JP3280765B2 - Congestion control method in packet network - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a congestion control method in a packet network, and more particularly, to a method in which a packet exchange in a network rapidly increases in a short time and a packet switch falls into a congestion state. The present invention relates to a congestion control method in a packet network which is performed to prevent a failure state such as a system failure from occurring or a congestion state of the exchange from spreading to the entire packet network.

[0002]

2. Description of the Related Art First, before describing a congestion control method performed in a conventional network, a packet network configuration and packet communication which are premised will be described. less than,
A packet network, which is a premise of the present invention, will be described by taking an ISDN packet network as an example. Figure 4 shows the ISD
FIG. 2 is a block diagram illustrating an N packet network configuration.

[0003] In the figure, 1 is an originating ISDN packet terminal, 2 is an originating ISDN exchange, 3 is an originating packet exchange, 4 is a packet relay network, 5 is a destination packet exchange, 6
Is a destination ISDN exchange, 7 is a destination ISDN packet terminal, 11 and 12 are ISDN channel control procedures, and 13 and 1
5 is X. 25 and 14 are packet relay protocols (in-network control + X.25). Here, X. 25 is a CCITT (International Telegraph and Telephone Consultative Committee) recommendation number, which defines interface conditions between a packet-type terminal and a packet-switched network.

Referring to FIG. In the case of an ISDN packet network, the subscriber terminal 1 is an ISDN subscriber
It is connected to the DN switch 2 and is accommodated in the packet switch 3 through the ISDN switch 2. Packet switch 3
Is connected to the packet relay network 4 and is connected to another packet switch 5 via the packet relay network 4 to enable packet communication between terminals.

[0005] Communication between packet switches via the packet relay network 4 is based on X.25. A packet relay protocol 14 is used in which a packet network control procedure is added to 25 protocols. In addition, between the subscriber terminal and the ISDN exchange and the packet exchange, ISDN channel control procedures 11 and 1 are performed.
2 is used to control the ISDN communication channel.

[0006] As described above, the packet switch comprises:
X.11, which is one of the CCITT recommendations and defines a packet-type terminal interface. The present invention provides a packet communication service between terminals using the 25 protocol for a communication procedure.

FIG. 5 is a diagram for explaining a communication phase of ISDN packet communication. In the figure, 1 is an originating ISDN packet terminal, 2 is an originating ISDN exchange, 3
Is the originating packet switch, 4 is the packet relay network, 5 is the destination packet switch, 6 is the destination ISDN switch, 7 is the destination I
SDN packet terminal; 11A to 14A: ISDN channel setting phase; 15A to 18A: ISDN channel release phase; 25CR packet, 22-
2 is X. 25CN packet, 22-3 is X.25. 25CA packet, 22-4 is X.25. 25CC packet, 33-1
33-2 is X.25. 25DT packet, 44-1 is X.25 DT packet. 25
CQ packet, 44-2 is X. 25 CI packets, 44
-3 to 44-4 are X. 25 CF packet.

[0008] Here, X. The packet communication conforming to the 25 protocol has three communication phases: a call setup phase, a data transfer phase, and a call release phase. Further, in the case of ISDN packet communication, X.400 is used. In addition to the 25 communication phases, there is a communication phase for ISDN channel control. X. Before the 25 call setup phase,
The ISDN channel setting phase (11A to 14A)
X. After the 25 call release phase, there are ISDN channel release phases (15A to 18A), respectively.

Hereinafter, the ISDN packet communication between terminals will be specifically described with reference to FIG. When the subscriber terminal 1 accommodating the ISDN network starts ISDN packet communication, first, a channel for performing packet communication with the packet switch 3 through the ISDN switch 2 by the ISDN channel control protocol (11A, 12A). Make settings.

[0010] When the setting of the communication channel is completed, the originating terminal 1 sends the X.509 to the originating packet switch 3. A 25CR packet (22-1) is transmitted. X. from the originating subscriber.
The originating packet switch 3, which has received the 25CR packet, receives the destination terminal 7 and receives it.
Toward X. 25CN packet (22-2), and the destination packet switch 5 A 25CN packet is transmitted to the destination terminal 7.

[0011] In the receiving terminal 7, when responding to the call setting from the calling side, X. 25 CA packet (22-3) is transmitted to the destination packet switch 5, and the destination packet switch 5 A packet (22-4) of 25 CC is transmitted to the originating side packet exchange 3, and the originating side packet exchange 3 similarly transmits the X.25 packet. 25. Send a 25CC packet to the calling subscriber 1 to complete the call setup connection; The 25 call setup phase ends.

