JP4834759B2 - Media server, session recovery method, and computer program - Google Patents

Description

  The present invention relates to a media server, a session recovery method, and a computer program that perform session failover when a failure occurs during automatic voice response.

  With landline telephones, end-user telephone charges vary with distance. Therefore, when providing an IVR (Interactive Voice Response) function that performs automatic voice response on a nationwide scale, media servers may be placed throughout the country to shorten the communication distance in order to keep end-user telephone charges low. Many. However, when media servers are arranged throughout the country, there is a problem that costs for equipment and maintenance increase.

  On the other hand, the Internet and IP (Internet Protocol) telephones have been popularized over the past 10 years, and SIP (Session Initiation Protocol), which is a protocol used for real-time services such as VoIP (Voice over IP) and instant messaging, has been standardized. IP telephone services (VoIP services) by communication carriers have been widely provided.

  Since the communication cost does not change depending on the distance in VoIP, if VoIP is used, it is possible to consolidate media servers scattered in various places in one place. In recent years, the quality of IP telephone services has become stable and the reliability has increased, so it is considered that media servers using VoIP will continue to be consolidated in order to reduce costs. However, if the media servers are consolidated, the media server will be affected nationwide at the time of failure, so that the media server is required to have higher reliability than ever before.

  As described above, it is expected that the media server will be required to have high reliability in the future. For this reason, for example, on the hardware side, there is an increasing number of cases where an expensive but high quality and high reliability mainframe is shifted to SI (system integration) using an inexpensive but high failure rate IA server. On the software side, development using inexpensive open source software has increased. Although the cost of open source software can be reduced, the quality may be inferior to that of a vendor product, and the probability of an error occurring on the application side may be high.

  As for system reliability, the problem of hardware failure can be solved by using an FT (fault tolerant) server. In the FT server, all hardware components such as CPU (central processing unit) and memory are duplicated, and even if one of the hardware components breaks down, the system will not stop at all and will continue to operate without instantaneous interruption. can do. If an FT server is introduced, there is no need to worry about hardware failure, but it is very expensive because it is expensive. In addition, the FT server can cope with a hardware failure, but there is a problem that the service is stopped when an application fails due to some error.

  Therefore, in a system that is not mission critical, in order to improve the reliability of the system, an HA (High Availability) configuration that makes the system redundant and minimizes a system stop due to a failure is often adopted. In the HA configuration (also referred to as a cluster configuration), when a failure occurs in the active server, the cluster monitoring application automatically starts the business application on the standby server. The HA configuration includes a shared disk type in which data is placed on a shared disk and the shared disk is used by a plurality of servers, and a data mirror type in which the disk of each server is mirrored between servers. With the HA configuration, a service interruption occurs, but even if either software or hardware fails, the operation can be continued by switching from the active server to the standby server. The HA configuration is a method for ensuring reliability that is widely used in the world, and has been used in non-real-time protocols such as HTTP (Hypertext Transfer Protocol) and DB (database). However, since it may take about 10 seconds to switch from the active system to the standby system, it is not suitable for use in a real-time service such as VoIP.

In many cases, the IVR service includes an IP telephone terminal on the end user side, a relay device that relays SIP packets, and a media server having an IVR function. The repeater is a server managed by the carrier, and has a role of searching for a callee and establishing a call. If the relay machine breaks down, all users who use the IP phone will be unable to talk.
In Patent Document 1 and Patent Document 2, session information is written in an external storage device every time a SIP server that is an active repeater establishes a session, and the standby repeater fails when the active repeater fails. It is described that the system is activated as an active system. In Patent Document 1, the standby relay machine reads information from the external storage device at the time of startup and restores all sessions. In Patent Document 2, the standby relay machine activated as the active system receives the SIP packet. Sometimes information is acquired from the external storage device, and only the received session information is restored.

JP 2007-81933 A JP 2008-103952 A

In general, telephone-type services include many fixed-line telephones and other extremely high-quality services, and end users often consider the quality as a matter of course. Therefore, it is a big problem that the call is disconnected at the time of failure.
On the other hand, companies that try to build a media server with VoIP have so far prioritized cost, and have often built an IVR system with inexpensive hardware that is not an FT server, inexpensive software such as open source software. It was assumed that the failure rate might be high and the reliability would be low.

  However, in the future, since there is a high possibility that a large-scale and high-reliability IVR system will be built on VoIP, it is expected that a session failover function for voice calls will be required. In other words, there is a strong demand for a highly reliable telephone system configuration that can withstand base aggregation, in which processing is taken over from the active system to the standby system without being disconnected even in the event of any application / hardware failure. .

  If the relay machine breaks down, all users who use the IP phone are unable to talk, and thus are basically duplexed. The techniques of Patent Document 1 and Patent Document 2 described above are techniques for restoring a SIP session in a relay device. When attempting to implement the technologies of Patent Document 1 and Patent Document 2 on a media server, there is a problem that the RTP (Real-Time Transport Protocol) session and IVR processing necessary for the media server cannot be restored.

  Moreover, when trying to implement the techniques of Patent Document 1 and Patent Document 2 on a media server, there is a problem that the load on the external device becomes too high. In Patent Document 1 and Patent Document 2, duplexing of repeaters is realized by an HA configuration. If it is a server that does not need to operate in real time, such as a repeater, if the information held by the server is updated, the updated contents are reflected on the shared disk each time, so that the standby system can be used in the event of a failure of the active system. The operation can be continued by referring to the contents of the shared disk in the system server. However, since the media server operates in real time, it cannot be duplicated with a simple HA configuration.

  In other words, the relay station receives only 5 to 6 packets even in a call of about 3 minutes. For example, assuming that 100 end users are talking, 500 to 600 SIP packets are received in 3 minutes. The load for storing the changed information on the shared disk is assumed to be about 3.3 accesses / second. Also, because of the SIP specification, there is no need to respond immediately after receiving a packet, and there is a margin of about 30 seconds.

  However, since the media server operates in real time, data is frequently updated. Since the media server receives a large number of voice packets of 50 packets / second in one call, for example, considering a state in which 100 end users are talking, about 5000 accesses / second to the shared disk Expected to be loaded. In this case, writing to the shared disk is not in time, and there is a high possibility that it cannot be handled by a simple HA configuration. In the case of a voice packet, if a delay of 10 milliseconds or more occurs in transmission, the voice is interrupted. However, voice packets are transmitted and received with a delay of 10 milliseconds or less while applying such a load. Is very difficult.

Furthermore, in Patent Document 1, all SIP packets are stored in an external storage device, and the load of writing and reading is high.
In addition, in Patent Document 1 and Patent Document 2, information necessary for recovery is added and deleted one after another, and there is a possibility of affecting disk utilization efficiency and external storage device response performance.

  In addition to the problems of Patent Documents 1 and 2 above, when it is impossible to achieve high reliability that can withstand the consolidation of bases as described above, and avoidance of hardware failure by FT server with low-cost pressure In addition, it is likely that application errors are likely to occur due to the increased usage of open source software, which may be lower quality than vendor products due to low price pressure, which may cause application errors, It is necessary to realize session failover of the media server in consideration of the fact that the service is required to have higher quality than other services.

  The present invention has been made in view of the above circumstances, and its purpose is to provide a case in which a failure occurs in a media server when voice data is provided via an IP network by an IVR function. An object of the present invention is to provide a media server, a session recovery method, and a computer program capable of recovering a SIP and RTP session and quickly restarting an IVR scenario from the time of occurrence of a failure.

In order to solve the above-mentioned problems, the present invention provides a media server in a voice automatic answering system having a plurality of media servers connected via a network, and a telephone terminal connected via the network and a call control signal A call processing unit that transmits and receives packets and establishes a call session with the telephone terminal, and voice transmission to the telephone terminal that has established a call or voice received from the telephone terminal according to a voice automatic response scenario A voice session that establishes a media session between the voice terminal and the telephone terminal that has established the voice, and the voice that is instructed to be transmitted from the voice auto-response process by the established media session. The encoded voice packet is transmitted to the telephone terminal and the voice packet received from the telephone terminal is transmitted. A voice processing unit that outputs a voice processing unit to the voice automatic response processing unit, and call processing recovery in which call identification information acquired from the call control signal transmitted and received by the call processing unit and information used for session recovery are set Audio processing restoration information in which information is written to an external storage device and the call identification information and information used for restoration of the media session established for the call identified by the call identification information are set in the external storage device , Further, the call specifying information, information specifying a current execution position on a scenario of an automatic voice response executed by the voice automatic response processing unit for the call specified by the call specifying information, and the voice and history information of the result of recognizing the voice received from said telephone terminal according to the scenario of voice response executed by the automatic response unit When the controller that writes the set automatic voice response processing information to the external storage device, the monitoring unit that detects a failure of another media server, and the failure of the other media server in the active system are detected by the monitoring unit, The address used by the other media server is set as its own address, and the call processing restoration information, voice processing restoration information, and voice automatic response having the same call identification information written by the other media server are written. A recovery processing unit that reads processing information from the external storage device, and the voice automatic response processing unit includes information for specifying a current execution position on a scenario from the voice automatic response processing information read by the recovery processing unit ; The history information is acquired, and the voice automatic response scenario is executed from the execution position specified by the acquired information and the history information . The audio processing unit establishes a media session based on the audio processing recovery information read by the recovery processing unit, and encodes audio instructed to be transmitted from the automatic audio response processing unit by the established media session. The voice packet is transmitted to the telephone terminal, and the voice packet received from the telephone terminal is output to the voice automatic response processing unit. The call processing unit is based on the call processing recovery information read by the recovery processing unit. A media server for establishing a call session with the telephone terminal and transmitting / receiving a call processing signal.

  Further, the present invention is the above-described media server, wherein information used for call recovery includes the calling party telephone number, the called party telephone number acquired from the transmitted and received call control signal, and the transmission of the call control signal. Alternatively, it includes the latest value of information whose value is updated upon reception.

