GB2354396A

GB2354396A - Accessing voice mail using Java and WWW

Info

Publication number: GB2354396A
Application number: GB0013177A
Authority: GB
Inventors: Zohar Sivan; Gilad Cohen; Hagai Krupnik; Zvi Mizrahy
Original assignee: International Business Machines Corp
Current assignee: International Business Machines Corp
Priority date: 1999-06-02
Filing date: 2000-06-01
Publication date: 2001-03-21
Also published as: GB0013177D0; IL135598A0

Abstract

A Java implemented system connects a telephony Voice-response system, such as a Voice-mail box or an IVR, and a terminal over the Internet. An IP-PSTN gateway is used to bridge the telephony world and the IP network. A Java-based client 17 communicates with a standard HTTP web server 16 for sending dialing information and other instructions to the Voice-response system. The web server communicates with the IP-PSTN gateway 13 using a servlet 14 or cgi-bin script running on the HTTP web server. Instructions pass to the Voice-response system and audio data is sent in real-time from the Voice-response system to the gateway, where it is compressed and forwarded to the client which decompresses the audio data and plays the messages. The user operates the Voice-response system in conventional manner, so that standard Voice-response functionality is maintained. The client is written in Java as an applet, so any Java-enabled web browser can run the client and connect to the HTTP web server. Dialing information and instructions which are sent from the applet to the Voice-response system, can be easily configured in the client.

Description

2354396 METHOD AND APPARATUS FOR ACCESSING A TELEPHONY VOICE-RESPONSE

SYSTEM FROM A TERMINAL USING A STANDARD WEB BROWSER

FIELD OF THE INVENTION

This invention relates to a voice-response system operable through the Internet.

BACKGROUND OF THE INVENTION

Voice-response systems are widely used in the telephony world. one particular example is a Voice-mail system that enables users to communicate even if the called party is not available, by recording a message in real-time which is saved in the called party voice-mail box on the voice-mail system. Telephony Voice-mail systems are designed to work within the telephony world, i.e. interaction is done using a telephone line and instructions to the Voice-mail system are sent as Dual Tone Multiple Frequency (DTMF) tones. A user can access the voice-mail system by dialing an access number from a telephone and interaction with the system is done using the telephone keypad, which generates DTMF tones.

A drawback of conventional Voice-mail systems is that when users need to access their Voice-mail system remotely e.g. from abroad, a long distance call or even International telephone call is needed. This call is much more expensive compared to local telephone calls.

Several alternatives to the traditional approach for accessing the Voice-mail system without a telephone exist. one method is to send the recorded messages to an E-mail address of the user, who can listen to them independently with an audio player on the terminal. The drawbacks of this method are the lack of interaction with the Voice-mail system and the fact that all data must be sent before a user can listen to the messages. If the audio messages are uncompressed, a standard player can be used by the client, but the download time per a given bandwidth is much longer. If compression is used, a corresponding decoder must be installed in the client's machine. An E-mail address is also required.

Another approach is to connect to a Voice-mail system from the web over the Internet, download the messages as audio files to the client and only then play them on the client terminal. IBM's DirectTalk/6000 product features this method. (DirectTalk/6000 is a trademark of International Business Machines Corp.) Again, the lack of real-time streaming and 2 interaction with the Voice-mail system (for skipping or deleting messages for example) constitute drawbacks to this approach.

with the advances in the Internet world and specifically in Voice-over-IP (VoIP), bridging the two worlds can offer substantial benefits to users as well as for the telephone companies. Users can use their Internet connection for utilizing telephony functions, and the telephone companies can offer enhanced services. The purpose of this disclosure is to describe a novel approach for accessing a telephony

Voice-mail system from a terminal utilizing web-based tools and access methods. The user is capable of getting the messages in real-time streamed over the Internet and can interact with the voice-mail system as if he is using a standard telephone.

A major component in an IP telephony network is the PSTN gateway, which bridges between the IP network and the legacy PSTN telephony network.

It enables telephone calls and fax transmission between Internet workstation and regular telephones or/and fax machines. Generally, the PSTN gateway bridges calls between PSTN telephones and voice over IP terminals, such as Microsoft's NetMeeting Internet telephone. (NetMeeting is a trademark of Microsoft Corp.) The PSTN gateway also enables two legacy telephones or fax machines to communicate over an IP intermediate. The gateway uses signaling protocol such as the ITU-T H.323 protocol suite for the call setup and capability negotiation, while the voice streaming is usually done with common encoding/decoding schemes such as ITU-T G711, G723.1 or ETSI-GSM.

