WO2001011835A1

WO2001011835A1 - Bandwidth management in a communications system using circuit-switched and packet-switched networks

Info

Publication number: WO2001011835A1
Application number: PCT/US2000/040569
Authority: WO
Inventors: Dale Scholtens; Anita B. Marsh; Michael J. Klemm
Original assignee: Tellabs Operations, Inc.
Priority date: 1999-08-06
Filing date: 2000-08-04
Publication date: 2001-02-15
Also published as: CA2379437A1; WO2001011825A3; WO2001011853A2; WO2001011825A2; WO2001011825A9; WO2001011853A9; AU7628400A; CA2379093A1; WO2001011836A1; CA2381464A1; AU7628200A; AU7628300A; CA2381467A1; WO2001011853A8; WO2001011853A3; AU7628100A

Abstract

Managing resources in a communications system can include terminating a call delivered over a virtual packet network connection and caching the packet network connection for possible use in carrying narrowband traffic from a subsequent circuit-switched call. The techniques can provide a way to efficiently manage packet-domain bandwidth used for carrying narrowband calls, while minimizing signaling demand on the broadband network.

Description

BANDWIDTH MANAGEMENT IN A COMMUNICATIONS SYSTEM USING CIRCUIT-SWITCHED AND PACKET-SWITCHED NETWORKS

BACKGROUND The invention relates to bandwidth management in a communications system using circuit-switched and packet-switched networks.

A traditional telephone exchange configuration provides circuit connections between remote locations. Many of the telecommunications networks currently used are synchronous digital networks. Digitized voice communications are transmitted synchronously over the networks at a fixed rate. Discrete time periods (time slots) are packed with the digital information for a particular call, and digital information for multiple calls can be packed sequentially to form a time division multiplexed (TDM) data stream. The connections may be provided, for example, using network switches having dedicated inter-switch connections. Because the number of inter-switch connections is static, the number of incoming circuits that can be routed to each output port of the exchange also is static.

Situations may arise in which the demand for connections to a particular location reaches its limit, while the demand for connections to another location is below its limit. In such cases, it would be advantageous to be able to reconfigure the network connections to allow more circuits to be connected to the location having the high demand. However, that is not possible in a system having dedicated inter-switch connections.

Packet-domain network architectures, such as asynchronous transfer mode (ATM) networks, allow connections to be made between endpoints without dedicated inter-switch connections. Fixed-size packets of data, known as cells, are transferred between the ATM switches, which are packet switches that provide virtual circuits between the end points of a network. The virtual circuits may be reconfigured depending upon data traffic volume. Hence, an ATM network can provide a more efficient way to connect end points in a network with rapidly changing connectivity requirements. On the other hand, although broadband packet networks are suited for carrying large amounts of information, some packet networks are not well-suited for handling the overhead required to establish and remove a high volume of packet connections.

SUMMARY

According to one aspect, a method of managing resources in a communications system includes terminating a call delivered over a virtual packet network connection and caching the packet network connection for possible use in carrying narrowband traffic from a subsequent circuit-switched call. Some implementations may include one or more of the following features. If the cached connection remains idle for a predetermined duration or for the average duration of a call, it can be removed. Whether a virtual packet network connection should be cached can depend on various factors, including an anticipated pattern of future narrowband traffic. According to another aspect, a method of delivering narrowband traffic includes caching a virtual connection in a packet network and using the cached connection to transport circuit-switched narrowband traffic over the packet network.

A particular cached connection can be selected to carry the narrowband traffic from among multiple cached connections in the packet network. The particular cached connection can connect gateways configured to perform adaptations between circuit-switched bearers and packet-switched bearers. The particular cached connection can be selected from among existing connections to a gateway that is hosting a call for the narrowband traffic.

The techniques also can include identifying a first set of gateways capable of servicing a call for the narrowband traffic and identifying a second set of gateways having an existing virtual connection through the packet network to a gateway that is hosting the call. The cached connection can connect to a gateway that is a member of both the first and second sets.

In another aspect, a call controller for use in a communications system can be configured to provide control signals to cause a packet network connection to be cached for possible use in carrying narrowband traffic for a subsequent circuit- switched call. A communications system including such a call controller also is disclosed.