  Further, the present invention is the above-described media server, and the information used for call recovery is updated with a call control signal received from the telephone terminal and with the transmission or reception of the call control signal. The restoration processing unit transmits the call control signal set in the call processing restoration information read from the external storage device to the call processing unit, and the call processing unit A call session with the telephone terminal is established based on the call control signal transmitted from the restoration processing unit.

  Further, the present invention is the above-described media server, wherein the information used for media session recovery includes addresses for media sessions on the calling side and called side, and information in which values specific to the media session are set. Including a value at the time of media session establishment of information whose value is updated with the transmission of the voice packet, and a start time of the media session, wherein the voice processing unit reads the voice processing read by the restoration processing unit Obtaining the start time from the recovery information and the value at the time of media session establishment of the information updated with transmission of the voice packet, and the current value of the information updated based on the elapsed time from the start time And the addresses for the media session on the caller side and the callee side read from the voice processing recovery information and the media session A media session of the call is recovered using the information in which a value specific to the application is set and the information in which the calculated value is set and updated in response to transmission of a voice packet. To do.

  The present invention is the above-described media server, wherein the control unit receives the call processing recovery information, call processing recovery information, and voice automatic response processing information including call identification information of a released call. It is deleted from the storage device.

The present invention also relates to a session recovery method used for a media server in an automatic voice response system having a plurality of media servers connected via a network, wherein the call processing unit is a telephone connected via the network. A call processing step of transmitting / receiving a call control signal packet to / from a terminal and establishing a call session with the telephone terminal; and a voice automatic response processing unit to the telephone terminal establishing a call according to a voice automatic response scenario A voice automatic response processing step for executing voice transmission and / or recognition of voice received from the telephone terminal, and a voice processing unit establishes and establishes a media session with the telephone terminal that has established a call. Voice encoded in the voice automatic response processing step by the media session A voice processing step of transmitting a packet to the telephone terminal and outputting a voice packet received from the telephone terminal to the voice automatic response processing unit; and a control unit from the call control signal transmitted and received in the call processing step The call processing recovery information in which the acquired call identification information and information used for call session recovery are set is written to an external storage device, and the call identification information and the call identified by the call identification information are established. Voice processing restoration information in which information used for restoration of the media session is set is written in the external storage device, and the call identification information and the call identified by the call identification information in the voice automatic response processing step Information for identifying the current execution position on the executed voice automatic response scenario, and the voice automatic response And information writing step of writing a voice response processing information set and history information of the result of recognizing the voice received from said telephone terminal according to the scenario of voice response that is executed in physical step in the external storage device, the monitoring A monitoring step for detecting a failure of another media server, and a recovery processing unit used by the other media server when the recovery processing unit detected a failure of another active media server in the monitoring step. Recovery that reads the call processing recovery information, voice processing recovery information, and voice automatic response processing information having the same call identification information written by the other media server from the external storage device while setting the address as its own address A processing step, and the voice automatic response processing unit reads the voice automatic read in the restoration processing step. A voice that acquires information specifying the current execution position on the scenario and the history information from the response processing information, and executes the scenario of the voice automatic response from the execution position specified by the acquired information and the history information The automatic response recovery step and the audio processing recovery establish a media session based on the audio processing recovery information read in the recovery processing step, and are transmitted in the audio automatic response recovery step by the established media session. A voice processing recovery step of transmitting a voice packet encoding the instructed voice to the telephone terminal and receiving a voice packet received from the telephone terminal; and the call processing unit is read in the recovery processing step Based on the call processing recovery information, a call session with the telephone terminal is confirmed. And a session recovery method characterized by having a call processing recovery step of transmitting and receiving call processing signal.

The present invention also transmits and receives a call control signal packet to and from a telephone terminal connected via a network to a computer used as the media server in an automatic voice response system having a plurality of media servers connected via a network. A call processing step for establishing a call session with the telephone terminal, and voice transmission to the telephone terminal that established the call or recognition of voice received from the telephone terminal, or according to a scenario of automatic voice response A voice automatic response processing step that executes both, and a media session established with the telephone terminal that has established a call, and a voice encoded in the voice automatic response processing step by the established media session Send the packet to the telephone terminal and Call processing recovery information for receiving a voice packet; call processing recovery information in which call identification information acquired from the call control signal transmitted and received in the call processing step and information used for call session recovery are set in an external storage device And writing to the external storage device voice processing restoration information in which the call identification information and information used for restoration of the media session established for the call identified by the call identification information are set, and In the voice automatic response processing step, the call specifying information, information specifying a current execution position on a voice automatic response scenario executed in the voice automatic response processing step for the call specified by the call specifying information, and recognizing a voice received from said telephone terminal according to the execution scenario voice response Results and information writing step of writing a voice response processing information to the external storage device is set and the history information, and the monitoring step of detecting a failure of other media servers, other media servers of the primary system in the monitoring step When a failure is detected, the address used by the other media server is set as its own address, and the call processing restoration information and voice processing restoration having the same call identification information written by the other media server are written. A recovery processing step of reading information and voice automatic response processing information from the external storage device; information specifying a current execution position on a scenario from the voice automatic response processing information read in the recovery processing step; and History information is acquired, from the execution position specified by the acquired information and the history information A voice automatic response recovery step for executing a voice automatic response scenario, a media session is established based on the voice processing recovery information read in the recovery processing step, and the voice automatic response recovery step is performed by the established media session. The voice packet encoded in the voice instructed to be transmitted in the voice terminal is transmitted to the telephone terminal, and the voice processing recovery step of receiving the voice packet received from the telephone terminal, and the call read in the recovery processing step A computer program characterized by executing a call processing recovery step of establishing a call session with the telephone terminal based on processing recovery information and transmitting / receiving a call processing signal.

  According to the present invention, even if the active media server fails during voice transmission, the standby media server continues to provide the IVR scenario from the time of the failure without interruption of the call. Can do. As described above, since the session failover of the media server can be realized, the quality is inferior when compared to the vendor product such as the case of using the hardware having a higher failure rate than the FT server or the open source software. Even when possible software is used, a highly reliable system can be realized. Therefore, it is possible to reduce the installation, operation, and maintenance costs because it is possible to realize the consolidation of media servers using inexpensive hardware and software.

1 is a functional block diagram of an IVR system according to a first embodiment of the present invention. FIG. It is a figure which shows the data structural example of the telephone number setting information by the embodiment. It is a figure which shows the example of a data structure of the IVR flow data by the embodiment. It is a figure which shows the example of a data structure of the call process recovery information by the embodiment. It is a figure which shows the data structural example of the audio | voice process recovery information by the embodiment. It is a figure which shows the data structural example of the IVR process information by the embodiment. It is a figure which shows the sequence at the time of the incoming call by the same embodiment. It is a figure which shows the sequence at the time of the failure generation by the embodiment. It is a figure which shows the monitoring flow by the embodiment. It is a figure which shows the processing flow at the time of the incoming call of the active media server by the embodiment. It is a figure which shows the processing flow at the time of the incoming call of the active media server by the embodiment. It is a figure which shows the processing flow at the time of the transmission of the active media server by the embodiment. It is a figure which shows the restoration processing flow of the standby | mediate media server by the same embodiment. It is a figure which shows the example of a data structure of the call process recovery information by 2nd Embodiment. It is a figure which shows the example of a data structure of the audio | voice process recovery information by the embodiment. It is a figure which shows the sequence at the time of the failure generation by the embodiment. It is a figure which shows the processing flow at the time of the incoming call of the active media server by the embodiment. It is a figure which shows the processing flow at the time of the transmission of the active media server by the embodiment. It is a figure which shows the restoration processing flow of the standby | mediate media server by the same embodiment.

  Hereinafter, exemplary embodiments of the present invention will be described with reference to the drawings.

[First Embodiment: Memory Information Sharing Method]
In the present embodiment, session failover of voice calls by the memory information sharing method is realized. In the memory information sharing method, an active media server in an IVR (Interactive Voice Response) system stores all SIP (Session Initiation Protocol) related information in an external storage device, and this active media server If a failure occurs, the call processing is continued using the SIP-related information stored in the external storage device by the standby media server.

  FIG. 1 is a functional block diagram of an IVR system according to an embodiment of the present invention. IVR automatically answers end-user calls and plays voices such as “Please press No. 1 for XX” to guide the user and transfer it to the operator, An automatic voice response system having an automatic voice response function for providing information to the user.

  As shown in the figure, the IVR system includes a telephone terminal 1, a relay server 2, a media server 4, an external storage device 6, and a call control device 7. The telephone terminal 1 and the relay server 2 are connected to a network N that is an IP (Internet Protocol) network. Although only one telephone terminal 1 is shown in the figure, a plurality of telephone terminals 1 are connected to the network N. The two media servers 4 are connected to the network N via the router 3, and the media server 4, the external storage device 6, and the transmission control device 7 are connected via a private network such as a LAN (Local Area Network). Is done.

  Although two media servers 4a and 4b are shown as the media server 4 in the figure, three or more media servers may be provided. Here, it is assumed that the media server 4a is an active system that is currently operating, and the media server 4b is a standby system that takes over processing when the active system fails. Note that when there are three or more media servers 4, the ratio of the number of active and standby servers is N to M (N and M are 1 or more).

  In a voice call session failover, if the active media server 4 fails, the operation is taken over by the standby media server 4 and the call is continued. For example, after an announcement from the active media server 4a “Please push # 1 for XX” and the end user inputs “1” to the telephone terminal 1, the media server 4a again “△△ Suppose that a failure of the media server 4a occurs in the middle of an announcement such as In this case, the standby media server 4b is started and resumes from the announcement in the middle of the failure, and the announcement ““ △, please press No. 2 ”is continued and the IVR process is continued. . Silence may occur for several seconds when switching from the active system to the standby system.

