JP2002165239A - Multiprocessor type call processing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a multiprocessor type call processing system which is superior in scale expandability and functional expandability, and capable of high-speed processing. SOLUTION: A divider 101 divides a message received from an external network among call processing units 1051, 1052, and so on which then efficiently process simple call processes by themselves and request host processors 1091, 1092, and so on to process about complicated processes or those not dealt with by themselves. A call state memories 1081, 1082, and so on of the call processing units 1051, 1052, and so on already store call state control tables for controlling call processing conditions, and call processing processors 1061, 1062 and so on reference the call state control tables to request the host processors 1091, 1092, and so on to process.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multiprocessor type call processing system in which call processing is distributed to a plurality of processors for processing.

[0002]

2. Description of the Related Art Hitherto, a multiprocessor type call processing system has been developed and used for distributing call processing to a plurality of processors. As an example of a multi-processor type call processing system, in order to provide a service quickly and efficiently with the same system to an area where the requirements of each of the channel capacity and the control system processing capacity are different, An example of a communication node system that adopts an architecture in which the control system control devices are separated and can be independently expanded is a “device distributed control method in a multiprocessor node system” (IEICE Technical Report SSE9).
9-106 / TM99-49, The Institute of Electronics, Information and Communication Engineers, 19
November 1999). In the system described in the above-mentioned document, since the control system device and the communication path system device are separately mounted as separate processing modules, it is possible to flexibly cope with the scale-up of the system performance. Also, if the failure range falls within the module,
There is an advantage that the whole system can be prevented from being stopped due to the failure, and the failure can be recovered only by replacing the processing module.

Also, Japanese Patent Application Laid-Open Nos. 62-95049, 6-77997, 10-233776.
JP, JP-A-10-243095, JP-A-200
0-197994, JP-A-2000-224649
The publication discloses a method of distributing call processing in a multiprocessor call processing system. The call processing distribution method described in these publications is such that each call processing unit periodically notifies the distribution unit of its own load status so that the distribution unit can recognize a low-load call processing unit, Thus, the distribution unit can recognize that the call processing unit having no notification is in an abnormal state, for example, a congestion state. The distribution unit distributes processing to a call processing unit that is not in an abnormal state based on these states, and attempts to equalize the load of call processing.

[0004]

However, in improving the system performance, there is often a trade-off between the improvement of the processing performance and the diversification of the processing. However, there is a problem that high speed cannot be sufficiently achieved simply by allocating processes according to load conditions and performing multiprocessor processing as in the invention described in (1).

An object of the present invention is to provide a multiprocessor type call processing system which is excellent in scale expandability and function expandability and can perform high-speed processing.

[0006]

According to the present invention, in a multiprocessor type call processing system in which call processing is distributed to a plurality of processors, a distribution means for distributing call processing, A plurality of call processing means for performing processing, and at least one host processor means, wherein each of the call processing means, when the processing of a call that cannot be processed by itself is distributed by the distribution means, the host processor means And the host processor performs the requested call processing. A multiprocessor call processing system is provided. Each of the call processing means requests the host processor to process the call when the processing of the type of call that cannot be processed is allocated by the distribution means, and the host processor performs the requested call processing. .

Here, each of the call processing means has first storage means for storing a call status management table in which the call requested by the host processor means and the host processor means requesting the call processing are associated with each other, Using the call state management table,
While requesting call processing to the host processor means, a subsequent message relating to the call may be transferred to the host processor means. Further, each of the call processing means, when requesting the call processing to the host processor means, to notify the call requesting processing to the distribution means and processing request contents representing the host processor means, the distribution means, the distribution means, The processing request content is stored in the second storage means, and the message following the call is transmitted to the corresponding host processor means by referring to the processing request content stored in the second storage means. May be. Further, each of the call processing means, when requesting the call processing to the host processor means, to notify the call requesting processing to the distribution means and processing request contents representing the host processor means, the distribution means, the distribution means, The processing request content is stored in the second storage means, and the message subsequent to the call is transmitted to the corresponding call processing means and host processor means with reference to the processing request content stored in the second storage means. May be configured.

[0008] When the call processing means makes a processing request to the host processor means, the processing request message may be added with information necessary for the processing request and transmitted. Further, when the host processor receives the processing request and performs the call processing, when it is determined that a message that cannot be processed by the call processing unit that has requested the call processing follows, the processing of the subsequent message is completely performed by the host processor. After the termination, a processing termination notification may be transmitted to the call processing means. Further, when the distribution unit or the call processing unit receives a request to cancel the call setting process, after the distribution unit or the call processing unit responds to the cancellation request, the call processing unit performs a corresponding call state management. You may be comprised so that the deletion process of the entry in a table may be performed. Further, the call processing means may be configured to perform call processing in a processing function unit.

