JP3495058B2 - Load balancing method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a load distribution type data exchange system in which sequentially generated calls are distributed to and processed by a plurality of processors, and the calls newly generated while maintaining the balance of the processing amount among the processors. The present invention relates to load balancing that determines a processor that performs processing.

[0002]

2. Description of the Related Art In a data communication system such as ISDN, various kinds of information such as voice, image, data, etc. are transmitted, and a variety of communication speeds, communication times and communication forms are required. In addition, the information described above includes information that occurs continuously and information that occurs in bursts. In a data exchange that processes information having such a property as a call in packet or cell units, In order to smoothly absorb the maximum peak point of call volume in the system and to enable expansion and maintenance operation that adapts to the increase of subscribers and calls after the system is installed, the calls that occur sequentially are distributed to multiple processors. A load balancing method for processing by processing is applied.

FIG. 7 is a diagram showing a configuration example of a data exchange to which a load balancing system is applied. In the figure, the data exchange 70 is disposed between the communication path 72 ₁ to 72 _n that are individually connected to the terminal 71 ₁ -71 _n, a relay circuit 73 connected to other data communication networks.

In the data exchange 70, the communication paths 72 ₁ ...
72 _n are connected to the plurality of exchange processing devices 7 via the path setting unit 74.
Connected to the corresponding inputs / outputs of 5 ₁ to 75 _N , these switching processors 75 ₁ to 75 _N are bus-coupled to the central controller 76 and the interoffice controller 77. The input / output of the inter-station control unit 77 is connected to the relay line 73, and the output of the central control unit 76 is connected to the control input of the path setting unit 74.

In the data exchange having such a structure,
The control device 76 is the exchange processing device 75. ₁~ 75_NOne of
It has a pointer in software that always points to one. Na
In the following, the value indicated by such a pointer is exchanged.
Device 75₁~ 75_NNumber attached to the reference number of
It is indicated by "1" to "N".

The central control device 76 receives a new packet or cell (hereinafter,
Any of these is simply called a "call". ) (Below,
When the word "call" is used, it is called "occurrence". ) Then, out of the terminal 71 ₁ -71 _n on the basis of a predetermined procedure, the terminal to be a destination of the call (hereinafter, referred to as "called terminal".)
Ask for. Further, the central control device 76 corresponds to the communication path corresponding to the called terminal and the exchange processing device corresponding to the above-mentioned pointer value (hereinafter referred to as "distributed processing device").
A path setting command for instructing connection with
And the call is transferred to the corresponding distributed processing device to instruct the start of processing, and the value of the pointer is updated based on a predetermined algorithm.

The path setting unit 74 connects the distributed processing device and the receiving terminal via a communication path, and the distributed processing device receives the call transferred from the central control unit 76 as described above. And performs call processing based on a predetermined communication procedure.

As described above, in the data exchange 70, the processing to be performed for the sequentially occurring calls is distributed to the exchange processors 75 ₁ to 75 _N under the control of the central control unit 76, so that the data exchange 70 operates like a non-load distributed data exchange. Since it is not necessary to install a high-speed processor that adapts to the number of terminals to be accommodated and the maximum call volume, it is possible to inexpensively and reliably by installing the number of exchange processing devices that corresponds to the number of terminals and call volume. Data is exchanged, the amount of processing is averaged, and the hardware and software configurations are standardized.

In recent years, many enterprises have adopted a center-end communication mode in which a single large host computer and many terminals are connected via a plurality of communication paths. The data exchange of is often applied to such a communication mode.

[0010]

By the way, in a data exchange adopting such a conventional load balancing system, the amount of transmission information contained in a call and the amount of processing actually required are increased every time a call occurs. Regardless of the situation, the switching processing device that should process the next call is selected by updating the above-mentioned pointer value in a cyclic manner. Therefore, in each exchange processing device, such a variation in the processing amount for each call is often ignored and the load distribution is biased.

Further, the person in charge of maintenance and operation grasps the amount of call processing held in each exchange processing device and the amount of call in each communication path during the other busy hours after the start of the service, and determines the predetermined amount. The bias of load distribution must be corrected by controlling the operation of the central control unit 76 by performing the operation described above, which hinders the efficiency of maintenance operation.