  Information necessary for switching to the standby system is SIP-related information, RTP (Real-Time Transport Protocol) -related information, and IVR-related information. SIP is a protocol for call control that is often used to find out a party to call with an IP telephone, and detailed specifications are described in RFC (Request for Comments) 3261. RTP is a protocol used for transmitting voice data and the like in real time, and voice data obtained by encoding analog voice is set in the RTP packet.

  In the memory information sharing method applied to the IVR system according to the present embodiment, the working media server 4 stores all SIP-related information in the external storage device 6, and basically the working media server is its own. SIP related information to be taken over by the standby media server is not stored in the memory. Therefore, even if the active media server 4 fails, the standby media server 4 does not need to restore SIP-related information on the memory.

  Further, RTP information for recovery is stored in the external storage device 6 at the start of the RTP session. Therefore, when the active media server 4 fails, the standby media server 4 refers to the information stored in the external storage device 6 and performs state return processing such as RTP port open, and takes over the session. For some information, it is necessary to estimate the current value from the stored value and restart RTP packet transmission.

  The information related to IVR is saved in the external storage device 6 as soon as it is updated. Therefore, when the active media server 4 fails, the standby media server 4 activates the IVR again based on the information stored in the external storage device 6. Therefore, the IVR process can be resumed from the time when the failure occurs.

  As described above, in the memory information sharing method, all the memory information of the active media server 4 is stored in the external storage device 6, and the standby media server 4 acquires information from the external storage device 6 when a failure occurs. Thus, the memory information is restored. In the case of this method, session failover is possible even if a failure occurs during the transfer of the bridge method. In addition, when switching to the standby system, it is not necessary to restore SIP-related information on the memory, so that it is possible to switch from the active system to the standby system in a short stop time.

Next, the configuration of each device will be described.
The telephone terminal 1 possessed by the end user is an IP telephone terminal, and establishes a call with the media server 4 by SIP, establishes a media session using RTP, and transmits and receives voice packets.
The relay server 2 performs address resolution or the like and relays the SIP signal.

  The media server 4 includes a communication unit 40, a monitoring unit 41, and a media processing program execution unit 42. The communication unit 40 transmits and receives packets (data) to and from other devices by IP. When the communication unit 40 takes over processing from the active system, the communication unit 40 takes over and uses the IP address used by the active media server 4. The monitoring unit 41 detects a failure of another media server 4.

  The media processing program execution unit 42 includes a call processing unit 43, a control unit 44, a main storage unit 45, an IVR processing unit 46, an audio processing unit 47, and a restoration processing unit 48. The main storage unit 45 is accessible from each unit in the media processing program execution unit 42, and stores various data used for call processing, voice processing, and IVR processing. The control unit 44 writes / reads call processing recovery information, voice processing recovery information, and IVR processing information to / from the external storage device 6. The call processing unit 43 performs the same call processing as that based on the SIP signal, but writes information that needs to be taken over at the time of failure out of various information used in the call processing to the external storage device 6 via the control unit 44. The call processing is performed with reference to information in the external storage device 6 as appropriate. The IVR processing unit 46 instructs the audio processing unit 47 to output audio data according to the IVR flow. The voice processing unit 47 performs voice data transmission / reception processing by RTP.

  Hereinafter, each part of the media server 4 included in the media server 4a is described by adding a to the reference numeral, and each part of the media server 4 included in the media server 4b is described by adding b to the reference numeral. For example, the monitoring unit 41 included in the media server 4a is described as a monitoring unit 41a, and the monitoring unit 41 included in the media server 4b is described as a monitoring unit 41b.

The external storage device 6 includes a storage unit 60, and a telephone number setting storage unit 61, an IVR flow storage unit 62, a call processing recovery information storage unit 63, a voice processing recovery information storage unit 64, and an IVR processing information storage unit in the storage unit 60. The data stored in 65 is read and written.
The transmission control device 7 instructs the media server 4 to transmit a call.

Next, each data stored in the external storage device 6 according to the present embodiment will be described.
FIG. 2 is a diagram illustrating a data configuration example of the telephone number setting information stored in the telephone number setting storage unit 61 according to the present embodiment. The telephone number setting information correlates the telephone number, the IVR ID that identifies the IVR scenario to be executed, the maximum number of channels indicating the number of simultaneously connectable calls, and the number of active channels indicating the number of currently connected calls. Consists of records. The telephone number is transmitted from the telephone terminal 1 and received when receiving the call to the media server 4 or when the media server 4 transmits and receives an instruction from the call control device 7 and receives the call to the telephone terminal 1. The original telephone number.

  FIG. 3 is a diagram illustrating a data configuration example of IVR flow data stored in the IVR flow storage unit 62 according to the present embodiment. In the figure, the IVR flow data indicates the correspondence between the IVR ID and the IVR flow scenario. The IVR flow scenario is data describing the order of the IVR processing to be executed for the end user, and the IVR processing unit 46 of the media server 4 reproduces the audio to be transmitted and the received audio according to the IVR flow scenario. Processes such as recognition are executed in order. The received voice is, for example, DTMF (Dual Tone Multi Frequency) or an end user's speech. DTMF is a method of transmitting numbers 1 to 9 and 16 types of codes such as *, #, A, B, C, and D as signal sounds in two audio frequency bands of a low group and a high group. The number pushed to the telephone terminal 1 can be transmitted to the media server 4.

  FIG. 4 is a diagram showing a data configuration example of call processing recovery information stored in the call processing recovery information storage unit 63 according to the present embodiment. The call processing recovery information holds information necessary for re-establishing a SIP session between the telephone terminal 1 and the media server 4, and includes a to tag, a from tag, a call, a CSeq of the media server, a CSeq of the telephone terminal, It consists of a record in which an outgoing / incoming flag, a server side telephone number, a terminal side telephone number, and various header fields are associated. Each record is generated in a call unit, that is, in a dialog unit. In SIP, a dialog is a period from when a telephone call is established to when the telephone call is terminated. A call can be uniquely identified by call identification information indicated by a combination of a to tag, a from tag, and a call.

  The to tag is a SIP tag in which a logical request destination (different from the IP address) is described. The from tag is a SIP tag in which a transmission source of a logical request (different from the IP address) is described. The callid is a globally unique random ID (identifier) and is recommended to be generated by the RFC1750 method. CSeq is a number incremented for each new request. The outgoing / incoming flag is a flag for determining whether the call is outgoing from the media server 4 or incoming from the telephone terminal 1. The various header fields are information on all or part of the header information of each transmitted / received SIP message. Details of the headers of the to tag, the from tag, callid, CSeq, and SIP message are described in RFC3261. Instead of the server-side telephone number and the terminal-side telephone number, the SIP URI (Universal Resource Identifier) of the media server 4 and the SIP URI of the telephone terminal 1 may be used.

  FIG. 5 is a diagram illustrating a data configuration example of the audio processing recovery information stored in the audio processing recovery information storage unit 64 according to the present embodiment. The voice processing restoration information holds information necessary for re-establishing voice communication by RTP between the telephone terminal 1 and the media server 4, and includes a to tag, a from tag, a call, an outgoing / incoming start time, an RTP sequence number of the media server. , SSRC, RTP port number and IP address, and RTP port number and IP address of the telephone terminal. Each record is generated in dialog (call) units.

SSRC is the source ID of the media stream included in the RTP packet, and is used to distinguish it from other RTPs. The RTP sequence number is a number for ensuring the order of RTP packets, and is incremented by 1 each time an RTP packet is transmitted. Here, in the established RTP session, it is used for the first RTP packet transmitted. The RTP sequence number is shown.
Details of the SSRC and RTP sequence numbers are described in RFC3350.

  FIG. 6 is a diagram illustrating a data configuration example of IVR processing information stored in the IVR processing information storage unit 65 according to the present embodiment. The IVR processing information indicates a history log of commands executed by the media server 4 for the end user, and includes a record in which a to tag, a from tag, a call, a command execution history, and end user input histories 1 to n are associated with each other. Generated in dialog (call) units.

  The command execution history specifies an IVR process such as guidance transmission or voice recognition which is instructed last in a series of IVR processes defined by the IVR flow scenario. The end user input histories 1 to n are data in which the input contents by the end user received from the telephone terminal 1 in the IVR process are described in order, and indicate input DTMF, selection contents specified by voice or DTMF, and the like.

Next, the operation of the IVR system according to the present embodiment will be described.
It is assumed that the telephone number setting information is stored in the telephone number setting storage unit 61 and the IVR flow data is stored in the IVR flow storage unit 62 before the operation of the IVR system. However, the initial value of the number of active channels in the telephone number setting information is “0”.

  FIG. 7 is a diagram showing a sequence example at the time of an incoming call from the telephone terminal 1 to the active media server 4 according to the present embodiment. Hereinafter, an IP packet in which the SIP X signal is set is referred to as an X packet. For example, an IP packet in which the INVITE signal is set is defined as an INVITE packet.

  First, the telephone terminal 1 transmits an INVITE packet to the relay server 2 (step S105). In the INVITE packet, an RTP port number and an IP address used by the telephone terminal 1 are set. The relay server 2 converts the telephone number set in the To field of the INVITE packet into the destination IP address and transmits it to the media server 4a (step S110). If the media server 4a determines that the incoming call is permitted, the media server 4a returns a 100 Tring message to the relay server 2 (step S115), and then transmits 180 Ringing to the relay server 2 (step S120). The relay server 2 relays 180 Ringing to the telephone terminal 1 (step S125), and the telephone terminal 1 allows the end user to hear a ringback tone.

  Subsequently, the call processing unit 43a of the media server 4a transmits 200 OK to the relay server 2 (step S130). In 200 OK, an RTP port number and an IP address used by the media server 4a are set. The relay server 2 relays 200 OK to the telephone terminal 1 (step S135). The telephone terminal 1 transmits ACK to the media server 4a (step S140). When the RTP session is established between the telephone terminal 1 and the voice processing unit 47a of the media server 4a along with the transmission of the ACK and transitions to a busy state, the voice processing unit 47a of the media server 4a converts the IVR voice into an RTP packet. The telephone terminal 1 transmits the end user's speech and DTMF to the media server 4a by RTP packet (step S145).