EP 889 627 in the name of AT&T Corporation discloses a remote access Voice-mail method and system for giving a client Voice-mail access over the Internet using a web graphical interface. A web server is coupled to an audio streaming server that is responsive to an audio signal received from a voice response unit for digitizing the audio signal and transmitting the digitized audio signal to the client via the web server. Whilst this reference addresses many of the drawbacks associated with PSTN-based systems, it is believed that the provision of the audio streaming server represents an unnecessary overhead which unduly complicates the manner in which the voice data is processed.

it would therefore be desirable to provide an IP-based Voice-response system offering high flexibility and connectivity to existing Voice-response systems, whilst requiring minimal changes.

3 it is an object of the present invention to provide a technique which alleviates the above drawbacks.

SUMMARY OF THE INVENTION

According to the present invention we provide a method for accessing a voice-response system coupled to a telephone network and accessible via a web server coupled to a computer network, said method comprising the following steps carried out by the web server:

1)receiving a voice-response request from a client, via the computer network, 2)transmitting via the computer network a Voice-response Java applet to the client, 3)receiving Voice-response access data from the client, 4)processing the Voice-response access data and sending corresponding DTMF commands to an IP-PSTN Gateway bridging the web server and the telephone network for onward delivery to the voice-response system, and 5)streaming voice data via the IP-PSTN Gateway to the client.

Also according to the present invention we provide a web server connectable to a computer network for accessing a Voice-response system coupled to a telephone network via an IP-PSTN Gateway, the web server comprising:

a memory storing therein a Voice-response Java applet and a web Java servlet or cgi-bin script, a request module coupled to the memory for receiving a Voice-response request from a client, via the computer network, and storing in the memory, a transmit module coupled to the memory for transmitting via the computer network a Voice-response Java applet to the client, an access module coupled to the memory for receiving voice-response access data from the client and storing in the memory, a processor coupled to the memory for processing the Voice-response access data and sending corresponding DTMF commands to the IP-PSTN Gateway for onward delivery to the Voice-response system, and 4 a voice streaming module coupled to the processor for receiving voice data from the Voice-response system via the IP-PSTN Gateway and forwarding to the client.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to understand the invention and to see how it may be carried out in practice, a preferred embodiment will now be described with regard to a Voice-mail system, by way of non-limiting example only, with reference to the accompanying drawings, in which:

Fig. 1A is a block diagram showing functionally the principal components in a system operating over the Internet and comprising a client, web server and IP-PSTN gateway triplet; is Fig. 1B shows functionally a detail of a servlet run by the web server; Fig. 2 is a flow diagram showing the principal operating steps carried out by the client; Fig. 3 shows pictorially an HTTP form used by an applet downloaded to the client; and Figs. 4A to 4D are a flow diagram showing the principal operating steps executed during the interaction of the client, web server and IP-PSTN gateway triplet.

DETAILED DESCRIPTION OF A PREFERRED E24BODIMENT

Fig. 1A showing functionally a system 10 comprising a Voice-mail system 11 connected to a Public Switched Telephone Network (PSTN) 12 coupled via an IP-PSTN gateway 13 to a servlet 14 stored within a memory 15 in an HTTP web server 16. A client computer 17 is connected to the web server 16 via a computer network 18 which allows a Java applet 19 also stored in the memory 15 to be downloaded f rom the web server 16 to a Java-based web browser 20 within the client computer 17. The Java applet 19 allows for decompressing and playing of voice messages, and for sending dialing information and instructions to the web server 16. The servlet 14 allows connection of the web server 16 to the gateway 13, thereby allowing access via the gateway 13 to the Voice-mail system 11 through the PSTN 12.

Messages in the Voice-mail system 11 are relayed to the client 17 for playing on an audio playback system 21.

Fig. 1B shows functionally a detail of the servlet 14, which comprises a request module 25 coupled to the memory 15 for receiving a Voice-mail request from a client 17, via the computer network is, and storing in the memory 15. A transmit module 26 is coupled to the memory 15 for extracting therefrom the Voice-mail Java applet 19 and transmitting via the computer network 18 to the client 17. An access module 27 coupled to the memory 15 receives Voice-mail access data from the client 17 and stores in the memory 15. A processor 28 is coupled to the memory 15 for processing the Voice-mail access data and sending corresponding DTMF commands to the IP-PSTN Gateway 13 for onward delivery to the Voice-mail system 11. A voice streaming module 29 is coupled to the processor 28 for receiving voice data from the Voice-mail system 11 via the IP-PSTN Gateway 13 and forwarding to the client 17.