Various implementations include one or more of the following advantages. Asynchronous bandwidth management can be provided for a system that allows circuit-switched traffic to be transported dynamically over packet-domain networks under Signaling System 7 (SS7) control. Service providers can consolidate circuit- switched and variable bit rate data services over a single, broadband network. High- quality narrowband services can be delivered over packet and cell networks. The system can provide better network utilization by ensuring that bandwidth is allocated where it is needed and not stranded in dedicated trunks. The techniques provide a way to efficiently manage packet-domain bandwidth used for carrying narrowband calls, while minimizing signaling demand on the broadband network.

Other features and advantages will be readily apparent from the following detail description, the accompanying drawings and the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a telephone connection through a hybrid ATM network and an associated signaling network. FIG. 2 is a simplified block diagram of an exemplary media gateway.

FIG. 3 illustrates various views through switching systems in the hybrid network.

FIGS. 4A and 4B are a flow chart of a method of selecting a gateway to service a narrowband call. FIG. 5 is a diagram illustrating selection of a gateway to service a narrowband call. DETAILED DESCRIPTION

A call control mechanism forming part of a switching system for carrying narrowband traffic, such as voice calls, modem data or facsimile data, over an asynchronous transfer mode (ATM) or other packet network connection is described below. The switching system supports narrowband calls over a packet backbone network, while interfacing seamlessly with an existing SS7-based public telephone network that uses circuit-switched technology. In particular, the system provides a way to efficiently manage packet-domain bandwidth used for carrying narrowband calls, while minimizing signaling demand on the broadband network. As shown in FIG. 1, a continuous call path is established starting with a narrowband SS7 call originating, for example, in a Public Switched Telephone Network (PSTN) 102A, using a virtual circuit over an ATM network 101, and completing on the terminating side in a narrowband circuit-switched SS7 call to the terminating subscriber through another circuit switched network 102B. The control mechanism interacts with the circuit-switched and packet-switched networks to correlate SS7 and ATM connections to establish a single continuous information path.

A large number of individual telephone circuits, such as DSO circuits, that are to be connected to the packet network 101 can be carried, for example, on fiber optic carriers 105 using time-division multiplexing (TDM) according to the Telcordia Synchronous Optical Network (SONET) standards. The TDM carriers 105 are coupled to access ports 116 (see FIG. 2) in media gateways 100A, 100B.

The gateways 100A, 100B can adapt the TDM telephone line signals to packet-based signals and vice-versa. Each gateway 100A, 100B can separate incoming TDM signals into individual DSO signal streams. The TDM telephone signals are circuit-switched, in other words, the bit stream can be divided temporally into individual DSO circuits. By contrast, in packet-based signals, the bit stream can be divided according to the destination address of each packet.

In one implementation, shown in FIG. 2, each gateway, such as the gateway 100A, includes a TDM switching matrix 117 that provides full switching capabilities. The switching matrices 117 permit the DSO circuits to be interconnected flexibly with narrowband channels appearing on the gateways. Echo cancellation and other digital signal processing functions can be performed in a digital signal processing portion 118 of each gateway. The DSO streams are adapted by an ATM adaptation layer 120 into ATM cells. The ATM adaptation layer 120 combines incoming DSO signals from a particular carrier 105 into payloads for ATM cells. A header can be provided as part of each cell and can be interpreted to identify which call the ATM cell is associated with. After the payload of a cell is loaded with data, the cell is inserted through ATM ports 121 into an ATM cell stream that traverses an ATM network 101. Each gateway includes a control section 119 that controls overall operation of the gateway. In one implementation, the gateways 100 A, 100B are implemented as Salix 7720 Class-Independent Switches manufactured by Tellabs Operations, Inc.

As illustrated in FIG. 1, each gateway 100A, 100B is connected to a respective ATM end point switch 1 15, which may be co-located with the gateway. The connection between a gateway and an ATM end point switch 115 and the connection between the ATM switch end point and the ATM network 101 are user-network interfaces (UNIs). Within the ATM network 101, there are a number of interconnected ATM switches 1 10. Those connections are network-network or network-node interfaces (NNI).

A call control network 126, which can form part of an existing telephone system, runs parallel to the ATM network 101. The call control network 126 primarily controls telephone switching equipment to connect the originating and terminating ends of a telephone call using Signaling System 7 (SS7) messages. A call controller 120 A, 120B is coupled to each gateway 100A, 100B and provides an interface between the gateway and the call control network 126. Circuit-switched traffic can be transported dynamically over the packet-domain network 101 under SS7 message control. Furthermore, multiple circuit-switched calls can be multiplexed over a single ATM connection.