  If the busy state continues for a long time, a re-INVITE packet or an UPDATE packet is periodically transmitted from the call processing unit 43a of the media server 4a or the telephone terminal 1 (step S150), and the received telephone terminal 1 or media server 4a The call processing unit 43a returns a response and confirms the continuation of the session.

  When the end user performs a call disconnecting operation, the telephone terminal 1 transmits a BYE packet to the media server 4a (step S155), and the call processing unit 43a of the media server 4a returns 200 OK (step S160). When the media server 4a disconnects the call, the call processing unit 43a of the media server 4a transmits a BYE packet to the telephone terminal 1, and the telephone terminal 1 returns 200 OK.

  In the case of transmission from the media server 4, the media server 4 performs the same operation as the telephone terminal 1 in the sequence shown in FIG. 7, and the telephone terminal 1 performs the same operation as the media server 4 shown in FIG.

FIG. 8 is a diagram showing a sequence when a failure occurs in the IVR system according to the present embodiment.
For example, when a SIP session is established between the telephone terminal 1 and the media server 4a by the procedure for making a call from the telephone terminal 1 and the procedure for making a call from the media server 4 shown in steps S105 to S140 in FIG. In step S205, the media server 4a writes the call processing recovery information in the call processing recovery information storage unit 63 of the external storage device 6 (step S210). When the media server 4a starts the IVR process (step S215) and establishes an RTP session with the telephone terminal 1 (step S220), the audio process recovery information is stored in the audio process recovery information storage unit 64 of the external storage device 6. Is written (step S225).

  The media server 4a executes IVR processing such as audio reproduction (step S230), and writes IVR processing information in the IVR processing information storage unit 65 of the external storage device 6. Further, when the SIP packet for confirming the continuation of the call is transmitted / received during the IVR process, the media server 4a reads the call process recovery information stored in the call process recovery information storage unit 63 of the external storage device 6. Or updating is performed (step S235). In other words, the media server 4a updates the IVR processing information every time the IVR processing proceeds, reads the call processing recovery information every time a SIP packet is received from the telephone terminal 1, and checks whether it is a new SIP packet. In this case, the call processing recovery information is updated.

  When the media server 4a fails during the execution of the IVR process, the media server 4b detects this (step S240), and reads the IVR process information from the IVR process information storage unit 65 of the external storage device 6 (step S245). The IVR process is activated from the middle of the scenario (step S250). Further, the media server 4b reads the voice processing restoration information from the voice processing restoration information storage unit 64 of the external storage device 6 (step S255), opens the RTP port indicated by the voice processing restoration information and sets the reception state at the same time. The RTP packet transmission is resumed to restore the RTP session (step S260), and the RTP session with the telephone terminal 1 is reestablished (step S265). When the media server 4b receives a SIP packet from the telephone terminal 1 (step S270), the media server 4b reads the call processing recovery information from the call processing recovery information storage unit 63 of the external storage device 6, and the received SIP packet is a packet for which a session has been established. (Step S275).

Next, a processing flow in the media server 4 will be described.
FIG. 9 is a diagram showing a monitoring flow according to the present embodiment.
In the figure, the monitoring unit 41 of the media server 4 activates monitoring of the life and death of other media servers 4 at a predetermined time interval (step S305). That is, the monitoring unit 41a of the media server 4a starts the life / death monitoring process of the media server 4b, and the monitoring unit 41b of the media server 4b starts the life / death monitoring process of the media server 4a.

  When the monitoring is activated, the monitoring unit 41 of the media server 4 exchanges information for monitoring the life and death with the monitoring unit 41 of the other media server 4 (step S310). The monitoring unit 41a of the media server 4a transmits a monitoring message to the monitoring unit 41b of the media server 4b, and the monitoring unit 41b of the media server 4b that has received the monitoring message performs hardware operations and media processing of the media server 4b. When the processing of the program execution unit 42b is normal, a monitoring response message notifying normality is returned. Similarly, the monitoring unit 41b of the media server 4b transmits a monitoring message to the monitoring unit 41a of the media server 4a, and the monitoring unit 41a of the media server 4a that has received the monitoring message operates the hardware of the media server 4a. When the processing of the media processing program execution unit 42a is normal, a monitoring response message notifying normality is returned.

  When both the active and standby media servers 4 are normal, that is, when the media servers 4a and 4b receive a monitoring response message for notifying normality in response to the monitoring message transmitted by themselves in step S310 (step S315: No failure detected), the processing continues for both the active and standby media servers 4a and 4b (step S320).

  When a failure of the standby media server 4b is detected, for example, the monitoring unit 41a of the media server 4a has not received a monitoring response message within a predetermined time after transmitting the monitoring message to the media server 4b. Alternatively, when a monitoring response message indicating a failure is received from the media server 4b (step S315: standby failure detection), the active media server 4a continues the processing (step S325). An operation server (not shown) connected to the media server 4 via the LAN and managing the operation of the entire IVR system transmits a mail addressed to the administrator and notifies the failure.

  When a failure of the active media server 4a is detected, for example, the monitoring unit 41b of the media server 4b has not received a monitoring response message within a predetermined time after transmitting the monitoring message to the media server 4a. Alternatively, when a monitoring response message indicating a failure is received from the media server 4a (step S315: failure detection of the active system), the monitoring unit 41b of the standby media server 4b sends the active media server to the communication unit 40b. In addition to instructing to take over the IP address of 4a, the media processing program execution unit 42b is instructed to execute a failure flow described later (step S330).

10 and 11 are diagrams showing a processing flow at the time of an incoming call in the active media server 4 according to the present embodiment.
10, the call processing unit 43a of the media server 4a receives the INVITE packet transmitted from the telephone terminal 1 and relayed by the relay server 2 as shown in steps S105 to S110 in FIG. 7 (step S405).

The call processing unit 43a of the media server 4a acquires the destination telephone number from the To field of the INVITE packet and writes it in the main storage unit 45a (step S410). Upon receiving an instruction from the call processing unit 43a, the control unit 44a notifies the external storage device 6 of the acquired telephone number and instructs to read out the maximum channel number and the active channel number. The control unit 44a receives the maximum number of channels and the number of active channels that the external storage device 6 has read out from the telephone number setting information in the telephone number setting storage unit 61 corresponding to the telephone number, and outputs it to the call processing unit 43a. The call processing unit 43a permits an incoming call when the number of active channels is smaller than the maximum number of channels (step S415).
If the number of active channels or the maximum number of channels corresponding to the telephone number cannot be read from the telephone number setting information, or if the number of active channels is equal to or greater than the maximum number of channels, the call processing unit 43a establishes a dialog. Rather, the call is terminated.

  When permitting an incoming call, the call processing unit 43a establishes a dialog according to the SIP specification and writes SIP information necessary for resuming the call into the main storage unit 45a (step S420). Specifically, it operates as follows.

  The call processing unit 43a writes the received INVITE packet to the main storage unit 45a, and also sets the from tag, callid, various header fields set in the INVITE packet, the RTP port number and IP address used by the telephone terminal 1, and the telephone terminal CSeq is read and written to the main memory 45a. Note that the RTP port number and IP address may be obtained from the ACK packet.

  For example, as shown in the processing of steps S115 to S140 in FIG. 7, the call processing unit 43a transmits 180 Tring, Ringing, 200 OK, and receives ACK according to the SIP specifications. The call processing unit 43a writes each of these transmitted / received SIP packets in the main storage unit 45a, and also stores the CSeq of the telephone terminal set in these SIP packets, the CSeq of the CSeq media server of the media server, and various header fields. Write to the storage unit 45a.

Note that when transmitting 200 OK, the call processing unit 43a writes the to tag set to 200 OK, the RTP port number and the IP address used by the media server 4a, to the main storage unit 45a.
The various header fields include the To, From, Record-Route, and Via header fields. However, the Record-Route and Via header fields may not be acquired.

  After the dialog with the telephone terminal 1 is established, the control unit 44a notifies the external storage device 6 of the telephone number acquired in step S410 and instructs addition of the number of active channels. The external storage device 6 identifies the telephone number setting information record in the telephone number setting storage unit 61 by the telephone number, and adds 1 to the number of active channels set in the identified record (step S425).

  Next, the control unit 44a that has received an instruction from the call processing unit 43a of the media server 4a, the to tag, the from tag, the call, the CSeq of the media server, the CSeq of the telephone terminal, and the INVITE packet stored in the main storage unit 45a. Server side telephone number which is the telephone number of the receiving side acquired from the To field of the terminal, telephone number of the terminal side which is the telephone number of the calling side acquired from the From field of the INVITE packet, various header fields, and an outgoing / incoming flag indicating incoming Is output to the external storage device 6 as registered call processing recovery information, and registration to the call processing recovery information is instructed (step S430). The external storage device 6 adds a record to the call processing recovery information in the call processing recovery information storage unit 63, and sets the registered call processing recovery information received from the media server 4a in the added record.

Further, the control unit 44a receiving the instruction from the call processing unit 43a of the media server 4a, the to tag, the from tag, and the call stored in the main storage unit 45a, the RTP port number and IP address used by the media server 4a, the telephone The RTP port number and IP address used by the terminal 1 are output to the external storage device 6 as registered voice processing recovery information, and an instruction to register the voice processing recovery information is given (step S435). The external storage device 6 adds a record to the audio processing recovery information in the audio processing recovery information storage unit 64, and sets the registered audio processing recovery information received from the media server 4a in the added record.
The call processing unit 43a of the media server 4a notifies the establishment of the dialog to the control unit 44a.

