JP2002135330A - Inter-node control system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an inter-node control system that warrants a bandwidth of control communication in the case of the control communication between node units using an IP network and sets the bandwidth used for the control communication between the node units depending on the processing state. SOLUTION: The inter-node control system warrants the bandwidth between the node units MGC and MG making communication in the IP network, periodically transmits a monitor packet by return and changes the setting of the communication bandwidth between the node units when a mean value of delay times exceeds a preset threshold value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an IP (Internet Protocol).
col) It relates to communication between node devices in a network. In particular, the present invention relates to an inter-node control method capable of efficiently and robustly controlling communication between node devices. The present invention provides a service (Voice over) for controlling between devices in a PSTN / IP integrated network in which a public telephone network and an IP network are integrated, and connecting a telephone network and an IP network to convert voice into IP packets and communicate. IP, VoIP)
It is suitable for use in control of devices in the above and transmission of control signals in a multiprocessor system that performs parallel processing by distributing a plurality of processors on the Internet.

[0002]

2. Description of the Related Art In recent years, an integrated PSTN / IP network integrating a public telephone network and an IP network has been studied. In particular, a VoIP service using a network architecture of a separated bearer transfer system and control system has been studied. Has been attracting attention.

FIG. 9 is a diagram for explaining a conventional VoIP service. A plurality of node devices are provided in the IP network, and TCP / IP (Transfer Control Protocol) is used as a communication protocol between the nodes.
l Protocol / Internet Protocol, Transmission Control Protocol /
Internet Protocol) is used. In FIG.
As a node device, two media gateways MG for connecting an IP network and a public telephone network PSTN, and a media gateway control MGC for controlling the connection are shown. IP
The network also has a gatekeeper GK that translates between IP addresses and telephone numbers. The gatekeeper GK may be provided in the media gateway control MGC. FIG. 9 shows the operation system OpS for controlling the IP network as one block, but this is actually realized by distributed control of each node device. A dedicated network called a common line network is used for setting up a call connection in the public telephone network, whereas a connection between the public telephone network and a terminal between IP networks is used.
In the VoIP service, it is necessary to set up a call connection from the public telephone network side to the media gateway control MGC and the media gateway MG via a common line network.
In this case, the control between the media gateway control MGC and the media gateway MG performs communication on the IP network and sets up a call connection. For more information on VoIP services, see Miyake et al., Prospects and Standardization of Wide Area IP Network Technology, IEICE, Vol. 83, No. 4, pp. 263-275, 2000.4.

[0004] Since IP packet-based communication using the TCP / IP protocol is a best-effort communication, the AT
Unlike guarantee type communication such as M, communication quality (QoS) is not guaranteed. Therefore, TCP
If an IP packet-based communication using the / IP protocol is used as a control communication route in a multiprocessor system including a media gateway MG and a media gateway control MGC, problems such as a delay in packet arrival time occur. For example, when generating a large amount of control signals from an arbitrary node to a certain node,
It is conceivable that the original control becomes impossible when the arrival delay time is significantly increased. In order to prevent such delay, RSVP (Resource Reservation Protocol)
l, network bandwidth reservation protocol). RSVP is a normal LAN or WAN
However, there is a problem that charging cannot be performed when used for IP-based communication, but there is no charging problem for control communication between the media gateway MG and the media gateway control MGC.

[0005]

However, in the conventional RSVP, when there are a large number of hosts, it is difficult to determine which host has priority and secures the bandwidth. This is because in normal IP-based communication, it is difficult to prioritize each process performed by the majority of users. In addition, RSVP tries to secure the bandwidth periodically, but there is no function to secure the bandwidth according to the processing status of the system, and there is no function to dynamically change the bandwidth in response to sudden traffic fluctuation. Absent.

The present invention solves such a problem and makes the control communication between the node devices robust, so that the band used for the control communication between the node devices is dynamically changed according to the status of the processing between the node devices. It is intended to set and efficiently control communication between node devices.

[0007]

An inter-node control system according to the present invention includes a plurality of node devices for distributing and processing control necessary for communication between each other in an IP network. In an inter-node control method including a unit for securing a band necessary for communication with another node device, the unit for securing the band periodically returns a monitoring packet between the node devices performing communication, and delays the monitoring packet. Means for setting and changing the communication band between the node devices when the average value of the time exceeds a predetermined threshold value.

