KR101206670B1 - Overload control in a communications network - Google Patents

Abstract

An adaptive overload system that controls the amount of traffic handled by a network access controller is disclosed for a network access controller that controls a plurality of network access points. Each network access point provides traffic access to the communication network and the system determines if an overload condition exists at the network access controller, and if present, a rate at which the network access point grants the traffic to the communication network. Creating one or more global constraints. The controller then multicasts one or more global constraints to one or more of the plurality of network access points. Each access point receiving each network global constraint then processes the global traffic constraint to determine a plurality of local constraints. The receiving network access point determines the local constraint by performing the following steps: determining a predetermined local gap interval to be imposed on the traffic; And determining an initial local gap interval that is different from the subsequent predetermined local gap interval. Wherein the initial local gap spacing is different between each of the plurality of network access points. The initial local gap interval is determined in a random or pseudo-random manner to ensure that the synchronization impact is eliminated in the network access controller that might otherwise occur in high call rate scenarios.

Description

OVERLOAD CONTROL IN A COMMUNICATIONS NETWORK}

TECHNICAL FIELD The present invention relates to overload control in a communication network, and more particularly, a rate at which an offered call provided in a VoIP network is received by a media gateway controller (MGC) from a media gateway (MG). It relates to an external overload control system for the MGC limiting but not limited to.

Remote voting (a telephone number is broadcast to allow users to vote) and similar mass calling schemes often result in very high call rates that start suddenly and last a relatively short time. do. To deal with this overwhelming congestion in traffic, it is not economical to provide sufficient network capacity and force overload control to be implemented in the communication network so that emergency services and other critical services can be supported. However, traditional methods for handling sudden congestion in the number of calls sent to a particular network address are not satisfactory with the development of the communication network itself.

Conventional methods of providing overload control in a communication network include a call gapping technique that seeks to limit the number of calls formed. Such call mapping techniques are well known to those skilled in the art and may block calls received within a predetermined time interval (the gap) following the first call that initiates the onset of the gap. It includes blocking. One example of a call mapping technique is described by way of example in US Pat. No. 6,259,776 under the name “System for Controlling Communication Overload Traffic,” the contents of which are incorporated herein by reference.

However, conventional communication networks are increasingly evolving to provide more functionality and support for media other than those provided by conventional PSTNs. For example, a call concentrator in the PSTN may be replaced by an access media gateway (MG) that converts conventional copper to provide access to IP media transport. In such a communication network, the MG is controlled by the MGC, which is similar to a local exchange processor implemented in a conventional PSTN and performs a traffic analysis function by analyzing dialed digits to determine the path of calls. More details on MG and MGC can be found in the MGC (MEGACO) charter standard document available from the IETF standard (url http://www.ietf.org/html.chaters/megaco-charter.html).

Whenever an incoming phone number is advertised by country and a significant number of customers try to call the phone number, there is an overload of concentrated calls to use incoming trunks to the destination main switching unit and destination local exchange. As a result, the exchange of normal service traffic may be disrupted. Several techniques have been proposed to address this problem in conventional PSTNs, such as, for example, US Pat. No. 6,259,776. U. S. Patent No. 6,259, 776 describes a communication network that includes an overload control configuration that restricts a call connection to a destination when traffic to a destination exceeds a certain level. The configuration includes a plurality of identical overload control functions executed at each of a plurality of nodes of the network and having respective gapping periods determined from the overload levels sensed at each node, wherein the overload control function includes a respective overload control function. The data defining the gaping periods are exchanged and adjusted towards the average gaping period to avoid substantial differences between each gaping periods from each node to any one desired destination.

While the overload control system described in US Pat. No. 6,259,776 provides an effective solution in a conventional SS7-type network, as shown in FIG. 1 of the accompanying drawings, multiple network access points (A ₁ ... A _N ) are single. It is less effective for IP networks or similar forms of communication that may be under the control of the network access controller X ₁ . In such a situation, a significant overload condition on the network is related to the maximum call processing capacity of the controller X ₁ with only a finite amount of resources available to handle calls to enter the network. This limitation on the number of calls granted to the network is shown schematically in FIG. 2.