  Receiving the dialog establishment notification, the control unit 44a reads the destination telephone number from the main storage unit 45a, notifies the external storage device 6 of the read telephone number, and instructs to read the IVR ID and the maximum number of channels. To do. The control unit 44a of the media server 4a receives the IVR ID and the maximum number of channels read from the telephone number setting information in the telephone number setting storage unit 61 by the external storage device 6 corresponding to the telephone number, and stores it in the main storage unit 45a. Writing is performed (step S440).

  Subsequently, the control unit 44a notifies the external storage device 6 of the received IVR ID and instructs the reading of the IVR flow scenario. The control unit 44a receives the IVR flow scenario read from the IVR flow data in the IVR flow storage unit 62 corresponding to the IVR ID by the external storage device 6, and writes it in the main storage unit 45a (step S445).

  Next, the control unit 44a notifies the external storage device 6 of the to tag, the from tag, and the callid acquired from the main storage unit 45a and instructs to read the command execution history. The control unit 44a receives the command execution history read from the IVR processing information in the IVR processing information storage unit 65 by the external storage device 6 corresponding to the to tag, the from tag, and the call, and writes the command execution history in the main storage unit 45a. (Step S450). If the command execution history cannot be read, it is determined that the current state is the initial state.

  In FIG. 11, the control unit 44a outputs the IVR scenario and the command execution history written in the main storage unit 45a to the IVR processing unit 46a, and instructs execution of the IVR flow. If the IVR processing unit 46a determines that there is no command execution history and is in the initial state, the IVR processing unit 46a executes the IVR scenario from the beginning. If the executed command is set in the command execution history, the IVR scenario is executed from the IVR process halfway indicated by the command (step S505).

  When the voice processing unit 47a receives an instruction from the control unit 44a and generates the SSRC and RTP sequence number used by the own media server 4a, the RTP port number and IP address used by the own media server 4a from the main storage unit 45a, and the telephone The RTP port number and IP address used by the terminal 1 are read out, and using these pieces of information, an RTP session is established with the telephone terminal 1 and transmission / reception of RTP packets is started (step S510). The control unit 44a writes the SSRC of the media server, the RTP sequence number set in the first transmitted RTP packet, and the outgoing / incoming start time indicating the time when the RTP session is started in the main storage unit 45a.

  When the RTP session is established and the RTP packet is transmitted / received, the control unit 44a transmits the to tag, the from tag, the call, the SSRC of the media server, and the media server 4a in the RTP session with the telephone terminal 1 from the main storage unit 45a. The RTP sequence number used first and the arrival / departure start time, which is the time when the RTP packet is first transmitted in the RTP session, are notified to the external storage device 6, and the update of the voice processing recovery information is instructed (step S515). The external storage device 6 identifies the record of the voice processing restoration information in the voice processing restoration information storage unit 64 by the to tag, the from tag, and the call, and starts the arrival and departure of the SSRC and RTP sequence number of the media server 4 in the identified record. Write time and. In addition, when the record corresponding to the to tag, the from tag, and the callid is not registered in the audio processing recovery information, a new record is added, and the SSRC and RTP sequence number of the media server 4 and the arrival / departure start time are added to the added record. And write.

  The media server 4a executes the IVR scenario. Specifically, the voice processing unit 47a transmits an RTP packet of the playback voice to the telephone terminal 1 in response to an instruction from the IVR processing unit 46a, and gives the end user guidance or receives the RTP received from the telephone terminal 1. The IVR processing unit 46a recognizes the DTMF or the utterance content indicated by the packet. The IVR processing unit 46a instructs the control unit 44a to add the recognition result to the IVR processing information every time it recognizes speech such as DTMF or end user utterance. Further, every time one command such as voice recognition or voice playback is executed, the IVR processing unit 46a controls to write information specifying the position of the currently executed IVR flow in the IVR processing information storage unit 65 as a command execution history. The unit 44a is instructed. Specifically, the control unit 44a notifies the external storage device 6 of the update of the IVR processing information by notifying the to tag, the from tag, and the call of the call currently being processed, the recognition result or the command execution history ( Step S520). The external storage device 6 specifies the record of the IVR processing information in the IVR processing information storage unit 65 using the to tag, the from tag, and the call, and writes the recognition result or the command history in the specified record.

If the call continues for a predetermined time such as 150 seconds or longer (step S525: YES), the media server 4a performs a session continuation confirmation process between executions of the IVR flow. That is, the call processing unit 43a of the media server 4a periodically transmits a re-INVITE packet or an UPDATE packet to the telephone terminal 1 when detecting that the call is continued for a predetermined time or more (step S530).
The control unit 44a updates the call processing recovery information with the CSeq of the media server updated by the call processing unit 43a transmitting the re-INVITE packet or the UPDATE packet. That is, the control unit 44a notifies the external storage device 6 of the to tag, the from tag, and the call of the call currently being processed and the CSeq of the updated media server, and instructs the update of the call processing recovery information. The external storage device 6 identifies the call processing recovery information record in the call processing recovery information storage unit 63 by the to tag, from tag, and call, and receives the CSeq of the media server described in the specified record. Update with

  Further, the call processing unit 43a of the media server 4a may receive a re-INVITE packet or an UPDATE packet from the telephone terminal 1. In this case, the call processing unit 43a reads the to tag, the from tag, the call, and the CSeq of the telephone terminal from the received re-INVITE packet or UPDATE packet, and outputs them to the control unit 44a. The control unit 44a notifies the external storage device 6 of the to tag, from tag, callid, and CSeq of the telephone terminal received from the call processing unit 43a, and instructs to update the call processing recovery information. The external storage device 6 identifies the call processing recovery information record in the call processing recovery information storage unit 63 by the to tag, from tag, and call, and receives the CSeq of the telephone terminal described in the specified record. Update with CSeq of terminal.

When the call is terminated due to the communication disconnection from the telephone terminal 1, a BYE packet is transmitted from the telephone terminal 1 to the media server 4a (step S525: NO, step S530: communication disconnection from the end user). When receiving the BYE packet transmitted from the telephone terminal 1, the call processing unit 43a of the media server 4a acquires a to tag, a from tag, and a call from this packet and outputs the acquired to tag to the control unit 44a.
The control unit 44a outputs the to tag, the from tag, and the call received from the call processing unit 43a to the IVR processing unit 46a, and instructs the end of the IVR processing. The control unit 44a ends the execution of the IVR flow scenario executed for the call specified by the to tag, the from tag, and the call received from the call processing unit 43a (step S540).
Further, the control unit 44a outputs the to tag, the from tag, and the call received from the call processing unit 43a to the voice processing unit 47a, and instructs the end of the RTP session. The voice processing unit 47a ends the session established for the call specified by the to tag, the from tag, and the call received from the call processing unit 43a (step S545).
The call processing unit 43a returns 200 OK to the telephone terminal 1 as a response to the BYE packet (step S550).

Alternatively, when the IVR flow is executed to the end, the call is terminated due to the communication disconnection from the media server 4a (step S525: NO, step S530: communication disconnection from the media server). The IVR processing unit 46a notifies the control unit 44a of the end of IVR flow execution.
When the control unit 44a notifies the call processing unit 43a of the end of IVR flow execution, the call processing unit 43a transmits a BYE packet to the telephone terminal 1 (step S555), and instructs the telephone terminal 1 to end the call. Furthermore, the control unit 44a outputs a stop command to the voice processing unit 47a to instruct the end of the RTP session. The voice processing unit 47a ends the session established with the telephone terminal 1 (step S560). Finally, the call processing unit 43a receives 200OK from the telephone terminal 1 as a response to the BYE packet (step S565).

  The control unit 44a of the media server 4a that has recognized that the call has been terminated by any of the above, reads the to tag, the from tag, and the call of the terminated call from the main storage unit 45a after step S550 or S565, and stores them externally. It transmits to the device 6 and instructs to delete call processing recovery information, voice processing recovery information, and IVR processing information. The external storage device 6 deletes the call processing recovery information record, the voice processing recovery information record, and the IVR processing information record specified by the to tag, the from tag, and call (step S570).

  Further, the control unit 44a of the media server 4a transmits the IVR ID of the IVR flow scenario being executed for the terminated call to the external storage device 6 and instructs the subtraction of the number of active channels (step S575). The external storage device 6 identifies the telephone number setting information record stored in the telephone number setting storage unit 61 by the IVR ID, and updates the number of active channels set in the identified record to a number obtained by subtracting 1 To do.

FIG. 12 is a diagram showing a processing flow at the time of outgoing call in the active media server 4 according to the present embodiment.
The control unit 44a of the media server 4a receives a call instruction from the call control device 7 (step S605). A server-side telephone number and a terminal-side telephone number are set in the outgoing call instruction.

The call processing unit 43a of the media server 4a notifies the external storage device 6 of the server-side telephone number acquired from the outgoing call instruction, and instructs to read out the maximum channel number and the active channel number (step S610). When the external storage device 6 receives the maximum channel number and the active channel number read from the telephone number setting information in the telephone number setting storage unit 61 corresponding to the telephone number, the control unit 44a compares these. The control unit 44a permits an incoming call when the number of active channels is smaller than the maximum number of channels, and instructs the call processing unit 43a to make a call.
If the server side telephone number is not set in the telephone number setting information, or if the number of active channels is equal to or greater than the maximum number of channels, the control unit 44a does not instruct outgoing and ends the process.

  The call processing unit 43a that has received a call origination instruction from the control unit 44a establishes a dialog with the telephone terminal 1 in accordance with the SIP specification, and writes SIP information necessary for resuming the call into the main storage unit 45a (step S615). . Specifically, it operates as follows.

The call processing unit 43a generates an INVITE packet and transmits it to the telephone terminal 1. The terminal-side telephone number received from the transmission control device 7 is set in the To field of the INVITE packet, and the server-side telephone number received from the transmission control device 7 and the generated from tag are set in the From field.
The control unit 44a writes the INVITE packet to the main storage unit 45a, and also reads the from tag, callid, various header fields, the RTP port number and IP address of the media server 4, and the CSeq of the media server that are set in the INVITE packet. Write to part 45a. Note that the RTP port number and IP address may be set in the ACK packet.