X. When the 25 call setup phase ends, X25. DT packet (33-1, 33-
2) Transfer to the data transfer phase. When releasing the connection, the call release packets (44-1 to 44-4) are transmitted between the terminals as in the case of the call setup.
To the call release phase, and release the connection.

[0013] Here, X. The 25CR packet is
X. Specified by the CCITT recommendation. 25 is a call setup packet of the X.25 protocol. 25CN is an incoming call packet, X.25CN. 25CA is an incoming call reception packet, X.25CA. 25C
C is a connection completion packet. What is 25CQ?
Recovery request packet; 25CI is a disconnection instruction packet, X.25CI. 25CF is a disconnection confirmation / recovery confirmation packet.

Incidentally, in the case of PVC (permanent virtual call), normal communication is performed only in the data transfer phase.
A PVC setup / release packet corresponding to call setup / release in SVC (switched virtual call) is used.

Here, the above SVC and PVC will be described briefly. A call of a selective connection performed by selecting an arbitrary partner instead of a fixed communication partner and setting a communication link each time is called SVC, and the communication partner is always fixed to one specific partner and the communication link is set. Is a PVC with a fixed connection in which is always set. Therefore SV
C means a selective connection call, and PVC means a fixed connection call.

[0016] The premise explanation has been completed above, and the congestion control method performed in the conventional network will be described below. In a packet network, for example, a sudden increase in intra-network packet traffic in a short time causes an increase in a processing load in a packet switch, a failure in normal packet transfer processing, and a congestion state. There is.

The congestion of the packet switch includes processor congestion caused by overload processing of the processors and buffers constituting the packet switch, and buffer congestion caused by insufficient buffers. A major cause of processor congestion is a sharp increase in call setup processing. This is because when a new call setup is performed, the processor is overloaded, such as a terminal number translation process, a communication path setting process, and a rationality determination process.

On the other hand, the main cause of buffer congestion is a rapid increase in data transfer processing. This is considered to be because the packet switching system is a store-and-forward system, and the packet of the subscriber terminal is temporarily stored in a buffer to perform communication processing. At the time of buffer congestion, when a packet is transferred, a transfer delay occurs, which greatly affects the service.

When a packet switch enters a congestion state, the possibility of occurrence of a failure such as a system down of the packet switch due to an increase in processing load increases, and at the same time, the influence of congestion generated in the packet switch causes other effects. There is a danger that it will spread to the packet switch and the subscriber line consolidation stage, eventually causing the entire packet network to fall into a congested state, and normal processing as a network cannot be performed.

Here, the subscriber line concentrator means ISD
ISDN to accommodate subscribers in N packet service
An exchange, which has a function of setting an ISDN channel with a subscriber and a function of transferring a packet from the subscriber transmitted on the set channel to the packet switch.

Hereinafter, a congestion control method performed in a conventional network will be described. The congestion control methods include a congestion detection method and a congestion restriction / cancellation process. The packet switch detects congestion according to each of the processor usage rate and the buffer usage rate, and reduces the load according to each usage rate. Therefore, a congestion restriction process is performed.

In general, congestion control is performed step by step by assigning priorities to communication types between terminals. A conventional congestion detection method when two types of communication are provided will be described with reference to FIG. FIG. 6 is a flowchart showing a conventional method of congestion state management.

Referring to FIG. The congestion detection includes threshold values A1 and A2 for the processor usage rate and the buffer usage rate.
(Steps S1 and S11) are provided. If the usage rate exceeds the threshold, primary congestion (steps S2, S1
As 2), the congestion restriction processing is performed on the communication having the lower priority (1-1, 1-2). Further, the usage rate increases and the thresholds B1 and B2 (where B1> A1, B2> A
2) When exceeding (steps S3 and S13), congestion restriction processing is performed on communication with a high priority as secondary congestion (steps S4 and S14) (2-1, 2-).
Step-by-step methods such as 2) are used.

The conventional congestion restriction processing will be described below with reference to FIG. FIG. 7 shows the same I as shown in FIG.
FIG. 2 is a block diagram showing a network configuration of SDN packet communication. Please refer to FIG.