Further, as described above, in the data exchange adopted in the center-end communication mode, the same data is exchanged by the periodical simultaneous communication due to the business days, business hours, and business characteristics of enterprises and other subscribers. When communication that transmits and receives a large amount of information at a time is performed in a burst, a specific exchange processing device falls into a shortage of processing amount and memory area, and the delay time of processing increases and a predetermined communication control procedure is performed. Communication quality may be deteriorated due to communication regulation, data discarding, and other processing based on the above. Further, such a congestion state has also been a factor that has a similar adverse effect on the processing of other calls that should be performed in parallel by the same exchange processing device.

It is an object of the present invention to provide a load distribution system that can evenly balance the loads of the processors.

[0014]

FIG. 1 is a block diagram showing the principle of the invention described in claim 1. In FIG. The present invention performs call setting processing for a callee terminal and performs load balancing for each call that occurs in each logical multiplexing unit in a data exchange network that divides transmission information and logically multiplexes and transmits the data. Based on the control device 11 that once forms a communication path between the load distribution destination and the terminal, and the call whose load is distributed by the control device 11 based on the communication procedure facing the destination terminal via the communication path. Exchange processing device 13 ₁ for performing call processing
In load-sharing data exchange and a to 13 _N,
The control device 11 includes, for each of the exchange processing devices 13 _{1 to} 13 _N ,
Of the storage means 15 for holding the total processing amount of calls distributed by the control device 11 and the exchange processing devices 13 _{1 to} 13 _N ,
The one corresponding to the minimum total processing amount stored in the storage unit 15 is set as the distribution destination, and the minimum total processing amount is the line speed of the communication path or the increment of the throughput caused by the call whose load is distributed to the communication path. The distribution control means 17 for adding either one is provided.

FIG. 2 is a block diagram showing the principle of the invention described in claim 2. The present invention performs call setting processing for a destination terminal for a call that occurs in each unit of logical multiplexing in a data switching network that divides transmission information, logically multiplexes and transmits the data, and balances the load. A control device 11 forming a communication path between a load distribution destination and a terminal, and a call processing for a call load-balanced by the control device 11 based on a communication procedure facing a destination terminal via a communication path. In the load balancing type data exchange equipped with the switching processing devices 13 _{1 to} 13 _N , the control device 11 has an operation mode storage means 21 that holds in advance the operation modes of the individual terminals that can be the distribution destinations of the load distribution. , Exchange processing equipment 13 ₁ to 1
Of the 3 _N , the distribution destination storage means 23 for preliminarily retaining the exchange processing devices that can be distribution destinations in association with the operating modes,
Distribution control means 25 for acquiring the operation form of the destination terminal from the operation form storage means 21, and setting the exchange processing device held in the distribution destination storage means 23 corresponding to the operation form as the distribution destination selection range. It is characterized by having.

[0016]

According to the load balancing method of claim 1, the storage means
15 is the exchange processor 13₁~ 13_NOf calls distributed to
Of these, the total amount of
Retained by each exchange processing device, when a call occurs, the distributed control means
17 is the exchange processing device 13 ₁~ 13_NOf the above
The switching processor with the smallest processing volume is the destination of the call.
In addition, the total processing amount of the exchange processing device
The line speed of the communication path formed between the terminal and
Through caused by load distribution of calls corresponding to the communication path
Either one of the put increments is added.

Such an increase in the line speed and the throughput is compared with the number of unprocessed calls used in the prior art instead of these, and the load of the calls is distributed to the switching processors 13 _{1 to} 13 13. _{Since N} substantially indicates the amount of resources (including the amount of processing) required, the exchange processing device 13 ₁
The load distributed to 13 _N is accurately equalized.

According to the load balancing method described in claim 2,
The destination storage means 23 is the exchange processing device 13. ₁~ 13_NMinutes of
Switch processing equipment that can become a distribution destination to the operating mode of individual terminals
Corresponding and holding in advance, when a call occurs, distributed control means
The storage means 21 stores the operation form of the terminal of the destination of the call.
And all the exchanges that are compatible with the operating mode.
The conversion processing device is acquired from the distribution destination storage means 23 and is applicable.
The selection range of the call distribution destination.

Therefore, the control means 11 positively distributes the load generated by the terminals operating in the same form to a plurality of specific switching processors, so that all the calls regardless of the operating form of the terminal are different. In comparison with the conventional example in which the load is distributed to the switching processing devices 13 _{1 to} 13 _N , the simultaneous communication for transmitting and receiving a large amount of information in a burst at the same time due to the operation mode of the terminals that differ depending on the subscriber is predetermined. It is distributed only to the exchange processing devices that have been set up, and the spread range of congestion due to such simultaneous communication is suppressed to a certain range.

