An intelligent network allows providing a subscriber of a telecom- munication network - such as a wired network or a mobile telephone network - with a host of different services. These services include e.g. a Private Num- bering Plan (PNP), enabling the use of private numbers, and a Personal Num- ber which allows the users to call a specific person using the same number, irrespective of the person's location and the subscriber connection used for receiving calls. A specific numbering block is defined for PBX subscribers in the numbering domain of the exchange. One example of this kind of an intelli- gent network is described in the recommendations of ITU-T Q-1200 series, of which Q-1210 - Q-1219 define the set of features known as CS-1 (Capability Set 1) and, correspondingly, Q-1220 - Q-1229 define the set of features CS-2.

The present invention suggests an improvement to the call set-up mechanism presented in recommendation Q-1214. This recommendation defines functions such as Service Control Function (SCF) and Service Switching Function (SSF). A network element performing the SCF is called a Service Control Point (SCP). Similarly, a network element performing the SSF is called a Service Switching Point (SSP).

A Basic Call State Model BCSM defined in connection with an intel- ligent network describes the different stages of call control and includes the points where call control can be interrupted for initiating an intelligent network service. It identifies those Event Detection Points in the call and the connec- tion process where service logic entities of the intelligent network can have an interactive relationship with basic call and connection management features.

One of the services offered by telecommunication operators is conference call. The problem with the general mechanisms for setting up a conference call is that they require considerable skill and care from the sub- scriber, because conference call set-up must be controlled using special keys # and * of the telephone set. Such problems relating to the convenience of use are often solved by means of an intelligent network architecture. The problem with CS-1 and CS-2 intelligent network architectures is that they do not include the possibility of merging call segments initiated by a Service Control Function SCF. For instance, CS-2 architecture entities CallSegmentAssociation, Call-

Segment, SSF FSM and the MergeCallSegments operation in the BCSM SDL -model are not possible for calls which the service control function SCF has initiated with the operation InitiateCallAttempt. In other words, within the ITU-T and ETSI draft recommendations this operation is not allowed, but the draft recommendations define the states in which the MergeCallSegments opera- tion is allowed, and this situation is not one of them.

The object of the present invention is thus to provide a method for producing services in such a way that the above-mentioned problems relating to production of services can be solved. The objects of the invention are achieved with a method which is characterized by what is stated in the char- acterizing parts of the independent claims. The dependent claims relate to the preferred embodiments of the invention.

The invention is based on complementing an intelligent network ar- chitecture so as to enable automatic merging of calls, without the subscriber controlling the merging with the keys of a terminal. At a predetermined time, the exchange sets up a conference call by calling the conference participants, i.e. the B subscribers. The answer-signal received from the telephone network about the call being connected to each B subscriber serves as a signal for the service control point SCP of the intelligent network. On the basis of the an- swer-signal, the SCP commands the SSP to merge the calls.

The invention will now be described in more detail, in connection with a preferred embodiment, with reference to the attached drawing in which Figure 1 provides an example of a signalling diagram illustrating the invention.

Figure 1 illustrates a method for setting up a conference call be- tween three B subscribers. The events in the figure proceed from top to bot- tom and they are described accordingly. The service control point SCP com- mands the service switching point SSP to set up a first call segment to a first B subscriber. Parameters CSA1, CS1 and LEGI have been appended to this InitiateCallAttempt message by way of an example. A leg is a connection to the subscriber. A CS (CallSegment) corresponds to one call. A CSA (CallSegmentAssociation) is logical joining of two or more call segments.

Other messages and parameters can be used as well. The symbols placed to the left of the SSP illustrate setting up and connecting of calls. A sur- rogate leg S means that there is no connection to a real subscriber. A call segment indicated by a dashed line means that the connection is being cre-

ated, and a call segment shown with a continuous line means that the connec- tion has been established. In connection with an InitiateCallAttempt message, the SCP also sends a RequestReportBCSMEvent message by which it asks the SSP to report when the event detection point has been reached (the con- nection has been established). With a Continue message, the SCP commands the SSP to initiate the defined events. As a result of this, the SSP sends an Initial Address Message IAM to the telephone network for setting up a call to a first B subscriber. The events are repeated for a second and a third B sub- scriber, after which three logical connections to three B subscribers have been established in the SSP call model.

The B subscribers 1, 2 and 3 each answer the call and the network submits for each B subscriber first an Address Complete Message (ACM) and then an ANswer Message (ANM). In connection with the latter, the event de- tection point mentioned above is triggered, and the SSP reports this to the SCP. The establishment of a conference call can be informed to B subscribers in such a way that the SCP sends the messages ConnectToResource, Play- Announcement and DisconnectForwardConnection to the SSP. The first one of these messages defines the resources needed for transmitting an an- nouncement, the second one causes the announcement to be read to a B subscriber and the third one releases the resources reserved for the an- nouncement. For each of these three messages, the SSP sends the SSP/IP (IP = Intelligent Peripheral) a corresponding command on the basis of which the SSP/IP reads an audio announcement to the B subscriber.

The phases indicated by a brace in Figure 1 are repeated for each B subscriber with n values 1,2 and 3. After receiving an ANswer Message and the subsequent notification of the triggering of the event detection point from all B subscribers 1, 2 and 3, the SCP commands the SSP to merge the call segments two at a time. The figure shows an example where segments 2 and 3 are merged first and then segments 1 and 2. The result is a single call seg- ment with no A subscriber, but a Surrogate Leg S in place of the A subscriber, and with three B subscribers. The reason for merging call segments two at a time is that known intelligent network messages (in this case MergeCallSeg- ments) include two call segments as parameters, so n call segments may be merged by performing n-l MergeCallSegment operations. If the intelligent network used recognises a message allowing several call segments to be merged by a single operation, such a message can naturally be employed.

The invention therefore requires for instance the following devia- tions from the ITU-T and ETSI draft recommendations. When two call seg- ments initiated by the SCP (for example CS2 and CS3) are merged, the pas- sive legs (the connections set up to real subscribers) are transferred from one call segment (CS3) to the target call segment (CS2) and the surrogate legs S in the source call segment (CS3) are deleted so that only one surrogate leg S will be left in the target call segment (CS2).

The embodiment described above is only one example of imple- menting the invention. For instance, it is not technically necessary to have all the call segments created before their merging begins. However, it is more convenient for the subscribers that the merging of the call segments only starts after they all have been created. Hearing a ringing tone during the call will thus be avoided. It is also technically possible to merge the call segments without an audio announcement (PlayAnnouncement) to the B subscribers.

However, in this case the B subscribers would need to constantly ask who are presently connected to the conference.

The method of the invention only requires fairly minor changes to a previously known intelligent network architecture. Although the described conference call is between three parties, it is apparent that the technology can be applied to an arbitrary number of conference parties. The invention is therefore not restricted to the preferred embodiments described above but it may vary within the scope of the claims.