Fig. 2 shows the principal operating steps carried out by the client 17 in order to access the Voice-mail system and Fig. 3 shows pictorially a Graphical User Interface (GUI) 30 used by the Java applet 19 run on the client. The client first connects to a known HTTP address that downloads a Web page containing the Voice-mail client Java applet. The applet is optimized in terms of minimal class size, such-that downloading time is only a few seconds. The Graphical User Interface 30 allows the user to enter details such as the voice-mail access number, user address number and user password. The user submits this information in order to initialize the call. If the call initialization is successful, the applet plays the audio prompt through the audio playback module 21, as can be heard in a regular telephone connection to the voice-mail.

The user can then submit additional telephone keypad DTMF signals using the applet's GUI. Different command buttons on the applet's GUI cause appropriate digital command strings to be sent to the servlet 14. The servlet 14 feeds them to the IP-PSTN gateway 13, which converts them to corresponding DTMF signals. These DTMF signals are sent to the voice-mail system in order to "navigate" through the mail system menus, thus providing the facility to play the current message, move on to the next message, delete the current message, etc. Each user interaction is followed by a new audio stream that is played in response to the user selection. The user can press a "hang-up" button to terminate the call at any given time. It will be understood that the applet is not restricted to the features described above and other features can be provided as required. Specifically, use of applets allows customization to a specific voice-mail system and further allows a single applet to access different Voice-mail systems by clicking 6 on different command buttons, which are pre-programmed to initiate different set of DTMF commands.

As further seen in Fig. 3, the Graphical User Interface 30 provides options for connection to voice-mail systems within different telephony networks. For example, the Voice-mail system may be part of the PSTN or part of the cellular telephone network. Alternatively, the web server may be part of an Intranet for use within an office, for example, in which case the Voice-mail system is maintained for different internal extensions thereof.

For each new dial request, the applet closes the current open HTTP connection with the Web server (if any), and opens a new one. This way "future" audio chunks are dropped and part of the inherent delay is not noticeable by the client. Embedded in the new HTTP connection header fields, is the connection type: call initialization, additional DTMFs or hang up. Also embedded, is a character string containing the DTMP signals that should be dialed. These are usually the telephone key pad digits 0-9 and the characters 11#11 and 1111. In some countries, it may also comprise alphabetic characters translated by the Gateway to the proper DTMF tones.

When a successful HTTP connection is established with the Web server, compressed audio is received by the client 17 as returned data in the body of the HTTP stream. Known standards for audio compression algorithms are the ITU-T G.723.1 or the ETSI-GSM family of speech encoders. A compression technique is chosen such that the compression rate allows transmission in real-time through common end-user modems, typically 28.8 Kbit per second.

The audio decoding and playback is done in real-time. To overcome IP network problems such as congestion stalls and variable arrival rates of Internet packets, a technique for audio prefetching and buffering is implemented in order to ensure high quality, continuous audio playback.

Since the connection with the Web server is performed using standard HTTP protocol, the Voice-mail Client applet 19 can connect with the servlet 14 even if either of them operates behind a firewall.

The web server The web server 16 can be any standard HTTP web server which supports cgi-bin scripts or servlets (all current HTTP web servers support at least one of these methods). Thus although servlets are used in the preferred 7 embodiment, it is to be understood that cgi-bin scripts can be equally well employed. The web server 16 processes HTTP requests from the client 17 to the Voice-mail system 11 and streams the voice data from the Voice-mail system 11 to the client 17 through the servlet. The web server 16 is assumed to be connected to the Gateway 13 through a communication network with very high bandwidth (like a LAN or WAN). Alternatively, the web server 16 and the IP-PSTN gateway 13 can be integrated within a single unit.

According to a preferred embodiment, the servlet method is used for its superior performance and better maintainability. Moreover, each instance of the servlet initiated by the web server 16 can support multiple users simultaneously.

The servlet A servlet is an Auxiliary module that responds to particular HTTP requests that arrive to the HTTP web server. In the configuration described above and shown in Fig. 1, the servlet receives an HTTP request from the applet run by the client 17. This HTTP request includes all the information that is needed to initiate a call to the voice-mail system. The servlet passes the information to the gateway 13 over a TCP socket using ITU-T H.323 or a proprietary protocol.