[0008] Communication between the node devices is preferably performed by TCP / UDP / IP. It is preferable that the means for securing a band secures a band necessary for communication in advance by RSVP and dynamically control the band by the means for changing the setting.

The means for changing the setting is based on the average value of the delay time of the monitoring packet and the sum of the number of packets and the packet length transmitted from each node device per unit time.
Means for determining the required bandwidth can be included.

[0010] Between the node devices, there is provided a first route for main communication in which a band required for communication is secured in advance, and a second route for sub-communication set separately from the first route. The means for changing the setting may include means for setting the second path for main communication and securing the required band obtained by the obtaining means when it is determined that the first path has a failure. Can be.

In this case, when the first route has a failure and the second route is used for main communication, the setting change means may change the third route different from the first route and the second route. Means for setting for secondary communication can be included. If the required bandwidth obtained by the obtaining means cannot be secured in the second route, a third route capable of securing the required bandwidth different from any of the first route and the second route is used for main communication. Means for setting can also be included.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram showing a configuration example when the present invention is used for a VoIP service. The existing public telephone network PSTN and the IP network are connected by a media gateway MG, in which the bearer service is converted from the public telephone network PSTN to the IP network or from the IP network to the public telephone network PSTN. A media gateway control MGC is further provided in the IP network, and performs connection control for making a call connection from an actual terminal to the terminal. That is, connection control is performed by MG-MGC communication via the IP network. Further, the gatekeeper GK provided in the IP network has a database for converting an IP address and a telephone number, and when a line is connected from the public telephone network PSTN to the IP network, or from the IP network to the public telephone network PSTN, A call connection is made based on the information of the gatekeeper GK. This gatekeeper GK is Media Gateway Control MGC
Or may be provided separately.

In this configuration, when a large number of calls are accepted, a large amount of control signals are transferred between the media gateway control MGC and the media gateway MG, which may cause a large delay. Therefore, in such a case, it is effective means to guarantee the band between the media gateway control MGC and the media gateway MG by using a band guarantee protocol such as RSVP. However, when a relatively small number of media gateway control MGCs are controlled between a large number of media gateways MG, a large number of calls are generated in the bandwidth between each media gateway control MGC and the media gateway MG. Assurance of the amount assumed as described above wastes resources. It is assumed that a large number of calls occur when some event occurs. Therefore, the bandwidth guarantee between the media gateway control MGC and the media gateway MG only needs to be prepared when necessary. Therefore, the monitoring packet is periodically looped back between MGC and MC, and when the average value of the delay time exceeds a predetermined threshold, MGC-MC
Change the communication band between MCs. As a result, a necessary band is always secured, and robust control communication between MGC and MC is realized.

FIG. 2 shows a specific processing flow. First,
k-th media gateway _{MG k (k = 1,2, ...} ) is in receiving the information added to the network media gateway control MGC, RS between pre MGC-MG _k
The bandwidth a _k is secured by the VP (S1). This bandwidth a _k can be arbitrarily determined based on the number of lines that can be accommodated by the media gateway MG _k and the service class.
Can be set with a policy. After this, MGC to MG _k
, A monitoring packet is sent periodically at a constant interval t, and is returned from MG _k to MGC. At this time, the round trip time RTT (Round Trip Time) via MGC-MG _k -MGC
Is measured (S2). From this measured value, the number of accepted calls or the congestion state of the MG is estimated. The average value of the measured RTT is the maximum specified delay time

[Outside 1] Is exceeded (S3),

(Equation 1)MGC to MG_kFor the reset message (Reestablish me
ssage) and send the band b _kTo (S4).

FIG. 3 is a diagram for explaining the message transfer between the media gateway control MGC and a plurality of media gateways MG. FIG. 4 is a diagram showing a message transfer for securing a bandwidth by a normal RSVP without considering the round trip time RTT. Data Flow, FIG. 5 shows a message transfer data flow for securing a bandwidth in consideration of RTT.

For RSVP, Media Gateway MG
Sends a path message to the media gateway control MGC periodically, and the media gateway control MGC sends a reservation message (Res
ervation message) to maintain the band securing state (FIGS. 3 and 4). Further, according to the present invention, the monitoring message is periodically returned (FIGS. 3 and 5), and when the average value of the RTT exceeds the maximum specified delay time, a re-set message is sent from the media gateway control MGC to the media gateway MG. And pass / pass as with conventional RSVP
Exchange messages and reservation messages to change bandwidth.

