude 1-in-1-out ・ Preferentially exclude MCUs in the lower layer (low NW cost) and have a small capacity, etc., and reduce unnecessary calculations as much as possible and calculate efficiently. More developmentally, it is conceivable to use machine learning together to accumulate performance information of efficient omission patterns and reflect it in the next calculation.

Another example of the MCU optimal placement calculation step will be described with reference to FIGS. 19-21. This example is an example when two meetings are held at the same time. In this case, as shown in FIG. 19, the MCU optimum arrangement is first calculated for the first conference by the algorithm detailed with reference to FIG. Similarly, as shown in FIG. 20, the MCU optimal placement is calculated for the second conference by the algorithm detailed with reference to FIG. Next, (A) the first meeting and the second meeting are regarded as one, and the same cost calculation is performed, and (B) the MCU installation candidate of the second meeting is limited to the place calculated as a result of the first meeting. And try to calculate. Then, (C) the cost of (A) and (B) is compared and the configuration of the minimum cost is adopted. In the examples of FIGS. 19 to 21, the cost of (A) is lower, but depending on the weight of the server operating cost, the total can be achieved by synergizing with the MCU installation location that is expected to be launched at another conference. Costs may be reversed.

Although the embodiments of the present invention have been described in detail above, the present invention is not limited thereto. For example, in the above embodiment, the conference system has been described as an application example of the multipoint communication system, but it can also be applied to other systems. Examples of applications include a web conferencing system, a real-time video distribution system, and a multicast distribution system for file transfer.

Further, in the above embodiment, as the management function of the MCU 50, an implementation example using the MCU management function 60, the network management node 81, the server infrastructure management node 82, and the MCU management node 83 has been described, but each function is centralized or distributed. The present invention can be carried out in other implementation forms such as conversion.

Further, in the above embodiment, the case where the network topology of the relay network 30 has a tree structure has been described, but the present invention can be implemented even with other network topologies. In this case, a cost calculation algorithm suitable for the network topology may be adopted.

10 ... Terminal 20 ... Access network 30 ... Relay network 35 ... Relay device 40 ... Edge server 50 ... MCU application 60 ... MCU management function 61 ... Conference reservation information reception unit 62 ... Infrastructure information reception unit 63 ... MCU selection calculation unit 64 ... Edge Server / application control transmission unit 64a ... Edge server control transmission unit 64b ... Application control transmission unit 70 ... Other network 81 ... Network management node 82 ... Server infrastructure management node 83 ... MCU management node

Claims (8)

A multipoint communication system that performs group communication between multiple terminals housed in a network.
A multi-point connection function that operates on multiple servers geographically distributed so that the network topology is a tree structure with the terminal side as the lower level in the network.
Prior to the start of the group communication, the route section is weighted for each layer of the tree structure based on the network configuration information including at least the position information of the terminal and the position information of the server and the resource information of the network and the server. The usage cost is calculated from the weighted route cost and the operating cost of the server on which the multipoint connection function operates for each arrangement pattern of the multipoint connection function that is set and assumed to be used in the group communication in the network. A multipoint connection function selection means for selecting one or a plurality of the multipoint connection functions related to the optimum arrangement pattern, using the arrangement pattern for which the calculation and the minimum utilization cost have been calculated as the optimum arrangement pattern.
When there are multiple multipoint connection functions selected by the multipoint connection function selection means, the multipoint connection functions can be linked with each other, and a single group communication can be performed from the terminal while straddling the multiple multipoint connection functions. A multipoint communication system characterized in that it is equipped with a multipoint connection function control means for controlling each multipoint connection function so as to behave as if it is implemented. When a plurality of group communications are performed in parallel, the multipoint connection function selection means calculates the optimum arrangement pattern for each group communication, and is common to the calculated optimum arrangement pattern for each group communication. A small usage cost was calculated by comparing the usage cost calculated assuming that the included multipoint connection function is used with the usage cost related to the optimum arrangement pattern calculated by considering each group communication as a unit. Select one or more of the multipoint connection functions related to the placement pattern
The multipoint communication system according to claim 1. The multipoint connection function selection means arranges the multipoint connection function in each layer in order from the highest layer of the tree structure, at least at least one of the points where the traffic related to group communication joins in the layer. , The usage cost is calculated for the placement pattern of the multipoint connection function in the layer, and the placement pattern for which the minimum usage cost is calculated is set as the optimum placement pattern.
The multipoint communication system according to claim 1. The multipoint connection function selection means calculates the usage cost in the lower layer while keeping the multipoint connection function of the layer in place for some branches in the arrangement of the multipoint connection function for each layer.
3. The multipoint communication system according to claim 3. The multipoint connection function selection means calculates the usage cost based on an initial arrangement pattern in which the multipoint connection function is arranged at all points where traffic related to group communication joins, and then some multipoints are obtained from the initial arrangement pattern. The usage cost is calculated based on the placement pattern excluding the connection function, and the placement pattern for which the minimum usage cost is calculated is set as the optimum placement pattern.
The multipoint communication system according to claim 1. The multipoint connection function selection means preferentially excludes the multipoint connection function having a small number of subordinate terminals.
5. The multipoint communication system according to claim 5. It is a multipoint communication method that performs group communication between multiple terminals housed in a network.
Multiple servers that operate the multipoint connection function are geographically distributed and deployed in the network so that the network topology has a tree structure with the terminal side as the lower level.
Prior to the start of the group communication, the multipoint connection function selection means has a hierarchy of the tree structure based on network configuration information including at least the position information of the terminal and the position information of the server and resource information of the network and the server. The weighting of the route section is set for each, and the weighted route cost and the server on which the multipoint connection function operates are set for each arrangement pattern of the multipoint connection function assuming use in the group communication in the network. The usage cost is calculated from the operating cost of the above, and the arrangement pattern for which the minimum utilization cost is calculated is set as the optimum arrangement pattern, and one or more of the multipoint connection functions related to the optimum arrangement pattern are selected.
When the multipoint connection function control means has a plurality of multipoint connection functions selected by the multipoint connection function selection means, the multipoint connection functions are linked with each other, and the terminal is straddled over the multiple multipoint connection functions. A multipoint communication method characterized by controlling each multipoint connection function so that it behaves as if a single group communication is being carried out. A multipoint communication program comprising operating a computer as the multipoint connection function selection means and the multipoint connection function control means according to claim 1.

Images