Upon successful initiation of the call, the gateway starts sending compressed audio bitstream to the servlet. This information is passed through User Datagram Protocol (UDP) datagrams which can be set directly or encapsulated in data packets employing another protocol such as Real Time Protocol (RTP), PTSP and so on. The servlet then opens an HTTP response to the applet/client and appends the arrived bitstream: to it.

The servlet then listens to follow-up requests from the client, which are handled in a similar manner to that described above. However, the servlet should match the requests that are coming from the same client.

This could be done by means of "session management".

While all data can be sent f rom the servlet to the gateway on a single TCP socket, audio bitstream arriving from the gateway to the servlet over LMP datagrams should use distinct UDP ports per each session.

Since the IP-PSTN gateway typically closes TCP sockets if no data is received during a specified time-out, i.e. non active sockets are closed after a while, the servlet initiates "Keep Alive" messages in order to keep 8 the socket alive. it should, however, be noted that the need to initiate "Keep Alive" messages is a function of the specific IP-PSTN gateway being used. If the IP-PSTN gateway has no problem of time-out, then there is no need to initiate "Keep Alive" messages. 5 Since HTTP connection introduces noticeable delays over the net, any bitstream that arrives at the servlet prior to the receipt of a follow-on request from a client and has not yet been sent to the client (owing to the HTTP communication bandwidth) should be discarded. This can be done by closing the UDP port and reopening it.

Any of the triplet applet-servlet-gateway can initiate a hang-up message that affects all three and closes the current session.

The IP-PSTN gateway The IP-PSTN gateway 13 is the heart of the telephony systems in companies that use IP telephony intensively. PSTN gateways vary and are categorized according to the hardware platform, operating system, telephony card/s and performance parameters. Small offices might need a single low-scale gateway running on a Pentiumm with a 'MicrosoftO NT operating system that uses up to 8 analog ports. (Microsoft and NT are trademarks of Microsoft Corp. and Pentium is a registered trademark of Intel Corp.) Large offices, carriers and call centers may require several high-scale gateways running on a RISC platform under the UNIX operating system and using several T1/E1 cards with a capacity of hundreds of ports. (UNIX is a registered trademark and is licensed exclusively through X/Open Company Ltd.) The IP-PSTN gateway 13 is used to bridge between the web Server 16 and the Voice-mail system 11 and supports several simultaneous connections.

It receives commands from the web server 16 and sends them to the Voice-mail system 11 as DTMF signals. When the connection is established, it reads voice data from the Voice-mail system 11, compresses it and transfers it to the web server 16.

Connection between the components By way of example, the proprietary interface between the IP-PSTN gateway 13 and the servlet may be implemented by two types of connection:

(1) Signaling and control over a well known predefined TCP port. Four messages are sent from the servlet to the IP-PSTN gateway:

INITIALIZE KEEP ALIVE DTMF Two messages are sent from the IP-PSTN Gateway 13 to the servlet:

A detailed description of each message is given below. The IP-PSTN gateway 13 is connected to the web server 16 via a single TCP connection, and is capable of accepting more than one connection request on this port. Consequently, the IP-PSTN gateway can support several web servers thus allowing the client to select a specific web server for effecting communication. By such means, telephone costs to the client can be reduced by locating the gateway as close as possible to the Voice-mail system and by the client accessing the Voice-mail system through his closest Internet Service Provider.

(2) voice streaming over UDP. Voice data is streamed from the IP-PSTN gateway 13 to the Web Server 16. The servlet defines which UDP port will be used to receive data from the gateway and informs the gateway so that the gateway knows to which UDP to send data. It is to be noted that each connection requires a different UDP port and it is the responsibility of the servlet to allocate them.

The interface between the servlet and the client Is applet is performed via standard HTTP in both directions: commands from the client Is applet are sent to the servlet and voice is streamed f rom the servlet to the client's applet.

Having described the various components of the system, there will now be described a typical implementation. Figs. 4A to 4D show the principal operating steps executed during the interaction of the client, web server and IP-PSTN gateway triplet, whilst the actions relating to the client have already been described in part with reference to Fig. 2.

a The user runs a web browser such a Netscape Navigator or Microsof tO Internet Explorer.