The call processing unit 43a receives 180 Tring, Ringing, 200 OK, and transmits ACK according to general SIP specifications, and establishes a dialog with the telephone terminal 1. The call processing unit 43a writes the transmitted / received SIP packets to the main storage unit 45a, and also stores the CSeq of the telephone terminal, the CSeq of the media server, and various header fields set in these SIP packets in the main storage unit 45a. Write.
The call processing unit 43a acquires the to tag, the RTP port number and the IP address used by the telephone terminal 1 from the 200 OK received from the telephone terminal 1, and writes them in the main storage unit 45a.
The various header fields include the To, From, Record-Route, and Via header fields. However, the Record-Route and Via header fields may not be acquired.

  After the dialog with the telephone terminal 1 is established, the control unit 44a notifies the external storage device 6 of the telephone number acquired from the call instruction received in step S610, and instructs the addition of the number of active channels. The external storage device 6 identifies the telephone number setting information record in the telephone number setting storage unit 61 based on the telephone number, and adds 1 to the number of active channels set in the identified record (step S620).

  Thereafter, the media server 4a performs the same processing as that after step S430 in FIG. However, in step S430, the registered call processing recovery information includes the outgoing / incoming flag indicating the outgoing call, the server side telephone number that is the telephone number of the outgoing side obtained from the INVITE From field, and the incoming side obtained from the To field of the INVITE packet. Is set as the terminal-side telephone number.

FIG. 13 is a diagram showing a restoration processing flow executed by the standby media server 4 when a failure of the active media server 4 according to the present embodiment is detected.
When a failure of the media server 4a occurs during the IVR process, the monitoring unit 41b of the standby media server 4b detects the failure of the active system as shown in FIG. The monitoring unit 41b of the media server 4b activates the recovery processing unit 48 (step S705). The recovery processing unit 48b causes the communication unit 40b to take over the IP address used by the communication unit 40a of the media server 4a, and the media server 4b is activated as the active system.

  The recovery processing unit 48b of the media server 4b specifies call processing recovery information stored in the call processing recovery information storage unit 63 of the external storage device 6 one record at a time, reads the specified record, and writes it to the main storage unit 45b. At the same time, the to tag, from tag, and callid are read from the record (step S710), and the following processing is performed. Note that only the to tag, the from tag, and the callid may be read from the external storage device 6.

  The control unit 44b notifies the external storage device 6 of the set of the to tag, the from tag, and the call acquired in step S710, and instructs to read out the audio processing recovery information and the IVR processing information (step S715). The control unit 44b includes a record of audio processing recovery information in the audio processing recovery information storage unit 64 read by the external storage device 6 corresponding to each set of to tag, from tag, and call, and an IVR processing information storage unit 65. The record in the IVR processing information is received and written to the main storage unit 45b.

  Further, when the control unit 44b acquires the server-side telephone number from the call processing recovery information record stored in the main storage unit 45b in step S710, the control unit 44b notifies the external storage device 6 of the acquired server-side telephone number and notifies the IVR ID. Is instructed to read. The control unit 44b receives the IVR ID read from the telephone number setting information in the telephone number setting storage unit 61 by the external storage device 6 corresponding to the server side telephone number. If the IVR flow corresponding to the IVR ID has not been read yet, the control unit 44b notifies the external storage device 6 of the received IVR ID and instructs the reading of the IVR flow scenario. The control unit 44b receives the IVR flow scenario read from the IVR flow data in the IVR flow storage unit 62 corresponding to the IVR ID by the external storage device 6, and writes the IVR flow scenario in the main storage unit 45b.

  When the IVR flow scenario can be specified by the command number, the control unit 44b of the media server 4b reads the command from the command execution history in the record of the IVR processing information stored in the main storage unit 45b in step S715 and externally The storage device 6 is notified, and reading of the IVR flow scenario is instructed. The external storage device 6 reads the IVR flow including the IVR process corresponding to the command from the IVR flow data in the IVR flow storage unit 62, and outputs it to the media server 4b.

  The voice processing unit 47b of the media server 4b reads the command from the command execution history in the record of the IVR processing information saved in the main storage unit 45b in step S715. The IVR processing unit 47b resumes the IVR process from the middle by executing the IVR scenario stored in the main storage unit 45b from the IVR process specified by the read command (step S720).

The voice processing unit 47b resumes the RTP session with the telephone terminal 1 using the value set in the record of the IVR processing information specified by the control unit 44b in step S715 (step S725).
Specifically, the control unit 44b reads out the outgoing / incoming start time and the RTP sequence number of the media server 4 from the record of the voice processing recovery information read out in step S715. When calculating the elapsed time from the outgoing / incoming start time to the current time, the control unit 44b divides the elapsed time by the RTP packet transmission interval to calculate the increment of the RTP sequence number. Since the RTP receiving side verifies whether the RTP sequence number is incremented by one, a predetermined constant may be added to the calculated increment, and the value may be within a range that does not cause an error. When the control unit 44b calculates the RTP sequence number to be set in the first RTP packet to be transmitted from the voice processing unit 47b by adding the calculated increment to the RTP sequence number read from the voice processing recovery information record, Notify the audio processing unit 47b. Also, the control unit 44b instructs the communication unit 40b to open the RTP port and IP address of the media server indicated by the read voice processing recovery information record.

  The voice processing unit 47b transmits the voice by the IVR process to the telephone terminal 1 by the RTP packet and receives the RTP packet transmitted from the telephone terminal 1. For the RTP packet transmitted to the telephone terminal 1, the SSRC of the media server 4, the RTP port number and the IP address of the telephone terminal 1 set in the record of the IVR processing information read in step S715 are used. The RTP sequence number notified from the control unit 44b is used for the RTP packet initially transmitted from the media server 4b. Also, the voice processing unit 47b receives from the telephone terminal 1 an RTP packet using the RTP port number and IP address of the media server 4 set in the record of the IVR processing information read in step S715.

  Subsequently, the recovery processing unit 48b determines whether there is a record that has not yet been read in the call processing recovery information stored in the call processing recovery information storage unit 63 of the external storage device 6 (step S730). If there is a record that has not been read (step S730: YES), the processing from step S710 is performed. On the other hand, if all the records have been read (step S730: NO), the media server 4b subsequently performs the processing from step S520 performed by the media server 4a. That is, it is only necessary to replace “a” described at the end of the reference numerals of each part with “b”.

  Note that the memory information sharing method of the present embodiment basically holds all data in the external storage device 6. Therefore, when the SIP signal is received from the telephone terminal 1, the call processing unit 43 of the media server 4 calls the call in the external storage device 6 specified by the to tag, the from tag, and the call tag acquired from the received SIP signal. The information is referred to and written in the record of the process restoration information and the voice process restoration information, and the call process is executed. However, in order to speed up the processing, the same information as that of the external storage device 6 is held in the main storage unit 45 of the media server 4, and when only data is read, the information is acquired from the main storage unit 45 and is stored in the data. The information may be updated by accessing the external storage device 6 only when there is an update.

[Second Embodiment: Memory Information Reconstruction Method]
Next, an IVR system according to a second embodiment of the present invention will be described.
In this embodiment, failover by the memory information reconstruction method is realized. In the memory information reconstruction method, the media server 4 holds information for executing SIP processing in its own memory, and the received packet itself such as an INVITE packet or a 200 OK packet is stored in the external storage device 6. Remember. When the active media server 4 fails, the standby media server 4 reads an INVITE packet or an ACK packet from the external storage device 6, transmits the packet to itself, and holds the active system before the failure. The contents of the previously stored memory are reconstructed in the standby memory, and the status recovery process is performed. Note that the standby media server 4 needs to implement special logic such as not returning a provisional response to the state return packet transmitted by itself. Note that storage and restoration of RTP-related information and IVR-related information are the same as in the memory information sharing method.

As described above, in the memory information reconstruction method, when the active media server 4 fails, the standby media server 4 holds the packet received by the active system by transmitting the packet received to the standby media server 4. This is a method for recovering memory information in the same manner as the media server 4 of the active system, and since the repair range is smaller than that of the memory information sharing method, the implementation is relatively easy.
Below, the difference with 1st Embodiment is demonstrated.

The configuration of the IVR system according to the present embodiment is the same as the configuration of the IVR system according to the first embodiment shown in FIG.
The IVR flow data, telephone number setting information, and IVR processing information stored in the external storage device 6 according to this embodiment are the same as those in the first embodiment.

  FIG. 14 is a diagram showing a data configuration example of call processing recovery information stored in the call processing recovery information storage unit 63 according to the present embodiment. As shown in the figure, the call processing recovery information includes to tag, from tag, callid, INVITE packet, ACK packet, 200 OK packet, media server CSeq, telephone terminal CSeq, outgoing / incoming flag, server side telephone number, terminal Each record is generated in correspondence with a call. An INVITE packet and an ACK packet are stored when the call arrives at the media server 4, and a 200 OK packet is stored when the call is made from the media server 4.

  FIG. 15 is a diagram illustrating a data configuration example of the audio processing recovery information stored in the audio processing recovery information storage unit 64 according to the present embodiment. As shown in the figure, the voice processing restoration information is composed of a record in which a to tag, a from tag, a call, a call start / end time, an RTP sequence number of the media server 4, an SSRC, an RTP port number, and an IP address are associated with each other. Each record is generated for each call. The RTP port number and IP address of the telephone terminal are described in the SIP message recorded in the call processing recovery information.