[0020]

Embodiments of the present invention will now be described in detail with reference to the drawings. In each of the embodiments described below, the hardware configuration is the same as that of the conventional example shown in FIG. 7, and the characteristic part relating to the present invention is the software of the central controller provided in place of the central controller 76. is there.

FIG. 3 is a diagram for explaining the operation of the embodiment corresponding to the invention described in claim 1. In the figure, in the present embodiment, a central control device 30 is provided instead of the central control device 76, and the present invention is applied to ISDN.

The central processing unit 30 includes a load distribution table 31 and a joining condition table 32 which are referred to by software. The load distribution table 31 shows the identification numbers “1” to “N” of the exchange processing devices 75 ₁ to 75 _N and the amount of load distributed to the exchange processing devices individually corresponding to these identification numbers, and The distributed load amount updated based on the algorithm described later is held. In the subscription condition table 32, the subscription condition consisting of the sum of the line speed and the contract throughput of all the contracted logical channels for the communication paths individually corresponding to all the accommodated terminals is the identification number of each communication path (here, it is simple. Therefore, it is equal to the sub-number given to the reference numbers of the communication paths 72 _{1 to} 72 _n .)
Are registered in this order in advance.

Regarding the correspondence relationship between this embodiment and the block diagram shown in FIG. 1, the central controller 30 and the path setting unit 74 correspond to the controller 11, and the exchange processor 71 ₁
.About.71 _N correspond to the exchange processing devices 13 _{1 to} 13 _N , the load distribution table 31 corresponds to the storage means 15, and the central control device 30 corresponds to the distribution control means 17.

The operation of this embodiment will be described below. When the interoffice controller 77 receives a packet from an adjacent station via the relay line 73 and recognizes that a call has occurred,
The contents of the packet are transferred to the central controller 30 (FIG. 3). The central control device 30 temporarily stores the contents of the packet transferred in this way, and thereafter, the exchange processing device 7
One of 5 ₁ to 75 _N (here, the switching processor 75 ₂ for simplicity) is randomly selected and the switching processor is requested to process the corresponding call (FIG. 3). .

The exchange processing device which has received such a request issues an incoming call request (FIG. 3), and the central control device 30 carries out an incoming call negotiation based on a predetermined communication procedure according to the incoming call request as in the conventional example. To do.

The central control device 30 inquires the destination terminal 71 _T whether or not to accept the incoming call in the course of the incoming call negotiation, and when the terminal permits the incoming of the corresponding incoming call, the load balancing table The minimum of the distributed processing amount registered in 31 (= ZZ (<... <YY <X
The identification number zz of the exchange processing device corresponding to (X) is obtained. The central control device 30 sets, via the path setting unit 74, a communication path connecting the exchange processing device 75 _ZZ corresponding to the identification number thus obtained and the destination terminal 71 _T ,
Also, the subscription condition corresponding to the terminal is acquired by referring to the subscription condition table 32. Further, the central control device 30 adds the identification number of the exchange processing device 75 ₂ to the joining condition thus obtained and sends it to the exchange processing device 75 _ZZ ((4) in FIG. 4)) ). Further, the central control device 30 obtains the sum of the line speed given by the subscription condition and the contract throughput of the contract logical channel, and when the former is lower than the latter, the line speed is the distributed load amount of the corresponding load distribution table 31. was added to, if the latter opposite falls below the former is added to the dispersion load of the load distribution table 31 to the appropriate sum contracts throughput contract logical channel.

The exchange processor 75 _ZZ sends a response completion notice to the exchange processor 75 ₂ based on the identification number added to the above-mentioned subscription condition (FIG. 3), and the exchange processor 75 ₂ responds to the notice. Terminate the temporary processing of the corresponding call.

The central control unit 30 informs the exchange processing unit 75 ₂ that the load of the corresponding call is distributed to the exchange processing unit 75zz.
Then, the contents of the above-mentioned packet are transferred to the exchange processor 75 _ZZ (FIG. 3). The exchange processor 75 _ZZ is
When receiving and recognizing such a notification, the packet transfer preparation is performed, and then the packet transferred from the central control device 30 is transferred to the terminal 71 _T (FIG. 3).