When the packet switch 3 enters the congestion state, the packet switch 3 recovers the congestion state and prevents the congestion state from spreading to the other packet switches 5 and the subscriber line concentrator (ISDN switch) 2. The SVC call during communication is disconnected, the PVC call is reset (it is a fixed connection call, so it is reset instead of being disconnected), and the other exchanges are used to restrict communication to the congested exchange. Perform congestion notification processing (11-1 to 11-2).

In the subscriber line concentrating stage 2 which has received the notification, the packet processing from the subscriber terminal 1 to the congested packet switch 3 is regulated at the layer 2 frame level (22). Is carried out. On the other hand, in the other packet switch 5, in response to packet communication from the subscriber terminal 7 to the packet switch 3 in a congested state, the SVC call during communication is disconnected, and the PVC call is reset (44). ).

When the congestion regulation process is successful and the congestion state of the packet switch in the congested state is alleviated, the congestion regulation is gradually released in the exchange and the packet exchanges to the other packet exchanges and the subscriber concentrator stage are performed. Thus, a notification of a restriction release request for packet communication for the exchange is sent. At the same time, when the other packet exchange and the subscriber line concentrator receive the restriction release notification from the exchange, the restriction release processing is performed stepwise.

[0028]

In the above congestion restriction processing, as described above, different congestion detection targets are determined as primary and secondary congestion in accordance with the priority of the communication type when congestion is detected. Although the restriction process is performed on the restriction target, the restriction process itself performs the congestion restriction process uniformly irrespective of the packet type. As a result, all the calls in communication are disconnected. Therefore, when congestion occurs, it is of course impossible to remedy a call during communication, and when a large number of terminals that have been disconnected have re-issued the call all at once, the exchange must be sent to the exchange again. Therefore, a processing load due to the re-calling is applied, and there is a risk that the congestion state is promoted.

The congestion restriction processing in the conventional method is performed by using the X.3 of Layer 3 Regulations are implemented at 25 levels. Specifically, X. In response to the 25 call setup packet (X.25CR), a call setup rejection packet (X.25CI) is returned,
The establishment of the packet call is refused, and the SVC call is disconnected and the PVC call is reset for the call in communication. However, a recovery request packet, a disconnection instruction packet, a disconnection confirmation / recovery confirmation packet (X.25CQ, X.25
25CI, X. 25CF) is a call release request and is unlikely to cause a processor overload, so no restriction is imposed.

If the restriction is uniformly applied to the packet of the subscriber as in the conventional method, the call during communication is cut off when congestion occurs, and the serviceability to the subscriber is reduced. Not only is bad, but the subscriber who is disconnected from the current call recalls, for the call setting process,
In addition, processor utilization may increase and congestion may be exacerbated. Many packet terminals are automatic calls, and these terminals automatically re-issue calls when the current call is disconnected. As described above, the possibility of congestion may increase. Is enough.

Therefore, an object of the present invention is to solve the above-mentioned problems, and to reduce the congestion state due to the overload of the packet switch to call setup.
When caused by a large number of packets
Is intended to improve the serviceability to subscribers without disconnecting a call during communication, and therefore, in a packet network in which there is no danger of causing an increase in processor utilization and congestion due to automatic repetition of subscribers. An object of the present invention is to provide a congestion control method. Congestion due to overload of packet switch
The condition was caused by a burst of data transfer packets
If the data packet or call is already
An object of the present invention is to provide a congestion control method in a packet network, in which a switch regulates a signal requesting release .

[0032]

In order to achieve the above object, according to the present invention, a congestion state due to an overload of a packet switch,
When the congestion is detected by congestion detection means as congestion due to call setting request signal traffic for a selective connection to be performed while setting a communication link each time by selecting an arbitrary partner instead of a fixed communication partner, or Is always fixed to one specific partner, as a congestion due to the setting request signal traffic for fixed connection always set the communication link, when detected by the congestion detection means,

In the call control phase, the exchange rejects the signal for newly requesting the call setting for selective connection or the signal for requesting the setting for fixed connection. In the subsequent transfer phase, the exchange does not regulate data packets already in communication or signals requesting call release. Another packet
The congestion condition due to overload of the switch
Call control when caused by concentration
Call setup for selective connection
Signals that require fixed connection settings.
The switch does not reject the setting and the call control
In the transfer phase following the
The switch does not respond to signals that require
We decided to do it.

In the following, the case of ISDN packet communication will be described as an example in order to describe specific means of realization. However, in the case of packet communication, the same method can be used. Like the 25 protocol,
The generality of the present invention is not lost if the communication is performed using a protocol having a call setup / call release phase and a data transfer phase in the in-channel.