According to the present invention, in a multiprocessor call processing system in which call processing is distributed to a plurality of processors, a distribution means for distributing the call processing, a third storage means, A plurality of call processing means for performing the processing of the call, and at least one host processor means, each of the call processing means, when the processing of a call that can not be processed by itself is sorted by the sorting means, Requesting the host processor to process the call, and notifying the call requesting processing to the host processor and the name of the host processor to the distribution means, and the host processor receiving the request from the call processing means. Processing the received call, and the distributing means comprises: the call received from the call processing means and the host processor. Stores the stage names to the third memory means, by referring to the stored contents of said third memory means,
A multiprocessor call processing system is provided, wherein a subsequent message addressed to the call is sent directly to the host processor means. Each of the call processing means requests the host processor means to process the call when the processing of a call of a type which cannot be processed by the distribution means is performed, and the call requesting processing to the host processor means and the host A processor means name is notified to the distribution means. The host processor processes a call requested by the call processor. The allocating means stores the call and the host processor name received from the call processing means in the third storage means, and refers to the storage contents of the third storage means to determine a subsequent call destination for the call. Is sent directly to the host processor means.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Before describing a multiprocessor call processing system according to an embodiment of the present invention,
First, the outline will be described. The multiprocessor type call processing system according to the embodiment of the present invention performs processing of a predetermined type of call provided in the front end (for example, processing of a simple call).
And a plurality of call processing units that perform the processing efficiently. At least one call processing other than the predetermined type (for example, complicated processing or processing not supported by the call processing unit) is performed. By sending a processing request to the provided host processor unit, a system with high processing efficiency is constructed. By adopting a multiprocessor configuration, scale scalability is also realized. Specifically, in order to perform high-speed processing in the call processing unit, a call state management table is arranged in the call processing unit to shorten the memory access time. Also, in order to maintain the consistency of the protocol operation, while requesting the call processing to the host processor, the message relating to the call is reliably transferred to the host processor.

Further, by configuring the message to be processed by the host processor to quickly reach the host processor without passing through the call processing unit, the processing start time in the host processor can be reduced. At the same time, the processing delay in the host processor decreases,
As a result, the processing performance of the system is improved. When the call processing unit requests the host processor for processing (switch request), the processing request message (switch request message) includes information (call state, message received by the call processing unit) necessary for the processing request. , Call processing unit number). As a result, when the switching request is approved by the host processor unit, it is not necessary to send the information necessary for the processing request again, so that the number of message exchanges between the call processing unit unit and the host processor unit can be reduced. ,
Therefore, the processing speed can be improved.

After the host processor processes the message, if the host processor determines that a message which cannot be processed by the call processing unit follows, the processing of the message is completed by the host processor. , A notification of the end of the process (switch request) is transmitted to the call processing unit. As a result, the number of times of switching between the host processor unit and the call processing unit unit can be reduced, so that the processing time can be reduced. Further, when the distributing unit or the call processing unit receives the cancel request message indicating the request to cancel the call setting process, it immediately responds to the cancel request, and the related process (clearing the corresponding call state table entry) is performed as described above. It is done after the response. Thereby, the processing delay for the cancel request message can be reduced.

Further, the call processing unit is constituted by a processing function unit (for example, a transfer processing function, a name resolution function, etc.). Thereby, when the processing performance for a certain function is insufficient, a call processing unit for realizing the corresponding processing function may be added, so that expansion according to the performance of each function can be realized. Further, when the implementation method of a certain function is changed, it can be dealt with only by exchanging the call processing unit, so that it is possible to flexibly cope with the change of the function implementation. In addition, by providing a common call state storage unit, when processing is requested to the host processor unit, messages are transmitted only between the distribution unit and the host processor unit. , The number of message transmissions and receptions can be reduced, and system resources can be used effectively.

Hereinafter, embodiments of the present invention will be described in detail. FIG. 1 is a block diagram of a multiprocessor type call processing system used in a multiprocessor type call processing system according to an embodiment of the present invention, and a block diagram of a system applied to a call processing method using a multiprocessor. FIG. In the present embodiment, SIP (Sessio
2 shows an example of a proxy server that performs n Initiation Protocol processing. Note that SIP is an application layer protocol and is a signaling protocol used to connect and disconnect a call between a plurality of SIP terminal devices.

In FIG. 1, a multiprocessor type call processing system includes a distribution unit 10 which constitutes distribution means.
1, a plurality of call processing units 105 ₁ , 105 ₂ ,... Arranged in the front end and constituting call processing means, respectively.
The host processor means comprises one or more host processors 109 ₁ , 109 ₂ ,... And a subscriber database 110. The call processing units 105 ₁ , 105 ₂ ... Perform processing of a predetermined type of call (for example, processing of a simple call), that is, perform only predetermined predetermined processing, and perform processing of a call other than the predetermined type ( For example,
Has a function of requesting at least one of the host processors 109 ₁ , 109 ₂ ,... For a complicated process or a process not supported by the call processing unit. . Also, each call processing unit 1
05 _1, 105 although ₂ ... to be processed can be set to each different as described below (for example, processing of the audio signals, ₂ call processing unit 105 processes the video signal), where the same thing Is assumed.

The distribution unit 101 receives a message received from an external network, and
5 _1, 105 2 _has a function of determining whether to be processed by ..., message transceiver 102 which constitutes a message receiving unit, and the message analyzing unit 103 constituting the message analyzing unit, the call processing unit The number correspondence table is stored and comprises a call processing unit state management unit 104 constituting a second storage means. The message transmitting and receiving unit 102 transmits and receives a message to and from the outside. The message analysis unit 103 analyzes the content of the message received by the message transmission / reception unit 102, and identifies a call targeted by the message (Call-ID:
Call ID) from the message (if necessary,
The source address and the destination address as well as the call ID are extracted). The call processing unit status management unit 104
(If necessary, not only the call ID but also the source address and the destination address) and the call processing units 105 ₁ , 105 ₂ ,
.. Or the correspondence between the call IDs (if necessary, not only the call IDs but also the source address and the destination address) and the host processors 109 ₁ , 109 ₂ ,. . The message analysis unit 103 can determine the transfer destination of the message by referring to the contents stored in the call processing unit state management unit 104.