When the switching processor 75 _ZZ completes the transfer of the above-mentioned packet, it notifies the central controller 30 of that fact together with the identification information of the corresponding call. Central controller 30
When receiving and recognizing such a notification, it starts the path release process, and is identified by the above-mentioned identification information among the communication paths formed via the communication path setting unit 74 based on the procedure of the process. One of the distributed load amounts registered in the load distribution table 31 is released, and the subtraction process opposite to the above-described addition is performed on the distributed load amount identified by the same identification information.

As described above, according to the present embodiment, the load distribution table 31 accurately indicates the processing amount of each call with respect to the calls that are not individually distributed and processed by the switching processors 75 ₁ to 75 _N. The line speed or the sum of the contracted throughputs of the contracted logical channels is accumulated, and the newly generated call is sequentially distributed and processed by the switching processors whose accumulated value is the smallest.

Therefore, the exchange processor 75 ₁ ...
The load distributed to 75 _N is accurately averaged, and the bias in load distribution that occurs in the conventional example is significantly reduced. In the present embodiment, the size of the line speed of the communication path and the sum of the contract throughputs of the contract logical channels of the communication path is determined, and either one of them is added to or subtracted from the distributed load amount of the load distribution table 31. However, the present invention is not limited to such a method, and the line speed may be always added or subtracted, for example, as long as the reduction in load distribution accuracy is allowed.

FIG. 4 is a diagram showing an embodiment corresponding to the invention described in claim 2. In FIG. In the figure, in the present embodiment, the present invention is applied to a data exchange adapted to a packet type switching system by providing a central control device 40 instead of the central control device 76.

The central control unit 40 includes terminals 71 _{1 to} 71 _n.
To the communication path identification number, the call volume indicating the maximum communication volume per unit time during communication of each communication path,
A subscriber data table 41 in which a group number for identifying the operation mode of each terminal is registered, and identification numbers of all exchange processing devices corresponding to each group number and serving as distribution destinations of calls corresponding to the individual group numbers. , A call volume table 42 in which the total value of the throughputs of calls that are distributed and not processed in each exchange processing device is registered, and the identification of the exchange processing device connected via the communication path setting unit 74 for each communication path. A communication path management table 43 for registering numbers is provided.

Regarding the correspondence between this embodiment and the block diagram shown in FIG. 2, the central control unit 40 and the path setting unit 74 correspond to the control unit 11, and the exchange processing unit 71 ₁
.About.71 _N correspond to the exchange processors 13 _{1 to} 13 _N , the subscriber data table 41 corresponds to the operation form storage means 21,
The call volume table 42 corresponds to the distribution destination storage means 23, and the central controller 40 and the communication path management table 43 correspond to the distribution control means 25.

FIG. 5 is an operation flowchart of this embodiment. The operation of this embodiment will be described below with reference to FIGS. In the present embodiment, the characteristic feature of the present invention resides in the processing procedure described later in which the central processing unit 40 decides which of the switching processing devices 75 ₁ to 75 _{N is} the switching processing device as the distribution destination of the call. . Also, the central control unit 40, replacement processor 75 _1-7 recognizes an occurrence of a call
The point that the path release processing is performed after the calls corresponding to 5 _N are distributed is the same as in the conventional example and the embodiment shown in FIG. 3, and the switching processors 75 ₁ to 75 _N and the inter-office controller 77 are The processing to be performed is also the same as that of the conventional example, and thus the description thereof is omitted here. Furthermore, the terminals 71 ₁ to
71 ₃ is assumed to be housed in the data exchange according to the present embodiment, which is contracted by a single enterprise as a subscriber.

The central controller 40 has terminals 71 _{1 to} 71 ₁ .
₃ one of the (in this case, the terminal 71 _t for simplicity.) When an incoming call to it is detected that has occurred, as shown in FIG. 6 (a), the terminal 71 _t in CCITT Recommendation _Q. It sends incoming (SET UP) message based on 930 series, performs an incoming call negotiation by receiving a response (CONNECT) message terminal 71 _t has transmitted in response to that message.

The 2 bits of the bit numbers 1 and 2 of the third octet constituting such a response message are shown in FIG.
As shown in (b) by shading, the identification number of the communication path is included as the information channel selection information. Central controller 4
0 is the subscriber data table 4 based on the identification number.
1 to obtain the group number (=% 1) (FIGS. 4 and 5), and call volume table 42 based on the group number.
By referring to FIG. 4 and FIG. 5, the exchange processing devices belonging to the corresponding group have the smallest total call volume (here, for the sake of simplicity, the exchange processing device 75 ₂
And ) As the distribution destination of the corresponding call (see FIG. 5).
), And the identification number (= 2) of the exchange processing device.