In a normal RSVP, a path message and a reservation message are periodically transmitted and received, and the band securing state between the MGC and the MC can be maintained. This regular MGC-
When the state maintenance between MCs is used, the setting can be returned to the initial state a _k when the reserved band b _k is not required.
Therefore, according to the present invention, the media gateway MG
Even if a delay between MG and MGC occurs because a large number of calls are accepted, the control communication can be made robust.

Generally, a typical large-scale network (WAN)
In the case of using RSVP, it is difficult to determine the bandwidth allocation to each user if the number of users is very large, whereas by using the extended RSVP applying the present invention to RSVP,
Since all the processing contents of each media gateway MG are grasped by the media gateway control MGC, it is effective even when performing band control. In other words, this is effective in that the media gateway / control MGC operates a band with a policy. Also, the fact that charging, which is generally pointed out as a problem of RSVP, cannot be performed, is not necessary for control communication in the MGC-MG.

The periodic monitoring packet and the extended RSVP according to the present invention can be used for logically duplicating a path between nodes and changing a path when a failure occurs. Before describing such an embodiment, a description will be given of how to improve the reliability of the system by duplicating a path.

In a network architecture of a separate bearer transfer system and control system proposed at present as an integrated PSTN / IP network, when each node is connected by an IP network, a control plane for controlling between MGC and MG is provided. (C-plane: Control
There are a plurality of logical planes such as a management plane (M-plane: Mantenance plane) for performing operation management of the MGC and MG, and a user plane (U-plane: User plane) for performing bearer transfer. In order to implement these logical aspects on an IP network, a method of realizing each aspect with an independent IP network and a method of sharing these multiple logical aspects with one IP network are considered. In the former case, since a plurality of networks are independently formed, the cost of the entire system increases. On the other hand, in the latter case, since the MGC and the MG transmit and receive packets without connection, the MGC-MG is equivalent to a single connection, and securing the reliability of the entire system including the MGC and the MG becomes an issue. .

In order to ensure the reliability of the system, in the conventional exchange, each functional unit such as a switch unit and a line accommodating unit is connected in a duplex manner by a system bus. On the other hand, the failure rate of a single-purpose general-purpose router is generally inferior to the failure rate of a conventional duplex switch in many cases. In order to use an IP network in which a plurality of routers intervene and to ensure carrier-class reliability in MGC-MC control, it is necessary to provide two routes between MGC and MG.

As a method of forming two routes between the MGC and the MG, it is conceivable to form the control surface by duplicating the other surface independently of the other surfaces. In this case, the reliability of the system is close to the reliability of the conventional switch, but since the network is formed independently, the cost of the entire system increases.

On the other hand, when the return of the periodic packet and the extended RSVP in the present invention are used, the same I
In a network configuration superimposed on the P network, MGC-MG
The route can be divided into two routes. That is, MGC-MC
Logically set two paths for main communication and sub-communication between them, detect a failure from the delay of the periodic packet, and
VP secures bandwidth and changes route. As a result, the two paths can be operated optimally according to the state of the network, and the reliability of the entire system can be improved by logically using the two paths.

FIG. 6 shows an embodiment in which the route between the MGC and MG is logically duplicated, and FIG. 7 shows a control flow of the route change.

In this embodiment, the extended RSVP
A different first route and a second route are set between the MGC and the MG. Here, the first route is a main communication route, that is, a main operation route, and the second route is a sub communication route, that is, a sub operation route. Then, the status of the two routes is monitored by periodic monitoring packets. If the first route is disconnected or the bandwidth becomes extremely short, the first route is determined to be in a failure state, and operation is performed using the second route that has been secured in advance. If the second route is disconnected or the bandwidth is insufficient, the second route is determined to be in a failure state, and a new route is set. The disconnection of the route can be determined by the monitoring packet not returning to the transmission source. An extreme shortage of the band can be determined by an abnormally long delay time in which the monitoring packet is returned and returned.

When the first route is determined to be in a failure state, and operation is to be performed on the second route that has been secured in advance, it is first determined whether or not a required bandwidth can be allocated to the second route. The determination can be made based on the average value of the delay time of the periodic monitoring packet and the sum of the number of packets to be transmitted by each node and the packet length (message amount).

When the required bandwidth is secured for the second route, a third route completely independent of both the first route and the second route is searched. Then, the first route is released, and the third route is set as a sub-operation route.