The message is composed of a header part and a body part. The header portion mainly includes protocol processing information indicating which protocol processing corresponds to this message (information such as whether to perform call connection or disconnection), and the message includes It contains address information (source address, destination address, call ID for identifying the call, address of the computer from which the message has been propagated, etc.) necessary to propagate the message correctly. Further, the body portion stores all information that is not directly related to SIP (for example, information for negotiating a medium used between SIP terminal devices). The types of messages are roughly classified into request messages and response messages.
The request message is a method (METHOD: processing performed by the client to the server, for example, INVI
There are as many as TE (requests call connection), ACK (confirms call connection and starts conference), BYE (disconnects call), INFO (transmits information related to call), etc. There are a plurality of response messages corresponding to the types of response codes (for example, information notification for the 100s, success for the 200s, redirection for the 300s, and failure for the 400s due to incorrect requests).

The call processing units 105 ₁ , 105 ₂ ,.
.. Takes the received message as an input, transits the corresponding call state, and performs the accompanying processing. Each call processing unit ₁₀₅ 1, ₁₀₅ 2, since ... are same configuration and explains their construction the processing unit 105 ₁ as an example, the call processing unit 105 ₁ constitutes a message transmitting and receiving means Message transmitting / receiving unit 10
7 _1, the call processor unit 106 ₁ and constituting the call processor unit, and a call state storage section 108 ₁ constituting the first storage means. Message transmitting / receiving unit 1
07 ₁ transmits / receives messages to / from the distribution unit 101 and the host processor units 109 ₁ , 109 ₂ ,.... Call state storage unit 108 ₁ stores a call state management table representing the call state. The subscriber database 110 stores subscriber data. The subscriber database 110 searches the destination of the message in the subscriber database 110 to check whether a transfer destination exists.

[0019] Figure 2 shows a call state management table stored in the call state storage unit 108 _1. Each entry of the call state management table includes a call ID, a call state, the number of the host processor requesting the process, and the number of messages transferred to the host processor unit. Note that “null” indicates that no processing request has been made to the host processor unit.
And messages received call processor unit 106 _1,
Processing is determined from the corresponding call state and call processing is performed. FIG.
A call processing unit ₁₀₅ 1, ₁₀₅ 2, message transceiver ₁₀₇ 1 ..., ₁₀₇ 2, shows the processing procedure from the message arrived at., Each call processing unit ₁₀₅ 1, 105 _2, the call processor unit ₁₀₆ 1 ..., ₁₀₆ 2, a process ... are shown in FIG. 3 performed. The processing procedures of the call processing units 105 ₁ , 105 ₂ ,... Are the same, and each performs the processing of FIG.

FIG. 3 the call processing unit 105 ₁ as an example
To illustrate the process, if the message transceiver 107 ₁ call processing unit 105 ₁ receives a message from the distribution unit 101, the call processor unit 106 ₁ analyzes the contents of the message (process step A1 in FIG. 3) ,
Searches for the corresponding call state from the call state storage section 108 ₁ (step A2). Here, (if necessary, the source address and destination address not call ID only) extracted call ID in the distribution unit 101, added to the message distribution unit 101 transfers to the call processing unit 105 _1, a call by processor unit 106 ₁ to refer to it, it can be configured to retrieve the appropriate call state from the call state storage unit 108 _1. Next, it is determined whether the "processing request to the host processor" corresponding call state (step A3), if the call is not in processing request to the host processor, processing in the call processors 106 ₁ is started (A5 to A122).

[0021] In this case, the call processor 106 _1, first determine the message type (step A5), the message is whether it is possible to process the call processing unit 105 ₁ itself, that determines whether a predetermined type of call (Step A6). If the call can be processed, the call state searched in step A2 is changed to perform call processing (step A7). After that, a message transmission process is performed accordingly. in this case,
First, update writes a call state after the transition to the call state storage section 108 ₁ (step A8), if there is a message to be transmitted, and ends by sending to the distribution unit 131 it (step A9). The distribution unit 131 transmits the message from the message transmission / reception unit 132 to the outside. In step A6, the received message is
If it is judged not to be processed in 05 ₁ itself, it requests processing to the host processor. In this case, first, the host processor which requests the processing is determined (step A10).
At this time, it is conceivable that the host processor is determined in consideration of the load state of the host processors 109 ₁ , 109 ₂ ,. Then, it asks the selected host processor whether a call processing request can be made, that is, sends a processing request notification (switch notification) to ask whether a call processing request (switch processing) is possible (step A11). ).

If the switching is rejected by the host processor, the process returns to step A10 (step A1).
2) Select another host processor unit and continue the process. When the switching is approved by the host processor unit (step A12), as shown by the broken line in FIG.
For the corresponding host processor, call state is information required for processing the request, and forwards the call processing unit number of the call the message processing unit 105 ₁ has received and the call processing unit 105 ₁ performs request processing ( Step A12
1). FIG. 4 shows the contents of the message used when making the request. As shown in FIG. 4, the request message includes a call ID,
Call state, constituted by the message and the call processing unit 105 ₁ of the call processing unit _{numbers 1} call processing unit 105 has received. INFO is a message indicating that information related to a call is transmitted. Incidentally, if the call addressed message corresponding to the call processing unit 105 ₁ before request processing is completed has reached the distributing unit 101, the message is also transferred to the host processor unit in step A121. Thereafter, it updates the call state by rewriting the applicable entry in call state management table of the call state storage section 108 in the ₁ (2) (Step A 122). At this time, the call state of the entry is “processing request to host processor”,
The host processor number of the corresponding host processor is written in the host processor number.