FIG. 16 is a diagram showing a sequence when a failure occurs in the IVR system according to the present embodiment.
In the figure, the processing from step S805 to S840 is the same as the processing from step S205 to S240 in FIG.
When the media server 4b detects a failure of the media server 4a, it reads the call processing recovery information from the external storage device 6 (step S845). The media server 4b internally transmits the same SIP packet received from the telephone terminal 1 to the own media server 4b using the read call processing recovery information. However, the response to the internally received SIP packet is not transmitted to the telephone terminal 1. Thereby, the media server 4b constructs information necessary for SIP processing and stores it in the main storage unit 45b (step S850).
Thereafter, the processing from step S855 to S880 is the same as the processing from step S245 to S270 in FIG.
In step S885, when a SIP packet is received from the telephone terminal 1, the media server 4b refers to the information that has been established in step S850 and stored therein, and confirms whether it is a SIP packet of an already established session ( Step S885).

  Next, a processing flow in the media server 4 according to the present embodiment will be described. The monitoring flow according to this embodiment is the same as that of the first embodiment shown in FIG.

FIG. 17 is a diagram showing a sequence example when an incoming call is received from the telephone terminal 1 to the active media server 4 according to the present embodiment.
In the figure, the processing from step S905 to S925 is the same as the processing from S405 to S425 shown in FIG.
In step S930, the control unit 44a that has received an instruction from the call processing unit 43a of the media server 4a stores the to tag, the from tag, the callid, INVITE packet, the ACK packet, the CSeq of the media server stored in the main storage unit 45a, The CSeq of the telephone terminal and the outgoing / incoming flag indicating an incoming call are output to the external storage device 6 as registered call processing recovery information, and an instruction for registration in the call processing recovery information is given (step S930). The external storage device 6 adds a record to the call processing recovery information in the call processing recovery information storage unit 63, and sets the registered call processing recovery information received from the media server 4a in the added record.

Furthermore, the control unit 44a that receives the instruction from the call processing unit 43a of the media server 4a registers the to tag, the from tag, the callid, the RTP port number and the IP address used by the media server 4a, which are stored in the main storage unit 45a. The voice processing restoration information is output to the external storage device 6 and the registration to the voice processing restoration information is instructed (step S935). The external storage device 6 adds a record to the audio processing recovery information in the audio processing recovery information storage unit 64, and sets the registered audio processing recovery information received from the media server 4a in the added record.
The call processing unit 43a of the media server 4a notifies the establishment of the dialog to the control unit 44a. Thereafter, the processing from step S440 in FIG. 10 is performed.

FIG. 18 is a diagram showing a processing flow at the time of outgoing call in the active media server 4 according to the present embodiment.
In the figure, the processing from step S1005 to S1020 is the same as the processing from S605 to S620 shown in FIG.
In step S1025, the control unit 44a that has received an instruction from the call processing unit 43a of the media server 4a stores the to tag, from tag, callid, 200OK packet, CSeq of the media server, and telephone terminal stored in the main storage unit 45a. CSeq, the server side telephone number, the terminal side telephone number, and the outgoing / incoming flag indicating transmission are output to the external storage device 6 as registered call processing recovery information, and an instruction to register the call processing recovery information is given (step S1025). The external storage device 6 adds a record to the call processing recovery information in the call processing recovery information storage unit 63, and sets the registered call processing recovery information received from the media server 4a in the added record.

Furthermore, the control unit 44a that receives the instruction from the call processing unit 43a of the media server 4a registers the to tag, the from tag, the callid, the RTP port number and the IP address used by the media server 4a, which are stored in the main storage unit 45a. It outputs to the external storage device 6 as voice processing restoration information, and instructs registration to the voice processing restoration information (step S1030). The external storage device 6 adds a record to the audio processing recovery information in the audio processing recovery information storage unit 64, and sets the registered audio processing recovery information received from the media server 4a in the added record.
The call processing unit 43a of the media server 4a notifies the establishment of the dialog to the control unit 44a. Thereafter, the processing from step S440 in FIG. 10 is performed.

FIG. 19 is a diagram showing a recovery process flow executed by the standby media server 4 when a failure of the active media server 4 according to the present embodiment is detected.
The processing of the media server 4b in steps S1105 to S1115 when a failure of the media server 4a occurs during the IVR processing is the same as the processing from steps S705 to S715 of FIG. However, in step S1110, the recovery processing unit 48b receives from the identified external storage device 6 the to tag, the from tag, and the call read from the call processing recovery information record, and the outgoing / incoming flag.

  When the outgoing / incoming flag read in step S1110 indicates an incoming call (step S1120: incoming call), the control unit 44b that has received an instruction from the recovery processing unit 48b reads the to tag, the from tag, and the call read in step S1110. The output to the external storage device 6 is instructed to read the INIVITE packet and the ACK packet (step S1125). The external storage device 6 identifies the call processing recovery information record in the call processing recovery information storage unit 63 by the to tag, from tag, and call, reads the INVITE packet and the ACK packet from the specified record, and outputs them to the media server 4b. To do.

  The recovery processing unit 48b internally transmits the INVITE packet read from the external storage device 6 to the call processing unit 43b. As a result, the call processing unit 43b performs the same processing as that when receiving the INVITE packet (step S1130). The call processing unit 43b writes the INVITE packet to the main storage unit 45b, as in the process in step S420 in FIG. 10, and also sets the from tag, callid, various header fields set in the INVITE packet, and the RTP used by the telephone terminal 1. The port number, IP address, and CSeq of the telephone terminal are read and written in the main storage unit 45b.

  Subsequently, the call processing unit 43b generates a 200 OK packet for the INVITE packet internally transmitted in step S1130, and performs the same processing as when the 200 OK packet is transmitted (step S1135). As the RTP port number and IP address of the 200 OK media server 4, the RTP port number and IP address in the record of the audio processing recovery information read in step S1115 are set. The to tag read in step S1110 is used for the To field of the 200 OK packet. Actually, the 200 OK packet is not transmitted. As a result, similar to the processing in step S420 of FIG. 10, the call processing unit 43b writes the generated 200OK packet to the main storage unit 45b, and also sets the to tag set to 200OK and the RTP port number used by the media server 4b. The IP address, the media server CSeq, and various header fields are written in the main memory 45b.

Subsequently, the recovery processing unit 48b internally transmits the ACK packet read from the external storage device 6 in step S1125 to the call processing unit 43b. Thereby, the call processing unit 43b performs the same process as the process at the time of receiving the ACK packet (step S1140). The call processing unit 43b writes the transmitted ACK packet to the main storage unit 45b and also writes various header fields to the main storage unit 45b, similarly to the processing in step S420 of FIG.
With the above processing, the media server 4b is in the same state as step S920 in FIG.

  On the other hand, if the outgoing / incoming flag read in step S1110 indicates outgoing (step S1120: outgoing), the to tag, from tag, and callid read in step S1110 are output to the external storage device 6, and the call processing recovery information is displayed. Reading is instructed (step S1145). The external storage device 6 identifies the record of the call processing recovery information in the call processing recovery information storage unit 63 by using the to tag, the from tag, and the call, reads the specified record, and outputs it to the media server 4b.

  Upon receiving an instruction from the recovery processing unit 48b, the call processing unit 43b generates an INVITE packet from the call processing recovery information record read from the external storage device 6. Thereby, call processing unit 43b performs the same process as the process at the time of INVITE packet transmission, but does not actually transmit (step S1150). The terminal-side telephone number in the read call processing recovery information record is read in the To field of the INVITE signal, the server-side telephone number and the from tag in the call processing recovery information record are read in the From field, and the call is read in the call. The call in the call processing recovery information record is 1 for the media server CSeq, the RTP port and address of the media server 4 are the RTP port number and IP read from the voice processing recovery information record read in step S1115. An address is set. The call processing unit 43b writes the INVITE packet to the main storage unit 45b, as well as the from tag, call, various header fields set in the INVITE packet, in the same manner as the process in step S615 of FIG. The RTP port number and IP address used by the media server 4b and the CSeq of the media server are read and written in the main storage unit 45b.

  Subsequently, the recovery processing unit 48b internally transmits 200 OK set in the call processing recovery information record read in step S1150 to the call processing unit 43b. Thereby, the call processing unit 43b performs the same process as the process at the time of receiving the 200 OK packet (step S1155). The call processing unit 43b writes the received 200 OK packet in the main storage unit 45b, as in the process in step S615 of FIG. 12, and also uses the to tag, the RTP port number and IP address used by the telephone terminal 1, the CSeq of the telephone terminal, Various header fields are written into the main memory 45b.

Subsequently, the recovery processing unit 48b generates an ACK packet and performs the same process as when the ACK packet is transmitted (step S1160). In the ACK packet, the same To field, From field, callid, and CSeq of the media server as in the INVITE signal are set. However, the ACK packet is not actually transmitted. As a result, similar to the processing in step S615 of FIG. 12, the call processing unit 43b writes the ACK packet to the main storage unit 45b and stores the from tag set to ACK and the CSeq of the media server in the main storage unit 45b. Write.
As a result, the media server 4b is in the same state as step S1015 in FIG.

  The voice processing unit 47b of the media server 4b resumes the IVR process from the middle using the record of the IVR processing information stored in the main storage unit 45b by the same process as step S720 of FIG. 13 (step S1165), and the voice process The unit 47b resumes the RTP session with the telephone terminal 1 by using the value set in the record of the IVR processing information specified by the control unit 44b in step S1115 by the same process as step S725 in FIG. (Step S1170).

  Subsequently, the restoration processing unit 48b determines whether there is a record that has not yet been read in the voice processing restoration information stored in the voice processing restoration information storage unit 64 of the external storage device 6 (step S1175). If there is a record that has not been read (step S1175: YES), the processing from step S1110 is performed. On the other hand, if all the records have been read (step S1175: NO), the media server 4b subsequently performs the processing from step S520 performed by the media server 4a. That is, “a” described at the end of the reference numerals of the respective units may be replaced with “b”.