In one implementation, each call controller 120A, 120B includes a narrowband call control (NCC) unit 122 and a bandwidth parceller (BP) unit 124. The NCC unit 122 and the BP unit 124 may be part of the same physical unit or may comprise separate physical units. In general, the NCC unit 122 keeps track of the narrowband trunk circuits that connect the switching system to neighboring systems. The BP unit 124 provides broadband call processing in support of narrowband calls, including exercising discretion over the placement of packet-domain calls to create the packet network connections needed to support the portion of each narrowband call that traverses the packet network.

Preferably, the BP unit 124 manages the set of packet-domain calls to reduce the number of connections through the packet network 101 actually being established and or removed. In other words, the BP unit 124 attempts to reduce the signaling load on the packet network 101. The BP unit 124 also can be configured to cause the time- variant bandwidth demand on the packet network 101 to track the historical bandwidth profile of the narrowband traffic.

FIG. 3 illustrates various views of the switching system as seen by the NCC unit 122 and the BP unit 124 of FIG. 1. In general, a set of narrowband trunk groups may span multiple gateways 100A, 100B. As seen through section A-A, the NCC unit 122 keeps track of the arrangement of physical trunks, including the physical

DSO circuits provisioned within various trunk groups. Each DSO circuit is assigned as a member of a single trunk group. The amount of bandwidth associated with a group, therefore, depends on the number of DSO circuits provisioned within the group. The total bandwidth as seen through section A-A is the sum of the bandwidths of the individual DSO circuits provisioned in the various trunk groups.

Viewed from the perspective of the NCC unit 122 in the originating call controller 120A, the appearance of an incoming trunk is fixed. A call on a particular DSO trunk is assigned resources from the hosting gateway. The NCC unit 122 in the originating call controller 120A and the gateway 100A are responsible for assuring that gateway resources are available for the call.

From the perspective of the NCC unit 122 as seen through section C-C, a single virtual trunk group is defined between each pair of switching systems.

As described in greater detail below, the BP unit 124 handles requests for connections from the associated NCC unit 122. As seen through section B-B, the BP unit 124 reserves gateway resources and keeps track of virtual connections established to other gateways through the ATM network 101. The BP unit 124 also keeps track of the packet-domain connections actually in use by the NCC unit 122.

In general, when a narrowband call is terminated, the BP unit 124 can disassociate the previously-allocated packet-domain resources from the call. However, to help reduce the total number of packet-domain connections that must be established and removed, the BP unit 124 can cache some or all packet-domain connections when the corresponding narrowband calls are terminated. By caching some of the packet-domain connections in the ATM network 101 even after the corresponding calls have been terminated, the signaling load on the network can be reduced. For example, when a new narrowband call requires that a path be established through the network 101, an existing, idle ATM connection can be used, thereby avoiding the overhead associated with establishing a new virtual circuit in the ATM network. The number and duration of ATM connections that are cached at any given time, however, should be balanced against the fact that caching packet-domain connections can reduce the bandwidth available for paths through other gateways.

In one implementation, each ATM connection is cached at least for a predetermined duration after handling a call that has been terminated. At the end of the predetermined duration, if the ATM connection remains idle, it is removed. In an exemplary implementation, the predetermined duration is set to a fixed length of time. Alternatively, the predetermined duration can be set equal to the average duration of a call, where the average is taken over some specified period of time. In that case, as well as in other cases, the predetermined duration may vary with time.

In another implementation, the BP unit 124 caches packet-domain connections so that the target number of cached connections between any given pair of gateways falls within a specified range. The minimum and maximum number of cached connections may vary with time and can be based, for example, on an anticipated pattern of narrowband calls and may be determined heuristically. If the maximum number of cached connections already exists and the last call on a particular connection is terminated, then the ATM connection associated with the terminated call may be removed. In general, the trunk groups may span multiple gateways within each switching system. Packet network connections between any pair of gateways in the same or different switching systems can be cached.

As described in greater detail below, the cached ATM connections can facilitate setting up a path through the packet network 101 for a particular narrowband call.

As indicated by FIGS. 4A and 4B, to establish a path for a narrowband call through the ATM network 101 to the destination point, the originating call controller 120A determines 200 that the call must be terminated at a gateway associated with another call controller in the network. The originating call controller 120 A selects an appropriate terminating call controller to handle the call, such as the call controller 120B, and signals that call controller. In one implementation, the call controllers 120A, 120B communicate using SS7 messages. The message sent to the terminating call controller 120B can include a connection descriptor that identifies the originating gateway 100A allocated to handle the call and that uniquely identifies the call.