Further, the central controller 40 refers to the communication path management table 43 based on the identification number of the communication path acquired from the response message as described above (FIG. 4).
By doing so, the identification number (= 2) of the corresponding exchange processing device is registered in the table (FIG. 4), and the exchange processing device and the terminal 7 of the callee are received via the communication path setting unit 74.
A communication path is formed with 1 _t (FIGS. 4 and 5).

Further, the central control unit 40 sets the total value (= 0) of the call volumes registered in the call volume table 42 corresponding to the identification numbers of the exchange processing devices of the distribution destinations to the subscriber data table 4
The call volume (= 5) registered together with the group number (=% 1) described above is added to 1 to distribute the call to the exchange processing device (FIG. 5), and conversely, it is activated when the processing of the call is completed. In the course of the path releasing process, the increment of the total value of the call volume corresponding to the above-mentioned addition is corrected, and the corresponding communication path formed via the communication path setting unit 74 is released.

As described above, according to the present embodiment, the calls originated simultaneously by a plurality of terminals belonging to the same subscriber such as a business entity are positively distributed to a plurality of exchange processing devices and processed. Even if a large amount of information is transmitted and received at the same time in a burst by regular simultaneous communication due to business days, business hours, and business characteristics of the business entity, the load is distributed to a specific exchange processing device. No bias occurs. Furthermore, since congestion and other ripples caused by the above-described burst communication are suppressed for the switching processing devices assigned to different groups, communication regulation, data discarding and other processing are performed by such congestion and communication is performed. Quality degradation is avoided.

In this embodiment, the call amount is registered in the subscriber data table 41, and the total value of the call amounts in the call amount table 42 is updated based on the call amount.
However, the present invention is characterized in that the range of selecting the exchange processing device at the distribution destination is individually set for each subscriber who operates the destination terminal. As an algorithm for finally selecting the exchange processing device of the load distribution destination in the above, any algorithm such as a round-robin method for cyclically updating the pointer indicating the load distribution destination may be used as in the conventional example.

Further, in this embodiment, subscribers of individual terminals
Although a single group number is given to the (communication path), the present invention is not limited to such a configuration, and for example, a plurality of different operations depending on the type of call originated by each subscriber and others. When taking a form, these operating forms may be registered in the subscriber data table 41 in a form that can be identified.

Further, in each of the above-described embodiments, the processing procedure for selecting the exchange processing device to which the call is distributed when the incoming call to the terminal occurs has been described. However, the present invention handles such an incoming call. Without limitation, even if an outgoing call occurs,
The same can be applied by recognizing the terminal of the caller or the communication path connected to the terminal based on the procedure of a predetermined call setting process in place of the above-mentioned incoming call negotiation.

In each of the above embodiments, I
The load distribution system according to claim 1 or 2 is applied to a data switch adapted to SDN or a packet switching system. However, the data distribution system is not limited to such a configuration and, for example, data adapted to a frame relay switching system or a cell relay switching system. It is similarly applicable to exchanges.

[0045]

As described above, according to the present invention, as a criterion for selecting the exchange processing device as the distribution destination of the generated calls, the line of the communication path formed between the exchange processing device and the destination terminal. Either the integrated value of the speed or the increment of the throughput caused by the load distribution of the call corresponding to the communication path is adopted, or the call is handled individually for the subscriber who operates the source or destination terminal. Set the exchange processing device that can be the target of load balancing.

That is, the distribution destination of the corresponding call is determined depending on the number of calls distributed to each exchange processing device, or the exchange processing device of the distribution destination is selected regardless of the difference in the destination subscribers. Compared to the conventional example, load distribution is achieved based on the amount of resources (including the processing amount) required by the switching processing device according to call load distribution, and the occurrence occurs according to the operation mode of each terminal. Since the simultaneous communication is distributed only to the predetermined exchange processing device, and the congestion spread range due to the simultaneous communication is suppressed to a certain range, the accuracy of the load distribution is improved and the load distribution is performed during the operation. The work of manually changing the ratio and the like becomes unnecessary.

Therefore, in the data exchange to which the present invention is applied, the quality of service can be made uniform and improved,
And the efficiency of operation and maintenance can be improved.