In step A3, if the corresponding call is "being processed by the host processor", the received message is transferred to the same host processor, and the processing is requested. Therefore, the number of transfer messages in the appropriate entry in the call state storage section 108 ₁ incremented by one (step A13), and determines a destination from the host processor number included in the entry (step A14), forward the message (Step A15). When the host processor finishes the requested processing, the host processor notifies the requested call processing unit of an end notification (switch request) indicating that the processing is ended. The processing result message shown in FIG. 5 is added to the switching request notification. That is, a call ID, a call state, and the number of transfer messages are added.

FIG. 6 shows a call processing unit 105.₁, 10
5₂,... Are host processors 109 ₁, 109₂,
... operation procedure after notification of processing switch request
Is shown. Each call processing unit 105₁, 105₂...
Call processor 106₁, 106₁,... Are shown in FIG.
The processing shown in is performed. Call processing unit 105₁For example
That is, the message transmitting / receiving unit 107 ₁Through
Call processor that has received the end notification (switch notification)
106₁Then, the call state storage unit 108₁Applicable call diversion
Number of messages sent and messages forwarded in processing results
The number is compared (step B1). If both are equal,
Notify the host processor of the switching approval (step B)
3), call state storage unit 108₁After processing in the call state management table of
The call state is written and updated (step B4). Step
In step B2, the call processing unit 105₁Transferred
The number of messages is the number of messages transferred from the host processor.
If the number of messages is less than
There is a potential. Therefore, if the number of transferred messages is different,
In this case, the host processor is notified of the rejection of termination (switching).
(Step B5). The call processing unit 10
5₁Switch even when processing load is too high to process
You can also tell refusal. Switch reject received
The host processor waits for a subsequent message and performs processing.
When the process is over, a switch request is made again.
Applicable call processing unit 105₁Notify.

FIG. 7 shows an operation example of the entire system. In this example, message corresponding to a call, but is transmitted to the call processing unit 105 ₂ corresponding to the call ID, can not process a call processing unit 105 _2, the message,
Shows an example of a request for processing to the host processor 109 _1. To illustrate the example, the SIP (Session Initiati
on Protocol). For the SIP specification, I
R issued by ETF (Internet Engineering Task Force)
FC2543.

In FIG. 7, a message is sent from an external network.
When message # 1 reaches the distribution unit 101 (1), the message
Message # 1 was received by the message transmitting / receiving unit 102
After that, the message analysis unit 103 calls from the message # 1.
Extracts the call ID (Call-ID) and manages the call processing unit status
Referring to 104, the call processing unit of the transfer destination is determined.
You. In this example, the call processing unit 105₂Forward to
(2). Call processing unit 105 ₂Now, call processing processor
Sa106₂Determines the message type (step in FIG. 3).
A5). As a result, the message is
105₂Is a type of call processing that is not supported by MET
If it is HOD (for example, INFO), the host processor
Decide to request processing from the server. Host processor
109₁Is determined as the transfer destination (step A10), and the host
Processor 109₁Processing request (switch) notification for
(3).

[0027] Upon receiving the processing request (switching) approval from the host processor 109 ₁ (4), the call processing unit 1
05 ₂ call state held, and forwards the message to request the processing. In this example, before the switching process to the host computer 109 ₁ is completed, a new message ♯2 addressed the call have been transferred (5)
(6). In such a case, ₂ call processing unit 105 holds the message ♯2 of the received (6), at the time of submission, call state, and the first message received in #
1 is transferred to the host processor 109 ₁ (7), and the held message # 2 is also transferred to the same host processor 10 1
It is transferred to the 9 ₁ (8). During processing request to the host processor 109 _1, if applicable call message addressed # 3 has been transferred to the ₂ call processing unit 105 (9) (1
0), ₂ call processing unit 105 call state storage section 108 ₂
Referring to the call state management table stored in, and transfers the message ♯3 to the appropriate host processors 109 ₁ (1
1).

[0028] If, before the message # 3 has been transferred to the host processor 109 ₁ arrives at the host processor 109 _1, end notification representing the end of processing from the host processor 109 ₁ (switching request) is notified (12) ,
Inconsistent message delivery occurs. However, the mismatch, it is possible to discover ₂ call processing unit 105 by comparing the number of transfer message in step B1, performs denial of termination (switching) (13). Host processor 109 ₁ has received the denial of completion, after the processing for the message # 3, again, end notification (switching request)
Notify. Now, because the message transfer count are coincident at step B1, end notification (switching) is approved, call state after processing to the call processing unit 105 ₂ are sent. The call processing unit 105 ₂ call state storage section 10
8. Update the corresponding entry in ₂ .

According to the above-described first embodiment of the present invention, since a plurality of call processing units 105 ₁ , 105 ₂ ... By adding the call processing unit, addition for each processing unit can be easily performed, and therefore, scalability can be realized. In this case, the predetermined type of processing may be processing of a plurality of functions. Further, the control system device is configured to include a plurality of call processing module units 105 ₁ , 105 ₂ ,
, And at least one host processor unit 109 ₁ , 109 ₂ ,... Corresponding to complex processing,
It is possible to construct a multiprocessor type call processing system which is excellent in scale expandability and function expandability and further increases control system processing capability.