The NCC unit 122 in the call controller 120B routes the call and identifies 202 a set of one or more narrowband circuits (e.g., DSO circuits) in the circuit-switched network 102B and their associated gateways 100B that are capable of servicing the call to the destination point. The NCC unit 122 sends 204 the identified information to the BP unit 124.

The BP unit 124 uses the information received from the NCC unit 122 to determine which particular gateway 100B and DSO circuit should handle the call. In particular, the BP unit 124 identifies 206 the set of gateways 100B (if any) having existing virtual connections to the originating gateway 100A that is handling the call. The existing virtual connections can include either an idle switched virtual circuit (SVC) in the network 101 or an unused channel in an active SVC. If one or more gateways 100B have a virtual connection to the gateway 100A that is hosting the call, the BP unit 124 identifies which of those gateways 100B (if any) is capable of handling the call based on the information previously received from the NCC unit 122. In other words, as shown in FIG. 5, the BP unit 124 identifies 208 the intersection between the set of gateways 100B previously identified by the NCC unit 122 as being capable of handling the call and the set of gateways that have an existing virtual connection to the hosting gateway 100A.

If more than one of the gateways 100B is capable of servicing the call and also has an existing virtual connection to the gateway 100A hosting the call, then the BP unit 124 selects 210 one of the gateways. For example, where load balancing among the gateways is desirable, the BP unit 124 would select the gateway that currently has the fewest number of calls active on its interface to the ATM network 101. In other words, the gateway 100B with the greatest available bandwidth can be chosen to handle the call. In other implementations, it may desirable to concentrate the traffic on only a few gateways so as to minimize the number of cached connections in the packet network.

If only one gateway 100B is identified as being capable of servicing the call and as having an existing virtual connection to the host gateway 100A, then the BP unit 124 selects 212 that gateway to service the call. The identification of the gateway 100B selected by the BP unit 124 is sent to

214 to the NCC unit 122. User-to-user signals can then be used to transfer 216 a packet connection descriptor from the terminating call controller 120B to the originating call controller 120A so that both ends of the connection are aware of the ATM connection being used for the call. If the trunk associated with the selected gateway is no longer available to handle the call, then the NCC unit 122 can request 218 the BP unit 124 to reiterate the selection process. In some implementations, the NCC unit 122 may request the BP unit 124 to use a round-robin or other technique to select the gateway to handle the call. In the event that there is no existing ATM connection between a gateway

100B that can service the call at the terminating side and the gateway 100A hosting the call, then the BP unit 124 establishes 218 a new packet connection between one of the gateways 100B identified by the NCC unit 122 as being capable of handling the call and the hosting gateway. For example, a packet connection from the gateway 100B having the greatest number of available trunks can be made to reduce the likelihood of glare at the terminating end while the packet-domain connection is established. The BP unit then informs the NCC unit 122 of the particular gateway 100B for which an ATM connection is available.

According to another implementation, if there is no existing packet-domain connection between the selected gateway 100B and the host gateway 100A, the BP unit 124 would send a message to the NCC unit 122 indicating that an existing connection is not available. Although a connection is not available, the BP unit 124 can register the request and use the information to anticipate future requests for connections through the ATM network 101. Thus, the BP unit 124 can adapt the number and/or duration of cached packet-domain connections to the existing traffic load. Upon receiving the message from the BP unit 124, the NCC unit 122 would be able to select a different gateway 100B and make another request to the BP unit 124 to establish a connection through the packet-domain network 101.

In general, the call controllers 120A, 120B can serve at either the originating or terminating ends of a call. Although the foregoing techniques have been described for a system having separate originating and terminating call controllers 120A, 120B, in some cases a single call controller may perform the functions of both call controllers.

Similarly, although the foregoing implementations have been described with respect to ATM networks, circuit-switched traffic can be transported over other packet-domain networks, such as frame relay, Ethernet and Internet Protocol (IP) networks, as well.

Various features of the system can be implemented in hardware, software, or a combination of hardware and software. For example, some aspects of the system can be implemented in computer programs executing on programmable computers. Each program can be implemented in a high level procedural or object-oriented programming language to communicate with a computer system. Furthermore, each such computer program can be stored on a storage medium, such as read-only- memory (ROM) readable by a general or special purpose programmable computer, for configuring and operating the computer when the storage medium is read by the computer to perform the functions described above. Other implementations are within the scope of the claims.