When the terminal is performing packet communication,
For example, when a simultaneous call from a plurality of terminals and a simultaneous call for a specific terminal occur in a short time, and when a fixed connection is frequently set, the processor of the packet switch includes the
SDN channel setting processing; Overloading occurs due to the 25 call setting process, and a congestion state occurs in which transfer processing of not only call setting packets but also other packets cannot be performed normally. Also, when the amount of data packet transfer from the terminal increases rapidly, the buffer capacity of the packet switch becomes insufficient, and the state becomes congested.

As described above, the congestion of the packet switch includes:
There are two types of processor congestion due to processor overload and buffer congestion due to buffer shortage. The relationship between the congestion factor and the ISDN packet communication phase will be described below.

In the ISDN packet communication phase, IS
DN channel setup / release phase; 25 call settings /
Release phase and X. There are 25 data transfer phases,
X. ISDN channel setup phase and X. In the case where the outgoing / incoming process is performed suddenly and frequently in the 25 call setting phase, and If the fixed connection setting process occurs frequently in the 25 data transfer phase, the processor usage rate increases due to the processing load on the processor, which causes processor congestion.

On the other hand, X. In the 25 data transfer phase, if the data transfer increases rapidly, a large amount of buffers are used for packet transfer, which causes buffer shortage and causes buffer congestion. Here, it can be seen that the factors of the processor congestion and the buffer congestion are independent, and congestion regulation targets necessary for alleviating the congestion are different.

According to the above description, processor congestion and buffer congestion are caused by independent factors, respectively. When congestion occurs, effective congestion for the congestion factor is set to alleviate the congestion state. You need to take control.

More specifically, call setting processing such as simultaneous calling and the setting of fixed connection are rapidly increased, and when processor congestion occurs, calling and setting of fixed connection are controlled and the processor is controlled. When it is necessary to reduce the burden, the data transfer increases sharply, the buffer required for packet transfer becomes insufficient, and when buffer congestion occurs, the packet transfer itself is regulated and the buffer shortage is eliminated. Need to be done. The present invention is different from the conventional method in that when processor congestion and buffer congestion occur, different congestion restriction processes are performed.

[0041]

According to the conventional method, effective regulation operation cannot be performed for congestion factors because congestion factors different from each other are unified. In the control method, by performing a restrictive operation that is effective against the cause of congestion, early recovery from the congestion state can be achieved. Further, in the congestion control method according to the present invention, when congestion occurs due to a factor such as a simultaneous call, only a new call is restricted, so that a call during communication is not disconnected as in the conventional method. . As a result, not only is the serviceability of the subscriber improved, but there is no recall due to disconnection of the call during communication, and congestion due to the recall can be prevented.

[0042]

FIG. 1 is a diagram for explaining a method for explaining an embodiment of the present invention. In the figure, 101 is a system monitoring unit, 102 is a system management function unit, 103-1 is a congestion control function unit (own processor), 103-2 is a congestion control function unit (other processor), and 104-1 is a packet transfer process (Own processor), 104-2 is a packet transfer processing unit (other processor), (11) and (12) are monitoring processes of processor usage and buffer usage, (13-1) is a congestion regulation processing command, 13-2) congestion notification processing of another processor, (14-1) and (14-2) congestion regulation processing instructions, (15) congestion notification processing to another module 105,
It is.

FIG. 2 is a flowchart showing a specific algorithm of the congestion control method according to the present invention. This flowchart includes steps S1 to S12. This will be described below with reference to FIGS.

The system management function unit 102 constantly monitors the processor usage rate and the buffer usage rate through the system monitoring unit 101 ((11), (12)). further,
The system management function unit 102 determines the congestion state by comparing the processor usage rate and the buffer usage rate with a preset congestion threshold (step S in FIG. 2).
5).

As a result of the determination, if the congestion state is found, the congestion control function unit 103-1 is notified of the congestion state and instructed to execute the congestion restriction operation (13-1). ). In the system management function unit 102, the congestion control function unit (103
-2) is notified that a certain processor is in a congestion state.

Further, the congestion control function unit 103-1
The processor usage rate and the buffer usage rate are determined, and the congestion restriction operation is determined (step S7 in FIG. 2). When the processor usage rate is equal to or more than the threshold value α and less than β and the buffer usage rate is equal to or more than γ and less than η (route 7-1 in FIG. 2),
Regulates only new call setup and fixed connection setup packets,
The call during communication (data packet) is determined to be a congestion restriction process in which the call is not restricted (however, α <β, γ <η).