Further, the call processing units 105 ₁ , 105 ₂ ,... Arranged at the front end access the call state management tables in the call state storage units 108 ₁ , 108 ₂ ,. Because call processing is realized at high speed,
It is possible to realize efficient call processing. Furthermore, since the call processing units 105 ₁ , 105 ₂ ,... Only need to provide a simple call processing function, the circuit scale can be reduced, and the processing performance can be improved at low cost by adding a plurality of such circuits. Therefore, the system can be realized at low cost. Also, since protocol extension can be handled by software change on the host processor,
Flexible function expansion can be realized.

Next, a second embodiment of the present invention will be described. The hardware configuration of the second embodiment is the same as that of the first embodiment, and is configured as shown in FIG. The second embodiment differs from the first embodiment in that the call processing units 105 ₁ , 105 ₂ ,.
Are requesting call processing to the host processors 109 ₁ , 109 ₂ ,..., As indicated by the broken line in FIG. 14, the distribution unit 101 transfers the message to the host processors 109 ₁ , 109 ₂ that are performing the request processing. ,...

FIG. 15 shows the call processing units 105 ₁ , 10
5 ₂ ,... Message transmitting / receiving sections 107 ₁ , 10
7 2 _is a flowchart illustrating a processing procedure from messages arrived to .... Each call processing unit 10
5 _{1, 105} 2, ... of the call processor unit 10
6 _{1, 106} 2, performs processing ... are shown in FIG. 15.
Each of the call processing processors 106 ₁ , 106 ₂ ,... Performs the processing shown in FIG.
The process ₁ will be described as an example. In FIG. 15, each process from step A1 to step A12 is the same as the processing operation of the first embodiment, and thus the description is omitted. Result of the determination in step A6, if unable to process the message received from the distribution unit 101 in the call processing unit 105 _1, the host processor ₁₀₉ 1, ₁₀₉ 2, ...
・ Request the processing to one of. The difference from the first embodiment is that after the processing request, the call processing processor 10
6 _1, in that updating rewrites the call processing unit state management unit 104 is in the sorting unit 101 (step C1). That is, _{after 1} call processing unit 105 is requested to process the host processor, the call processor unit 106
₁ indicates the contents stored in the call processing unit state management unit 104,
The request to the host processor to process the call and the number of the requested host processor are written.

After the processing of step C1, the distribution unit 101
When a message addressed to the call arrives, the message analysis unit 103 sends the message to the call processing unit state management unit 10.
Referring to FIG. 4, as shown by the broken line in FIG. 14, in order to request the corresponding host processor to process the corresponding call,
Forward the message directly. If the corresponding call is "being processed by the host processor" in step A3, an error occurs (step E1). This is step C
This is because the operation of the distributing unit 101 is changed by _{No. 1} , so that a message cannot be transferred to the call processing unit 1051. Note that each of the call processing units 105 ₁ , 105 ₂ ,
Are not used in the present embodiment.

When the processing in the host processor is completed normally, the processing operation of the distribution unit 101 is changed.
The message must not be forwarded to the host processor. FIG. 16 shows the call processing unit 105 ₁
Shows the operation procedure after receiving the processing result from the host processor. Steps B2 to B5 are the same as the processing procedure of the first embodiment, and a description thereof will be omitted. When the processing in the host processor ends normally and an end notification is received, in step D1, the contents stored in the call processing unit state management unit 104 are changed so that a message is not transmitted to the host processor that has completed the processing. .

FIG. 17 shows an operation example of the entire system according to the second embodiment. In this example, the message corresponding to a certain call is the call processing unit 10 corresponding to the call ID.
5 ₂ but transmitted to can not process the call processing unit 105 _2, the message, the host processor 109
₁ shows an example in which processing is requested to ₁ . The second shown in FIG.
17 is different from the processing operation of the first embodiment shown in FIG.
Of after the call processing unit 105 ₂ is subjected to a process request (switching) approval from the host processor 109 ₁ (4), (5) to the sorting unit 101, the fact that processing request processing for the call to the host processor 109 ₁ (Switching) notification is transmitted to change the call processing unit state management unit 104 of the distribution unit 101 (step C1 in FIG. 15). By this operation, after the message # 3 arrives at the distribution unit 101 in (11), the message # 3
3, as indicated by a broken line in FIG. 14, is transferred directly from the distribution unit 101 to the host processor 109 _1. However, since after the process switching from the host processor unit 109 ₁ to the call processing unit 105 _2, and then transfer the message forwarded to the host processor 109 ₁ (16) to the call processing unit 105 _2.

The process which has been requested to the host processor 109 ₁ is completed, _{after 2} call processing unit 105 approves the ends (switch) (13), the process ends (switching)
Is notified to the distribution unit 101 (14), and the distribution unit 101 updates the call processing unit status stored in the call processing unit status management unit 104 (D1). By this operation, thereafter, the message for the corresponding call is transferred only to the call processing unit units 105 ₁ , 105 ₂ ,....