Claims

What is claimed is:

1. A method of managing resources in a communications system comprising: terminating a call delivered over a virtual packet network connection; and caching the packet network connection for possible use in carrying narrowband traffic from a subsequent circuit-switched call.

2. The method of claim 1 including removing the cached connection if the cached connection remains idle for a predetermined duration.

3. The method of claim 1 including removing the cached connection if the cached connection remains idle for the average duration of a call.

4. The method of claim 1 including determining whether a virtual packet network connection should be cached, wherein the determination depends on an anticipated pattern of future narrowband traffic.

5. The method of claim 1 including determining whether a virtual packet network connection between a pair of gateways should be cached, wherein the determination depends on whether a number of currently cached virtual packet network connections between the pair of gateways falls within a targeted range, wherein the gateways are configured to perform adaptations between circuit-switched bearers and packet-switched bearers.

6. A method of delivering narrowband traffic comprising: caching a virtual connection in a packet network; and using the cached connection to transport circuit-switched narrowband traffic over the packet network.

7. The method of claim 6 including selecting a particular cached connection to carry the narrowband traffic from among a plurality of cached connections in the packet network.

8. The method of claim 7 wherein the particular cached connection connects gateways configured to perform adaptations between circuit-switched bearers and packet-switched bearers.

9. The method of claim 8 wherein the particular cached connection is selected from among a plurality of existing connections to a gateway that is hosting a call for the narrowband traffic.

10. The method of claim 6 including: identifying a first set of gateways capable of servicing a call for the narrowband traffic; and identifying a second set of gateways having an existing virtual connection through the packet network to a gateway that is hosting the call, wherein the cached connection connects to a gateway that is a member of both the first and second sets.

11. A call controller for use in a communications system, wherein the call controller is configured to provide control signals to cause a packet network connection to be cached for possible use in carrying narrowband traffic for a subsequent circuit-switched call.

12. The call controller of claim 1 1 configured to provide control signals to cause the cached connection to be removed if the connection remains idle for a predetermined duration.

13. The call controller of claim 11 configured to determine whether a virtual packet network connection should be cached, wherein the determination depends on an anticipated pattern of future narrowband traffic.

14. The call controller of claim 11 configured to determine whether a virtual packet network connection between a pair of gateways should be cached, wherein the determination depends on whether a number of currently cached virtual packet network connections between the pair of gateways falls within a targeted range, and wherein the gateways are configured to perform adaptations between circuit-switched bearers and packet-switched bearers.

15. A communications system comprising: at least one circuit-switched network including first and second interface points; a packet network; first and second sets of gateways coupled respectively to the packet network, wherein the first set of gateways is further coupled to the first interface and the second set of gateways is coupled to the second interface, and wherein the gateways are configured to perform adaptations between circuit-switched bearers and packet- switched bearers; and at least one controller configured to provide control signals to cause a connection in the packet network to be cached for possible use in carrying narrowband traffic from a subsequent circuit-switched call.

16. The system of claim 15 wherein the call controller is configured to provide control signals to cause the cached connection to be removed if the connection remains idle for a predetermined duration.

17. The system of claim 15 wherein the call controller is configured to determine whether a virtual packet network connection should be cached, wherein the determination depends on an anticipated pattern of future narrowband traffic across the packet network.

18. The system of claim 15 wherein the call controller is configured to determine whether a virtual packet network connection between a particular gateway in the first set and a particular gateway in the second set should be cached, wherein the determination depends on whether a number of currently cached virtual packet network connections between the particular gateways falls within a targeted range.

19. An article comprising a computer-readable storage medium including computer-executable instructions for causing a computer system to: cache a virtual connection in a packet network; and use the cached connection to transport circuit-switched narrowband traffic over the packet network

20. The article of claim 19 including instructions to cause the computer system to: identify a first set of gateways capable of servicing a call for the narrowband traffic; and identify a second set of gateways having an existing virtual connection through the packet network to a gateway that is hosting the call, wherein the cached connection connects to a gateway that is a member of both the first and second sets.

21. An article comprising a computer-readable storage medium including computer-executable instructions for causing a computer system to: terminate a call delivered over a virtual packet network connection; and cache the packet network connection for possible use in carrying narrowband traffic from a subsequent circuit-switched call.