If the processor usage rate is equal to or greater than β and the buffer usage rate is equal to or greater than η (route 7-2 in FIG. 2), it is determined that a congestion control state in which all packets are regulated. Further, the congestion control function unit 103-1 instructs the packet transfer processing unit 104-1 to reject a new call setting packet and a fixed connection setting packet from the terminal in the case of the former congestion restriction state. It instructs it to return and to perform the process of making it unregulated for other packets (14-1).

On the other hand, in the latter congestion restriction state, the communication from the terminal is instructed to disconnect the SVC call and reset the PVC call. In the packet transfer processing unit 104-1, the congestion control function unit 103-1
The congestion restriction operation is performed as instructed. In the congestion control function unit 103-1, the packet transfer processing unit 104
A congestion notification is transmitted to the other exchange (module) 105 through -1 (15).

In the congestion control function unit 103-2 of the non-congestion processor in the own exchange, when the congestion processor is notified from the system management function unit 102 (13-2), the packet transfer processing unit 104-2 is notified. A command is issued to implement the congestion control (14-2). Packet transfer processing unit 10
In 4-2, a congestion restriction process is performed. The system management function unit 102 monitors the congestion state at a certain period as described above, and performs congestion control based on the obtained congestion state.

An example of a specific congestion control process in the ISDN packet network will be described below with reference to FIG. FIG. 3 is an explanatory diagram illustrating an example of a specific congestion restriction process in the ISDN packet network.

Here, it is assumed that the ISDN channel setting and the X.25 are set.
For example, in a packet switch having a multiprocessor configuration in which different processors are provided with a call control function such as 25 call setup / call release and a packet communication processing function between terminals, the usage rate of a multiprocessor having a call control function is reduced. ISDN channel setting from the subscriber side and the inter-office side and X. The following describes an example of a congestion control method in which 25 call setting packets are restricted and data packets are not restricted.

FIG. 3 shows the configuration of an ISDN packet communication network and the configuration of a processor in a packet switch. 3, reference numeral 3 denotes an originating packet switch, 3-1 denotes a service control processor (having a call control function), 3-2.
Is a packet communication control (accommodating subscriber) processor, 3-3
Is a relay network connection processor, 11-1 to 11-3 are congestion notifications, 22 is an ISDN channel setting message, and 33 is X. 25 is a call setup packet (from the subscriber side), and 44 is a call setup packet (from the inter-office side).

In FIG. 3, when a service control processor 3-1 having a call control function enters a congestion state, an ISDN channel setting message and an X.509 message are transmitted to alleviate congestion. 25 call setup packet and X.25 It is necessary to regulate 25 fixed connection setting packets. First, the service control processor 3-1 having the call control function includes the subscriber accommodation processor 3-2 and the relay network connection processor 3-3.
To notify that the processor is in the congestion state.

ISDN channel setup message (22)
About X. In the subscriber accommodation processor 3-2 having the 25-protocol function, the ISDN channel control protocol is not terminated, so that it is impossible to regulate the protocol. Therefore, the processor 3-1 having the call control function
Then, an ISDN channel control protocol is once communicated with the subscriber terminal 1 to notify the ISDN exchange 2 that the new terminal for setting a new ISDN channel from the subscriber terminal 1 is regulated. I was notified
In the SDN exchange 2, the new ISD from the subscriber terminal 1
The N channel setting message (22) is restricted from rejecting new channel setting.

X. from the subscriber terminal 1 X.25 call setup packet and X.25 25 fixed connection setting packet,
Contrary to the SDN channel setting message, the ISDN exchange 2 uses X. No regulation is possible because the 25 protocols are not terminated. Subscriber accommodation processor 3-2
At X. from the subscriber terminal 1. 25 call setup packet and X.25 For the 25 fixed connection setting packet (33), a call setting rejection and a restriction process of rejecting the fixed connection setting are performed. However, X. 25 data packets and X.25. Since the 25 call release packet is not considered to be a factor of congestion, no restriction processing is performed.