[Brief description of drawings]

FIG. 1 is a principle block diagram of the invention according to claim 1.

FIG. 2 is a principle block diagram of the invention described in claim 2.

FIG. 3 is a diagram showing an embodiment corresponding to the invention described in claim 1.

FIG. 4 is a diagram showing an embodiment corresponding to the invention described in claim 2.

FIG. 5 is an operation flowchart of the present embodiment.

FIG. 6 is a diagram illustrating incoming call negotiation.

FIG. 7 is a diagram showing a configuration example of a data exchange that employs a load balancing method.

[Explanation of symbols]

11 Control device 13 Exchange processor 15 storage means 17,25 distributed control means 21 Operation mode storage means 23 distribution destination storage means 30, 40, 76 Central control unit 31 load balancing table 32 Subscription condition table 41 Subscriber data table 42 Call volume table 43 Communication path management table 70 data exchange 71 terminals 72 Communication path 73 trunk line 74 Path setting section 75 Exchange processor 77 Inter-station control device

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Ikuo Kodama 1015 Kamiodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture, Fujitsu Limited (72) Inventor, Eiichiro Takahashi 1015, Kamikodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture, Fujitsu Limited ( 72) Inventor Satoshi Hinomura 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation (72) Inventor Akira Hamada 1-1-6 Uchiyuki-cho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation (72 ) Inventor Yutaka Yamazaki 1-7-12 Toranomon, Minato-ku, Tokyo Oki Denki Kogyo Co., Ltd. (72) Inventor Yuji Ozaki 1-7-12 Toranomon Minato-ku, Tokyo Oki Denki Kogyo Co., Ltd. (72 ) Inventor Mitsuru Fujita 5-7-1 Shiba, Minato-ku, Tokyo NEC Corporation (72) Inventor Akihiko Suwa 5-7-1 Shiba, Minato-ku, Tokyo Japan Ki Co., Ltd. (56) Reference JP-A-4-79644 (JP, A) JP-A-3-97394 (JP, A) JP-A-3-254252 (JP, A) JP-A-2-194747 (JP , A) JP 62-226264 (JP, A) Satoshi Hinomura and 2 others, Load balancing method for bursty high-speed communication systems, IPSJ research report, Japan, IPSJ, May 1993 14 Sun, Vol. 93, No. 36, pp. 19-26, 93-DPS-60-3 (58) Fields investigated (Int.Cl. ⁷ , DB name) H04L 12/56

Claims (2)

(57) [Claims] 1. A call setting process for a source or destination terminal for an individual call that occurs in each unit of logical multiplexing in a data exchange network that divides transmission information and logically multiplexes and transmits the data. A control device (11) that forms a communication path between the load distribution destination and the terminal, and a call that is load-balanced by the control device (11), via the communication path. In a load balancing type data exchange equipped with an exchange processing device (13 _{1 to} 13 _N ) that performs call processing based on a communication procedure facing the terminal, the control device (11) includes the exchange processing device (13 ₁ ~ 13 _N ) every
The storage means (15) for holding the total processing amount of calls distributed by (11), and the storage means among the exchange processing devices (13 _{1 to} 13 _N ).
The one corresponding to the minimum total processing amount held in (15) is set as the distribution destination, and the line speed of the communication path or the throughput generated by the call whose load is distributed to the communication path is set to the minimum total processing amount. The load balancing method, comprising: a distribution control means (17) for adding any one of the increments. 2. Transmission information is divided and logically multiplexed for transmission.
That occurs in each unit of logical multiplexing in the data exchange network
For each call, the call to the source or destination terminal
Performs the setting process and balances the load.
Control device that forms a communication path between the distribution destination of the
Set (11), For the call whose load is distributed by the control device (11),
Communication procedure facing the terminal via the communication path
An exchange processing device (13) that performs call processing based on₁~ 13 _N) And
In the provided load balancing type data exchange, The control device (11) includes Operation mode of individual terminals that can be the distribution destination of the load distribution
An operation mode storage means (21) for holding The exchange processing device (13₁~ 13_N) Of the
Possible exchange processing equipment individually corresponding to the above operation mode
Distribution destination storage means (23) held in advance, The operation mode of the terminal at the destination is stored in the operation mode storage means.
Distributed destination storage acquired from (21) and corresponding to its operation mode
The exchange processing device held by the means (23) is selected as the distribution destination.
And a distributed control means (25) for selecting a range.
Load balancing method.