If the required bandwidth cannot be secured by the second route, a search is made for a third route that is completely independent of the first route and the second route and that can secure the required bandwidth. Then, the first route is released, the third route is set as the main operation route, and the second route is set as the sub-operation route as it is.

FIG. 8 shows the evaluation results of the embodiment shown in FIGS. 6 and 7 and the conventional IP network regarding the inoperability rate of the system when one MGC and a plurality of MGs are connected. In this evaluation, the number of connecting routers between MGC and MG was set to 1, 5 respectively,
The MGC, MG, and router unavailability rates are 10 ^-8 , 10 ^-7 , 1
^0-4 was assumed.

FIG. 8 shows that the system non-operating rate in this embodiment is 10 ^-6 or less, which is different from the case where MGC and MG are operated by a mere IP network, in that a significantly superior system non-operating rate can be satisfied. Understand. Therefore, even when a control command such as an architecture of a bearer transfer system / control system separation type is realized in the IP network, the reliability of the carrier class can be ensured.

[0031]

As described above, according to the present invention,
Communication control between node devices on an IP network such as the Internet can be efficiently performed. In particular, the bandwidth is set in consideration of the processing amount between the node devices when the telephone network is connected to the Internet. It is possible.

For example, when the present invention is applied to a system for connecting a telephone network to the Internet, when communication between node devices is equal at the time of initial setting of the entire system, by using an extended RSVP, It is possible to equally allocate the bandwidth between the node devices. Also, in normal times, when it is expected that a large amount of communication between any node devices occurs, by measuring the delay time of the monitoring packet and securing the necessary bandwidth from the result, There is an effect that the entire system can be operated efficiently.

Further, by providing two paths logically, the reliability of the whole system can be improved, and by using two paths according to the state of the network, the whole system can be operated efficiently. .

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration example when the present invention is used for a VoIP service.

FIG. 2 is a diagram showing a specific processing flow.

FIG. 3 is a view for explaining message transfer between an MGC and an MG.

FIG. 4 is a diagram showing a message transfer data flow for securing a bandwidth by normal RSVP without considering RTT.

FIG. 5 is a diagram showing a message transfer data flow for securing a bandwidth in consideration of RTT.

FIG. 6 is a diagram showing an embodiment in which a path between MGC and MG is logically duplicated.

FIG. 7 is a control flowchart of a route change.

FIG. 8 is a diagram showing a non-operating rate of the system when one MGC is connected to a plurality of MGs.

FIG. 9 is a diagram illustrating a conventional VoIP service.

[Explanation of symbols]

 PSTN public telephone network MG media gateway MGC media gateway control GK gatekeeper OpS operation system

Claims (7)

[Claims] 1. An Internet Protocol (IP) network comprising a plurality of node devices for distributing and processing control required for communication with each other, each of which is required for communication with another node device. In the inter-node control method including a means for securing a suitable bandwidth, the means for securing the bandwidth periodically returns a monitoring packet between the communicating node devices, and the average value of the delay time is predetermined. An inter-node control method comprising means for changing the setting of the communication band between the node devices when the threshold value is exceeded. 2. The inter-node control system according to claim 1, wherein communication between the node devices is performed by a transmission control protocol / user datagram protocol / internet protocol (TCP / UDP / IP). 3. The node according to claim 1, wherein the means for securing a bandwidth secures a bandwidth required for communication in advance by a network bandwidth reservation protocol (RSVP), and dynamically controls the bandwidth by the setting changing means. Control method. 4. The setting changing means includes means for obtaining a required bandwidth from an average value of the delay time and a total number of packets and packet lengths transmitted from each node device per unit time. Item 2. The inter-node control method according to Item 1. 5. A first path for main communication in which a band required for communication is secured in advance and a second path for sub-communication set separately from the first path are provided between the node devices. The means for changing the setting, when it is determined that there is a failure in the first path, the second path for main communication,
5. The inter-node control method according to claim 4, further comprising means for securing a required band obtained by said obtaining means. 6. The method according to claim 1, wherein the setting change unit is configured to, when the first route has a failure and the second route is used for main communication,
6. The inter-node control method according to claim 5, further comprising means for setting a third route different from the first route and the second route for sub-communication. 7. The setting change means, when there is a failure in the first route and the required bandwidth obtained by the obtaining means cannot be secured on the second route, the first route and the second 6. The inter-node control method according to claim 5, further comprising means for setting a third route different from any one of the routes and capable of securing the required band for main communication.