FIG. 18 shows another example of the processing operation of the entire system according to the second embodiment. In this example, (4)
The call processing unit 105 ₂ is connected to the host processor 109 ₁
After receiving the switching approval from, before forwarding the call state and message # 1 (10), the message # 3 is transferred from the distributing unit 101 to the host processor 109 _1. The case message # 3 host processor 109 ₁ that has been transferred and allowed to hold up (10) and (11) respectively call state, a message # 1, and the message # 2 is transferred, the message Processing is performed in numerical order. This message number uses sequence number header information (CSeq) defined in SIP.
Further, the distribution unit 101 can also add the message to each call. Before the host processor 109 ₁ is to forward the message # 4 to ₂ call processing unit 105 in the same way (20), (19), the message # 5 from the distribution unit 101 to the call processing unit 105 ₂ is transferred . Also in this case, the call processing unit 105 ₂
Performs the processing of message # 5 after message # 4 is transferred in (20). As described above, according to the second embodiment, the host processors 109 ₁ , 109 _1
The message to be processed in the call processing unit 1
05 _{1, 105} 2, the host processor ₁₀₉ 1 without passing through the ..., ₁₀₉ 2, to reach directly., The host processor ₁₀₉ 1, ₁₀₉ 2, it is possible to reduce the processing start time at ... Therefore, as compared with the first embodiment, the host processor units 109 ₁ , 109 ₂ ,.
, Processing delay is reduced, and as a result, the processing performance of the system can be improved.

Next, a third embodiment of the present invention will be described. The hardware configuration of the third embodiment is the same as that of the first embodiment, and is configured as shown in FIG. The third embodiment is different from the first embodiment in that the call processing units 105 ₁ , 105 ₂ ,.
, Is requesting the host processor 109 ₁ , 109 ₂ ,... To perform call processing, as shown by the broken line in FIG. 12, the distribution unit 101 transmits the message to the call processing unit 105 ₁ ,.
105 _2, not only ..., the host processor 109
₁ , 109 ₂ ,...

FIG. 8 shows the call processing units 105 ₁ and 105 ₁ .
_2, ... of the message transmitting and receiving unit ₁₀₇ 1, ₁₀₇ 2,
Is a flowchart showing a processing procedure after a message arrives at. Each call processing unit 105 ₁ , 10
The call processor units 106 ₁ , 10 ₂ ,...
6 ₂ performs processing ... are shown in FIG. Each of the call processing processors 106 ₁ , 106 ₂ ,... Performs the processing shown in FIG. 8, so the processing of the call processing unit 105 ₁ will be described as an example. In FIG. 8, each processing of steps A1 to A12 and step A13 is performed by the first processing.
The description is omitted because it is the same as the processing operation of the embodiment. Result of the determination in step A6, if unable to process the message received from the distribution unit 101 in the call processing unit 105 _1, the host processor ₁₀₉ 1, 10
Request processing to any of 9 ₂ ,. Wherein a difference from the first embodiment, after the processing request, the call processor unit 106 _1, a point to update rewrites the call processing unit state management unit 104 is in the sorting unit 101 (step C1). That is, _{after 1} call processing unit 105 is requested to process the host processor, the call processor unit 106 _1, the storage contents of the call processing unit state management unit 104, that and request that requested processing for the call to the host processor Write the number of the specified host processor.

After the processing in step C1, the distribution unit 101
When a message addressed to the call arrives, the message analysis unit 103 sends the message to the call processing unit state management unit 10.
4 See, as indicated by a broken line in FIG. 12, not only the call processing unit 105 ₁ which distributes the processing of the corresponding call, also forwards the message to the appropriate host processors. Also, if "being processed in the host processor" the call concerned is that in step A3, the call processor unit 106 _1, after the number of transfer messages in the appropriate entry in the call state storage unit 108 ₁ is incremented by one, The message is discarded (step C2). This in step C1, since the by changing the operation of the distributing section 101, because the call processing unit 150 ₁ without transferring the message has arrived to the host processor the same message is applicable.

When the processing in the host processor is completed normally, the processing operation of the distribution unit 101 is changed.
The message must not be forwarded to the host processor. The call processing unit 105 ₁ in FIG. 9 shows an operation procedure after receiving the result of processing from the host processor. Steps B1 to B5 are the same as the processing procedure of the first embodiment, and a description thereof will be omitted. When the processing in the host processor ends normally and an end notification is received, in step D1, the contents stored in the call processing unit state management unit 104 are changed so that a message is not transmitted to the host processor that has completed the processing. .

FIG. 10 shows an example of the operation of the entire system according to the third embodiment. In this example, the message corresponding to a certain call is the call processing unit 10 corresponding to the call ID.
5 ₂ but transmitted to can not process the call processing unit 105 _2, the message, the host processor 109
₁ shows an example in which processing is requested to ₁ . The third shown in FIG.
The processing operation of the entire system according to the embodiment is different from the processing operation according to the first embodiment shown in FIG.
Of after the call processing unit 105 ₂ is subjected to a process request (switching) approval from the host processor 109 ₁ (4), (5) to the sorting unit 101, the fact that processing request processing for the call to the host processor 109 ₁ Is transmitted to change the call processing unit state management unit 104 of the distribution unit 101 (step C1 in FIG. 8). By this operation, the message # in (10)
3 arrives at the distribution unit 101, the message # 3
, As indicated by a broken line in FIG. 12, ₂ call processing unit 105 from the sorting unit 101 and the host processor unit 1
09 ₁ is transmitted (bicast) to both.