N User asks for a specific URL on the Web Server. This hit downloads the client's applet to the Web browser.

5 The user fills the following details into an HTML form or applet GUI:

0 Voice-mail telephone number a User extension 2 User code 3 When the user presses a SUBMIT button, this information is collected by the applet from the HTML form and is sent to the Web server.

a The Web server runs an instance of the servlet and passes the request from the client's applet to the servlet.

a when the servlet accepts the session initialization request, it tries to establish a TCP socket connection with the dedicated process running in the PSTN gateway, if the socket has not already been opened. There are two ways in which this can be achieved. Thus, when the f irst user for a specific web server initiates a request to the web server, if there is no open connection, then one is opened now. Alternatively, the servlet may be programmed to effect the TCP socket connection when the server is first initiated, regardless of whether or not a user request has been received.

a The servlet sends an INITIALIZE command to the IP-PSTN gateway. The command includes the information contained in the Applet GUI as well as a session ID, thus allowing several Applet GUIs to be processed simultaneously, each in respect of a different client session. The HTTP web servlet allocates a new dedicated UDP port for the voice streaming and sends the port number in the INITIALIZE request to the servlet.

a When the IP-PSTN gateway receives the INITIALIZE command it registers the connection to this servlet and tries to dial the Voice-mail system telephone number.

0 The IP-PSTN gateway sends an INITIALIZE-ACK acknowledgement to the servlet.

ff When the Voice-ma-41 system answers, the PSTN gateway sends the User extension and the User Code into the voice-mail system as DTMF tones.

0 The IP-PSTN gateway waits for the establishment of a connection with the servlet over the UDP port. This port is used for the voice streaming.

0 The HTTP web servlet accepts the INITIALIZE-ACK response and starts listening to the UDP port.

Connection Duration During communication between the Gateway-client-webserver, the following actions are performed:

1. The PSTN gateway reads voice data from the telephony card, compresses the data using any suitable compression scheme (such as GSM), and streams the compressed frames to the servlet. The servlet receives these voice frames and streams them over HTTP connection to the client's applet.

2. The client's applet decodes the voice packets and plays it via the audio device.

3. During the connection, KEEP-ALIVE messages are sent from the servlet to the PSTN gateway over the TCP connection.

4. At any time during the connection, the user can send DTMF commands to the Voice-mail system. This is done by typing the DTMF sequence into a special text box on the applet GUI (Graphical User Interface).

Alternatively, a separate DTMF keypad could be used, or the GUI command buttons could be programmed to emit the appropriate DTMF sequence. After pressing the Submit button, the command is sent via HTTP to the servlet. The servlet identifies the session and sends this DTMF sequence to the PSTN gateway over its corresponding TCP connection. The PSTN gateway sends the DTMF tones to the Telephony card.

Connection Termination Connection can be terminated if one of the following occurs:

The HTTP connection between the client's applet and the servlet has been terminated or lost. This may happen if, for example, the user presses the "Hang-Up" button, leaves the current URL location to another location, or terminates the browser session. In this case the servlet-gateway voice streaming session is ended. The servlet sends a HANG-UP command to the PSTN gateway, which causes the gateway to disconnect the call, and to close the corresponding voice streaming socket to the servlet. As noted above, the TCP signaling socket remains open.

The telephone call to the Voice-mail system has been terminated (either by the PBX or the Voice-mail system itself). In this case the PSTN gateway closes the corresponding voice streaming socket to the servlet.

When the servlet traps this disconnection, it terminates the user-specific thread for the instant user.

12 A KEEP-ALIVE time-out is detected by the PSTN gateway. This causes the PSTN gateway to close all the sockets to the servlet. When the servlet traps this disconnection, it enters an idle state, after closing all the client sessions.

The invention has been described with regard to the client-server connection. However, it will be appreciated that the invention allows a new service to be provided not only by the Internet Service Provider, but also by the telephone company or any other Service Provider. For example, a Service Provider could use an Interactive Voice Response (IVR) system according to the invention to route a customer's inquiry to the correct destination.

Furthermore, as noted above, the invention is also applicable to Intranet systems, such as for internal office the client use and the like. In this case, the web server is a conventional computer to which and which is configured to run a Java servlet and download a Java computers are connectedapplet to one or more clients upon connection therewith.

In the method claims that follow, alphabetic characters used to designate claim steps are provided for convenience only and do not imply any particular order of performing the steps.