  As described above, according to the present embodiment, the active media server writes information necessary for SIP and RTP recovery to the external storage device when the session is established, and when the IVR process is started, The update status of the IVR process is written to the storage device in real time, and no information regarding RTP is written. Accordingly, without overloading the access to the external storage device, the media server can write information that needs to be taken over to the external storage device when a failure occurs, and does not cause a delay in the voice packet. Note that SIP update information may be written to an external storage device even after a session is established. Even if the active media server fails during a call, the standby media server can continue to provide an IVR scenario from the point of time when the failure occurs without interrupting the call. As described above, since the session failover of the media server can be realized, a highly reliable system can be realized.

  The telephone terminal 1, the relay server 2, the media server 4, the external storage device 6, and the transmission control device 7 described above have a computer system inside. The operation processes of the telephone terminal 1, the relay server 2, the monitoring unit 41 and the media processing program execution unit 42 of the media server 4, the external storage device 6, and the transmission control device 7 are computer-readable in the form of a program. The above processing is performed by the computer system reading and executing this program stored in a recording medium. The computer system here includes a CPU, various memories, an OS, and hardware such as peripheral devices.

Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

  The present invention is not limited to the embodiment described above, and various modifications can be made without departing from the spirit of the present invention, and it goes without saying that these are also included in the scope of the present invention. Yes.

DESCRIPTION OF SYMBOLS 1 ... Telephone terminal 2 ... Relay server 3 ... Router 4, 4a, 4b ... Media server 40, 40a, 40b ... Communication part 41, 41a, 41b ... Monitoring part 42, 42a, 42b ... Media processing program execution part 43, 43a, 43b: Call processing unit 44, 44a, 44b ... Control unit 45, 45a, 45b ... Main storage unit 46, 46a, 46b ... IVR processing unit 47, 47a, 47b ... Voice processing unit 48, 48a, 48b ... Recovery processing unit 6 ... External storage device 60 ... Storage unit 61 ... Telephone number setting storage unit 62 ... IVR flow storage unit 63 ... Call processing recovery information storage unit 64 ... Voice processing recovery information storage unit 65 ... IVR processing information storage unit 7 ... Transmission control device

Claims (7)

The media server in an automatic voice response system having a plurality of media servers connected via a network,
A call processing unit that transmits and receives a call control signal packet to and from a telephone terminal connected via the network, and establishes a call session with the telephone terminal;
An automatic voice response processing unit that performs transmission of voice to the telephone terminal that establishes a call and / or recognition of voice received from the telephone terminal according to a scenario of automatic voice response;
A media session is established with the telephone terminal that established the call, and a voice packet in which voice transmitted from the voice automatic response process is instructed by the established media session is transmitted to the telephone terminal; A voice processing unit that outputs a voice packet received from the telephone terminal to the voice automatic response processing unit;
Writing call processing recovery information in which call specific information acquired from the call control signal transmitted and received by the call processing unit and information used for call session recovery are set in an external storage device, and the call specific information, Voice processing restoration information in which information used for restoration of the media session established for the call identified by the call identification information is set is written in the external storage device, and the call identification information and the call identification information depending on the current information for specifying the execution position, the scenario of voice response executed by the voice response unit on scenario voice response executed by the voice response processing unit for a call identified by the voice response processing information set and history information of the result of recognizing the voice received from the telephone terminal to the external storage device Te And a control unit to burn them,
A monitoring unit for detecting a failure of another media server;
When the monitoring unit detects a failure of another active media server, the address used by the other media server is set as its own address, and the same call written by the other media server is written. Call processing recovery information having specific information, voice processing recovery information, and a recovery processing unit that reads voice automatic response processing information from the external storage device,
The voice automatic response processing unit acquires information specifying the current execution position on the scenario and the history information from the voice automatic response processing information read by the recovery processing unit , and the acquired information and the history information run the scenario of the voice response from the execution position specified by a,
The audio processing unit establishes a media session based on the audio processing recovery information read by the recovery processing unit, and encodes audio instructed to be transmitted from the automatic audio response processing unit by the established media session. Sending voice packets to the telephone terminal, and outputting voice packets received from the telephone terminal to the voice automatic response processing unit,
The call processing unit establishes a call session with the telephone terminal based on the call processing recovery information read by the recovery processing unit, and transmits and receives call processing signals.
A media server characterized by that.   The information used for call restoration includes the latest telephone number, the destination telephone number acquired from the transmitted / received call control signal, and the latest information whose value is updated when the call control signal is transmitted or received. The media server of claim 1, comprising a value. The information used for call recovery includes the latest value of the call control signal received from the telephone terminal, and the information whose value is updated when the call control signal is transmitted or received,
The restoration processing unit transmits the call control signal set in the call processing restoration information read from the external storage device to the call processing unit,
The call processing unit establishes a call session with the telephone terminal based on the call control signal transmitted from the recovery processing unit;
The media server according to claim 1. The information used for media session recovery includes addresses for media sessions on the calling and called sides, information in which values specific to the media session are set, and information whose values are updated when a voice packet is transmitted. Including a value at the time of media session establishment and a start time of the media session,
The voice processing unit acquires the start time from the voice processing restoration information read by the restoration processing unit, and a value at the time of media session establishment of the information updated with transmission of the voice packet, Calculate the current value of the information to be updated based on the elapsed time from the start time, read from the voice processing recovery information, the address for the caller and callee media session, and the media session specific Recovering the media session of the call using the information with the value set and the information set with the calculated value and updated with the transmission of the voice packet;
The media server according to any one of claims 1 to 3, characterized in that: The control unit deletes the call processing recovery information including the call identification information of the released call, the voice processing recovery information, and the voice automatic response processing information from the external storage device;
The media server according to any one of claims 1 to 4, wherein the media server is characterized. A session recovery method used for the media server in an automatic voice response system having a plurality of media servers connected via a network,
A call processing step for sending and receiving a call control signal packet to and from a telephone terminal connected via a network, and establishing a call session with the telephone terminal;
A voice automatic response processing unit that performs voice transmission to the telephone terminal that established the call and / or recognition of the voice received from the telephone terminal according to a voice automatic response scenario;
A voice processing unit establishes a media session with the telephone terminal that has established a call, and transmits a voice packet obtained by encoding the voice transmitted in the voice automatic response processing step by the established media session to the telephone terminal. A voice processing step of transmitting and outputting a voice packet received from the telephone terminal to the voice automatic response processing unit;
The control unit writes the call processing recovery information in which the call identification information acquired from the call control signal transmitted and received in the call processing step and the information used for recovery of the call session are set in an external storage device, and the call Write audio processing recovery information that sets specific information and information used for recovery of the media session established for the call specified by the call specific information to the external storage device, and further, the call specific information, Information specifying the current execution position on the scenario of the voice automatic response executed in the voice automatic response processing step for the call specified by the call specifying information, and the voice automatic response executed in the voice automatic response processing step voice self set and history information of the result of recognizing the voice received from the telephone terminal in accordance with the scenario An information writing step of writing the response processing information to said external storage device,
A monitoring step in which a monitoring unit detects a failure of another media server; and
When the recovery processing unit detects a failure of another active media server in the monitoring step, the address used by the other media server is set as its own address, and the other media server writes However, a recovery processing step of reading out call processing recovery information, voice processing recovery information, and voice automatic response processing information having the same call identification information from the external storage device,
The voice automatic response processing unit acquires information specifying the current execution position on a scenario and the history information from the voice automatic response processing information read in the recovery processing step, and the acquired information and the A voice automatic response recovery step of executing the voice automatic response scenario from the execution position specified by the history information ;
The audio processing recovery establishes a media session based on the audio processing recovery information read in the recovery processing step, and encodes the audio that is instructed to be transmitted in the audio automatic response recovery step by the established media session. A voice processing recovery step for transmitting the voice packet to the telephone terminal and receiving the voice packet received from the telephone terminal;
The call processing unit establishes a call session with the telephone terminal based on the call processing recovery information read in the recovery processing step, and performs call processing recovery step for transmitting and receiving call processing signals;
A session recovery method comprising: A computer used as the media server in an automatic voice response system having a plurality of media servers connected via a network,
A call processing step of transmitting / receiving a call control signal packet to / from a telephone terminal connected via a network and establishing a call session with the telephone terminal;
An automatic voice response processing step for performing transmission of voice to the telephone terminal that has established a call and / or recognition of voice received from the telephone terminal according to a scenario of automatic voice response;
A media session is established with the telephone terminal that established the call, and a voice packet encoded with the voice transmitted in the voice automatic response processing step by the established media session is transmitted to the telephone terminal; A voice processing step for receiving voice packets from the telephone terminal;
Writing call processing recovery information in which call specific information acquired from the call control signal transmitted and received in the call processing step and information used for recovery of a call session is set in an external storage device, and the call specific information, Voice processing restoration information in which information used for restoration of the media session established for the call identified by the call identification information is set is written in the external storage device, and the call identification information and the call identification information depending on the information specifying the current execution location on scenario voice response executing in said voice response processing steps for a call identified, scenarios voice response to said executed in interactive voice response processing step by voice response processing in which the history information of the result of recognizing the voice received from said telephone terminal Te An information writing step of writing the multicast to the external storage device,
A monitoring step for detecting a failure of another media server;
When a failure of another active media server is detected in the monitoring step, the address used by the other media server is set as its own address, and the same call written by the other media server is written. Call processing restoration information having specific information, voice processing restoration information, and restoration processing step of reading out voice automatic response processing information from the external storage device;
The information specifying the current execution position on the scenario and the history information are acquired from the voice automatic response processing information read in the restoration processing step, and the execution specified by the acquired information and the history information An automatic voice response recovery step for executing the automatic voice response scenario from a position;
A media session is established based on the voice processing restoration information read in the restoration processing step, and a voice packet obtained by encoding the voice instructed to be sent in the voice automatic response restoration step by the established media session is A voice processing recovery step of transmitting to the telephone terminal and receiving a voice packet received from the telephone terminal;
A call processing recovery step for establishing a call session with the telephone terminal based on the call processing recovery information read in the recovery processing step, and transmitting and receiving a call processing signal;
A computer program for executing

Images