The process which has been requested to the host processor 109 ₁ is completed, _{after 2} call processing unit 105 approves the ends (switch) (16), the process ends (switching)
Is notified to the distribution unit 101 (17), and the distribution unit 101 updates the call processing unit status stored in the call processing unit status management unit 104 (D1). By this operation, thereafter, the message for the corresponding call is transferred only to the call processing unit units 105 ₁ , 105 ₂ ,.... As described above, according to the third embodiment, the host processors 109 ₁ , 10 ₁
9 _2, message call processing to be processed by ... unit ₁₀₅ 1, ₁₀₅ 2, the host processor ₁₀₉ 1 without passing through the ..., ₁₀₉ 2, to reach directly.,
Since the processing start time in the host processors 109 ₁ , 109 ₂ ,... Can be shortened, compared with the first embodiment, the host processor units 109 ₁ , 109 ₂ ,.
, Processing delay is reduced, and as a result, the processing performance of the system can be improved.

FIG. 13 is a block diagram of a multiprocessor type call processing system according to a fourth embodiment of the present invention, and is a configuration diagram of a system applied to a call processing method using a multiprocessor. An example of a proxy server is shown similarly to FIG. 1, and the same parts as those in FIG. 1 are denoted by the same reference numerals. In the first and second embodiments,
The types of call processing performed by the respective call processing units 105 ₁ , 105 _2, ... Are set to be different from each other, but the fourth embodiment assumes the same type.

The difference from the first to third embodiments is that a common call state storage unit 134 constituting a third storage unit is used instead of the call processing unit state management unit.
And the distribution unit 131 includes a message transmitting / receiving unit 132
And a message analysis unit 133, and the other configuration is the same. The common call processing storage unit 134 stores all call states managed by this system. In the fourth embodiment, the distribution unit 131 detects a corresponding call state from the received message, and
05 ₁ , 105 ₂ ,..., The corresponding call state and the received message are distributed and delivered. After processing the message, the call processing units 105 ₁ , 105 ₂ ,... Return the processed call state to the distribution unit 131 and store the same in the common call state storage unit 134.

The call processing unit 105₁, 10
5₂,... Are the host processor 109₁, 10
9₂, ..., if requested, call
And the requested host processor 109₁, 109₂,number
To the distribution unit 131, and the common call state storage unit 134
To store. Call processing unit 105₁, 105₂,
.. processing of the call corresponding to the host processor 109₁,
109₂, ..., the corresponding call is
That the message analysis unit 133 is in the common call state.
It can be understood by searching the state storage unit 134. Therefore,
The message analysis unit 133 searches the common call state storage unit 134
A subsequent message for the call
Is transmitted only to the host processor (the first
As in the third to third embodiments, the call processing unit
Socket 105₁, 105₂Deliver a message to ...
Need not be done). According to the fourth embodiment, the host
Processor unit 109₁, 109₂Processing request to ...
Sometimes, the distribution unit 131 and the host processor unit 10
9₁, 109 ₂Message transmission only between…
Therefore, the distribution unit 131 and the call processing unit unit 105
₁, 105₂, ... Reduce the number of messages sent and received
System resources can be used effectively.
You.

The above-described embodiments are not limited to various modifications.
Is possible. For example, in each of the above embodiments,
Processing unit 105₁, 105₂... cancel
It does not correspond to predetermined processing such as
The host processor processes the cancel request message.
Sa109₁, 109₂, ..., the host
Processor 109₁, 109₂Cancel until ...
After the request message is transmitted, the host processor 10
9₁, 109₂, ..., cancel processing
The result is sent to the call processing unit 105.₁, 105₂,···What
Processing is performed to convey. In contrast, sorting
Units 101 and 131 or call processing unit unit 105₁,
105₂, ... request to cancel the call setup process
If it is received, it is assigned to the sorting units 101 and 131 or the call
Processing unit 105₁, 105 ₂...
Responds to the cancel request and performs the related processing (applicable
Clearing the call state management table entry)
Socket 105₁, 105₂... will be done after the response
You may do it. As a result, the cancellation request message
Processing delay for certain messages such as messages
You.

The host processors 109 ₁ and 109 _1
2, when a ... is call processing at the request of the call processing, the host processor ₁₀₉ 1, ₁₀₉ 2, the call processing unit ₁₀₅ 1, ... has requested to the call processing, ₁₀₅ 2, ...
If it is determined that there are messages that cannot be processed by the host processor, the processing of those messages may be completed by the host processor, and then a processing end notification may be transmitted to the call processing unit that requested the processing. For example, the request cannot be processed by the call processing unit that has requested the call processing until a request message for processing termination permission is transmitted from the host processor to the requested call processing unit and a response message is received from the call processing unit. It may be determined that the message follows. Thus, the number of times of switching between the host processor and the call processing unit can be reduced, and the processing time can be reduced.

The call processing units 105 ₁ , 10 ₁
5 _2, the ... processing functional unit (e.g., transfer processing function,
(Name resolution function, etc.). Thereby, the extension according to the performance of each function can be realized. Therefore, when the processing performance for a certain function is insufficient, a call processing unit for realizing the corresponding processing function may be added.
Further, since it is possible to flexibly cope with a change in the implementation of a function, it is possible to cope with a change in the implementation method of a certain function only by replacing the unit.

[0050]

According to the present invention, scalability, and
It is possible to provide a multiprocessor type call processing system which is excellent in function expandability and capable of high-speed processing.

[Brief description of the drawings]

FIG. 1 is a block diagram of a multiprocessor call processing system used in first to third embodiments of the present invention.