Claims

13 CLAIMS 1 A method for accessing a Voice-response system coupled to a telephone network and accessible via a web server coupled to a computer network, said method comprising the following steps carried out by the web server:

1)receiving a Voice-response request from a client, via the computer network, 2)transmitting via the computer network a Voice-response Java applet to the client, 3)receiving Voice-response access data from the client, 4)processing the Voice-response access data and sending corresponding DTMF commands to an IP-PSTN Gateway bridging the web server and the telephone network for onward delivery to the Voice-response system,and 5)streaming voice data via the IP-PSTN Gateway to the client.

2. The method according to Claim 1, wherein in step (c) the Voiceresponse data is received by the web server via the computer network.

3. The method according to Claim 1, wherein in step (c) the Voice-response access data is received as standard HTTP requests sent by the Java applet and include dialing information and instructions.

4. The method according to any preceding Claim, wherein in step (d) a servlet or a cgi-bin script is used to interconnect the web server and the IP-PSTN gateway.

5. The method according to any preceding Claim, wherein in step (e) voice messages extracted from the Voice-response system and fed to the IP-PSTN gateway are streamed in real-time by the web server to the client.

6. The method according to any preceding Claim, wherein in step (e) web server streams compressed voice messages from the IP-PSTN gateway to the client.

7. The method according to any preceding Claim, further including the step of sending KEEP-ALIVE messages to the IP-PSTN gateway.

8. The method according to any preceding Claim, further including the step of decoding the Voice-response access data so as to determine an 14 identity of the client, thereby allowing multiple client access simultaneously.

9. The method according to any preceding Claim, further including the steps of detecting a loss in communication with the client and sending a HANG-UP command to the IP-PSTN gateway.

10. The method according to any preceding Claim, further including the step of terminating in response to detecting a disconnection by the IPPSTN gateway.

11. The method according to any preceding Claim, wherein the Java applet is adapted to allow connection to different Voice-response systems.

12 A web server connectable to a computer network for accessing a Voice-response system coupled to a telephone network via an IP-PSTN Gateway, the web server comprising:

a memory storing therein a Voice-response Java applet and a web Java servlet or cgi-bin script, a request module coupled to the memory for receiving a voice-response request from a client, via the computer network, and storing in the memory, a transmit module coupled to the memory for transmitting via the computer network a Voice-response Java applet to the client, an access module coupled to the memory for receiving Voice-response access data from the client and storing in the memory, a processor coupled to the memory for processing the Voice-response access data and sending corresponding DTMF commands to the IP-PSTN Gateway for onward delivery to the Voice-response system, and a voice streaming module coupled to the processor for receiving voice data from the Voice-response system via the IP-PSTN Gateway and forwarding to the client.

13. The web server according to Claim 12, being adapted for connection to the Internet.

14 The web server according to Claim 12 or 13, being adapted for connection to an Intranet.

15. The web server according to any Claim 12-14, wherein the request module forwards command strings to the IP-PSTN Gateway for conversion thereby to DTMF signals.

16. The web server according to any Claim 12-15, wherein the access module is responsive to standard HTTP requests including dialing information and instructions-

17. The web server according to any Claim 12-16, wherein the servlet is adapted to extract voice messages from the Voice-response system via the IP-PSTN gateway and to stream said voice messages in real-time to the client from the IP-PSTN gateway. 10

18. The web server according to any Claim 12-17, wherein the servlet is adapted to receive compressed voice messages from the IP-PSTN gateway for sending to the client.

19. The web server according to any Claim 12-18, wherein the servlet is 15 adapted to send KEEP-ALIVE messages to the IP-PSTN gateway.

20. The web server according to any Claim 12-19, wherein the servlet is adapted to decode the Voice-response access data so as to determine an identity of the client, thereby allowing multiple client access 20 simultaneously.

21. The web server according to any Claim 12-20, wherein the servlet is adapted to detect a loss in communication with the client and to send a HANG-UP command to the IP-PSTN gateway.

22. The web server according to Claim 12-21, wherein an instance of the servlet is adapted to terminate in response to detecting a disconnection by the IP-PSTN gateway.

23. The web server according to any Claim 12-22, wherein the Java applet is adapted to allow connection to different Voice-response systems.

24. The computer program product for operating a web server, comprising a computer processor, to provide a remote client with access to a voice-response system, said computer program product comprising:

a computer readable medium; computer program instructions for implementing the method of Claim 1-11.