X. from the relay network 4 side. X.25 call setup packet and X.25 Regarding the fixed connection setting packet (44), the relay network connection processor 3-3, which terminates the packet relay network protocol, sends the X.25 fixed connection setting packet (44). Restriction processing for 25 call setting rejection and fixed connection setting rejection is performed. However, X. Concerning the 25-call setup packet, the congestion notification is sent to the opposing packet switch 5, and the X.25. 25 call setup packet and X.25
25 For the fixed connection setting packet, call setting rejection and
It is also possible to execute a restriction process of rejecting the fixed connection setting. As in the case of the subscriber accommodating processor, the X.300. 25 data packets and X.25. No restriction processing is performed on the 25 call release packet.

As described above, in the embodiment, the case where the processor utilization rate of the processor having the call control function is increased has been described. However, it is to be added that the other processors can perform the regulation processing in the same manner. Keep it.

[0058]

According to the present invention, when the packet switch enters a congestion state, not only can effective regulation processing for the congestion factor be performed, but also only new call setting packets and fixed connection setting packets can be transmitted. Having a congestion restriction state in which data is restricted and data packets are unrestricted, and by executing the processing, not only the call during communication is rescued and the serviceability to the subscriber is improved, but also the re-call due to the call disconnection. Prevents the congestion from being promoted.

In the above description, the case of ISDN packet communication has been described as an example. However, in the case of packet communication, it can be realized by the same method.
Call setup /
Finally, it should be added that the generality of the present invention is not lost if the communication uses a protocol having a call release phase and a data transfer phase.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of a method for explaining an embodiment of the present invention.

FIG. 2 is a flowchart showing a specific algorithm of the congestion control method according to the present invention.

FIG. 3 is a block diagram showing a configuration of an ISDN packet communication network and a processor configuration in a packet switch.

FIG. 4 is a block diagram showing an ISDN packet network configuration.

FIG. 5 is a diagram illustrating a communication phase of ISDN packet communication.

FIG. 6 is a flowchart showing a conventional method of congestion state management.

FIG. 7 is a block diagram showing a network configuration of ISDN packet communication.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Originating ISDN packet terminal, 2 ... Originating ISDN exchange, 3 ... Originating packet switch, 4 ... Packet relay network, 5
... Destination packet exchange, 6 ... Destination ISDN exchange, 7 ...
Incoming ISDN packet terminal, 101 ... System monitoring unit,
102: System management function unit, 103-1: Congestion control function unit (own processor), 103-2: Congestion control function unit (other processor), 104-1: Packet transfer processing unit (own processor), 104-2 ... Packet transfer processing unit (other processor), 105 ... other module

Continuation of front page (72) Inventor Takaaki Fukuda 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation (72) Inventor Akira Hamada 1-16-1 Uchisaiwaicho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone (72) Inventor Toshihiko Shimamori 5-7-1 Shiba, Minato-ku, Tokyo NEC Corporation Inside (72) Inventor Akihiko Suwa 5-7-1 Shiba, Minato-ku, Tokyo NEC Corporation ( 56) References JP-A-4-217144 (JP, A) JP-A-2-104150 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04L 12/56

Claims (2)

(57) [Claims] 1. A congestion state due to an overload of a packet switch causes a high utilization rate of a processor constituting the packet switch.
Low buffer usage required for packet transfer
From those caused by the frequent occurrence of call setup packets
Therefore , as a result of congestion due to call setting request signal traffic for a selective connection to be performed while selecting a communication partner instead of a fixed communication partner and setting a communication link each time, the packet switching is performed.
A setting request signal for a fixed connection in which a congestion state due to overload of the switching equipment is detected by the congestion detecting means, or a communication partner is always fixed to one specific partner and a communication link is always set. As traffic congestion,
When a congestion state due to overload of the packet switch is detected by the congestion detecting means, a signal for newly requesting call setting for selective connection or a signal for requesting setting for fixed connection in a call control phase. On the other hand, the exchange rejects the setting, and in the transfer phase following the call control phase, the exchange does not regulate data packets already in communication or signals requesting call release. Congestion control method in a packet network. 2. A congestion state due to overload of a packet switch.
However, the utilization rate of the processors that make up the packet switch is low.
High buffer usage required for packet transfer
From, caused by the concentration of data transfer packets
And thereby , as buffer congestion due to an increase in data transfer processing,
Congestion detection means by congestion due to overload of packet switch
In the call control phase, a new call setup for selective connection
Signal to request the setting for fixed connection
The switch does not reject the setting and the call control
During the transfer phase following the
Switching packets or signals requesting call release
Congestion in packet networks characterized by regulation
Control method.