FIG. 2 is a diagram showing a call state management table used in the embodiment of the present invention.

FIG. 3 is a flowchart illustrating a process according to the first embodiment of the present invention.

FIG. 4 is a diagram showing a request message used in the embodiment of the present invention.

FIG. 5 is a diagram showing a processing result message used in the embodiment of the present invention.

FIG. 6 is a flowchart illustrating a process according to the first embodiment of the present invention.

FIG. 7 is an explanatory diagram illustrating a process according to the first embodiment of the present invention.

FIG. 8 is a flowchart illustrating a process according to the third embodiment of the present invention.

FIG. 9 is a flowchart illustrating a process according to the third embodiment of the present invention.

FIG. 10 is an explanatory diagram illustrating a process according to a third embodiment of the present invention.

FIG. 11 is an explanatory diagram of a multiprocessor call processing system used in the first embodiment of the present invention.

FIG. 12 is an explanatory diagram of a multiprocessor type call processing system used in a third embodiment of the present invention.

FIG. 13 is a block diagram of a multiprocessor call processing system used in a fourth embodiment of the present invention.

FIG. 14 is an explanatory diagram illustrating a process according to the second embodiment of the present invention.

FIG. 15 is a flowchart illustrating a process according to the second embodiment of the present invention.

FIG. 16 is a flowchart illustrating a process according to the second embodiment of the present invention.

FIG. 17 is an explanatory diagram illustrating a process according to the second embodiment of the present invention.

FIG. 18 is an explanatory diagram showing a process according to the second embodiment of the present invention.

[Explanation of symbols]

101 ... distributing unit 102,132,107 ₁ constituting the distributing means, ₁₀₇ 2 ... message transceiver 103,133 constituting the message transmitting and receiving means, message analyzer 104 which constitutes a ... message analyzing means, ..Call processing unit status management units 105 ₁ , 105 ₂ forming second storage means Call processing unit units 106 ₁ , 106 ₂ forming call processing means Calls forming call processing processor means Processing processor sections 108 ₁ , 108 ₂ ... Call state storage sections 109 ₁ , 109 ₂ forming first storage means. Host processor sections 110 forming host processor means 110. Subscriber database 134. A common call state storage unit constituting the third storage means

Claims (9)

[Claims] 1. A multiprocessor call processing system in which call processing is distributed to a plurality of processors, wherein a distribution means for distributing the call processing, a plurality of call processing means for processing a predetermined type of call, At least one host processor means, wherein each of the call processing means requests processing of the call to the host processor means, when the processing of a call of a type which cannot be processed by itself is distributed by the distribution means, A multiprocessor call processing system, wherein the host processor performs the requested call processing. 2. The first call processing means according to claim 1, wherein said first call processing means stores a call state management table in which a call requested by said host processor means and a host processor means requesting call processing are associated with each other.
Wherein while requesting call processing to the host processor means using the call state management table, a subsequent message relating to the call is transferred to the host processor means. Item 2. The multiprocessor call processing method according to Item 1. 3. When each of the call processing means requests a call processing to a host processor means, the call processing means notifies the distributing means of a call requesting the processing and a processing request content representing the host processor means. Stores the processing request content in a second storage means, and refers to the processing request content stored in the second storage means, and transmits a message subsequent to the call to the corresponding host processor means. 3. The multiprocessor call processing method according to claim 1, wherein 4. When each of the call processing means requests a call processing to the host processor means, the call processing means notifies the distribution means of the call requesting the processing and the processing request content representing the host processor means. Stores the processing request content in the second storage means, refers to the processing request content stored in the second storage means, and stores the subsequent message of the call in the corresponding call processing means and host. 4. The multiprocessor call processing system according to claim 1, wherein the call is transmitted to a processor. 5. The processing method according to claim 1, wherein when the call processing unit sends a processing request to the host processor unit, information necessary for the processing request is added to the processing request message and transmitted. A multiprocessor call processing system according to claim 1. 6. When the host processor receives a processing request and performs call processing, when it is determined that a message that cannot be processed by the call processing means that has requested the call processing follows.
The multiprocessor call processing according to any one of claims 1 to 5, wherein a processing end notification is transmitted to the call processing means after all the processing of the subsequent message has been completed by the host processor means. method. 7. When the distributing unit or the call processing unit receives a request to cancel the call setting process, the call processing unit corresponds to the request after the distributing unit or the call processing unit responds to the cancel request. 7. The multiprocessor call processing system according to claim 1, wherein an entry in the call state management table is deleted. 8. The apparatus according to claim 1, wherein said call processing means is configured to perform call processing in a processing function unit.
The multiprocessor call processing system according to any one of the above. 9. A multiprocessor call processing system in which call processing is distributed to a plurality of processors for processing, a distribution means for distributing call processing, a third storage means, and processing of a predetermined type of call. A plurality of call processing means, and at least one host processor means, wherein each of the call processing means, when the processing of a call that cannot be processed by itself is distributed by the distribution means, the host processor means Requesting the processing of the call, and notifying the call requesting the processing to the host processor means and the name of the host processor means to the distribution means, the host processor means processing the call requested by the call processing means. The distributing means, the call received from the call processing means and the name of the host processor means to the third Stores in 憶 unit, the third by referring to the storage contents of the storage means, a multi-processor call processing system subsequent messages addressed to said call and transmits directly to the host processor unit.