In FIG. 2, the x axis represents the rate of calls provided to the network access controller by the network access point and the y axis represents the number of calls granted to the network by the network access controller. The total number of calls offered by all of the access points A ₁ ... A _N , actually approved by the controller X ₁ as a function of the total provided rate, is shown by the solid line ( Thus this shows the call rate approved for the network). If the provided rate is relatively low, the approved rate is increased to match the provided currency quantity. However, the controller has only limited resources and as the use of those resources increases, the controller eventually becomes overloaded. This occurs in the portion labeled A in FIG. 2, at which point the controller needs to reject a certain percentage of the received off-hook signal to keep the response time relatively low.

As the number of new calls offered per second, i.e., the provided call rate, increases beyond point A, the admission rate does not increase rapidly, so that the approval rate of the call to the network is at a given rate of the provided call. To the maximum value (L _M ). As a result, when the number of provided calls reaches the rate Lc, all resources of the controller are used to reject the call and new calls will not be accepted.

The internal control mechanism of the network access controller is reflected in the diagram shown in FIG. Internal control of the access controller provides the ability to reject some or all of the provided loads, and does not provide the provided loads with the ability to regulate any external restrictions (such as gapping). .

The dashed line in FIG. 2 shows the response time of the network access controller to a signal (eg, an off-hook signal) that the network access controller receives from a network access point in its control area. Initially, before reaching the overload point, the controller will slowly increase the response time while reliably handling more and more provided calls. The access controller's ability to reject provided calls needs to be combined with effective external constraints if the access controller must regulate its response time. The rate provided to the access controller in order to maximize the processing capability (throughput) of the access controller, L _M adaptive in some form in order to ensure that the relatively maintained close to each time the traffic supplied to the network access point exceeds L _M You need to implement external limit control.

One form of adaptive external control known to those skilled in the art is that provided by a call gapping overload system. The call mapping process allows the load provided to the access controller by the network access point to remain in the vicinity of the L _M , which allows the access controller's response time to remain relatively constant. However, if no external control is implemented, or if the external control is not sufficient to limit the provided call rate to near L _M , and then the provided rate increases until reaching L _C , then an overload controller is implemented. The internal overload control process substantially reduces the throughput of the access controller to zero, which results in the traffic not being allowed to the network at all.

Conventional currency mapping processes also have other limitations. For example, if the process is applied in a system where multiple network access points are controlled by a single access controller (known in the art as a very high “fan-in”), then traffic is externally restricted. The rate approved by (i.e., applied by a network access point) is too slow to respond to a command from the control point to change the approved rate (i.e. used by a gapping process to implement the external restriction. The gap spacing cannot be changed). This slow response by the network access point may result in slowing and unstable the overload control servo loop.

Other causes of delay that contribute to this slow response problem are:

Delay in sending the control message from the access controller to the network access point due to the large amount of control messages to be transmitted;

The first provided call is always accepted by an external restriction in a conventional call mapping process when a call restriction is initially imposed, and then this is an active interval that is executed simultaneously with all of the network access points impose a call gap. interval) having a timer generates a synchronization effect; And

If a gap interval update is already applied to a gapped access point, it includes a delay waiting for the existing gap interval (interval) timer to expire before the updated gap can affect the approved rate.

As such, the techniques imposed by conventional call mapping are no longer effective in situations where significant overload conditions occur in an access controller that controls many, more preferably very many, for example, thousands of access points.

One object of the present invention is to provide an adaptive external overload control technique for the communication network in which the access controller controls a plurality of network access points configured to grant traffic to the communication network. In particular, one object of the present invention is to provide a method for the network that prevents concentrated overload calls in the access controller from problems that may occur when an access controller such as a Media Gateway Controller (MGC) detects an overload condition. By providing an overload control system, it is intended to mitigate and / or eliminate. In particular, it is an overload control system that prevents synchronously repeated overloads from occurring in the network. Here, the term "focussed overload" refers to traffic that is concentrated on a specific access controller, as opposed to the process for authorization for the network controlling the overload at its destination. To a specific address or part of an address within). The traffic itself includes voice and / or data traffic (eg, in the form of short message service (SMS) or electronic mail).

Another object of the present invention is to externally provide a call rate provided by an access controller by multicasting a scalable call gap interval defined by the access controller to a plurality of network access points configured to provide calls to the access controller. It is to provide a regulation overload control system. Each network access point can change the adjustable call gap according to individual characteristics. An example of an individual characteristic of a network access point is, for example, the number of circuits on which the access point receives traffic requiring approval for the network.

Another object of the present invention is to regulate the traffic according to the specific network destination address of the traffic when authorized by each network access point.

Composition of the Invention

Forms and preferred features of the invention are provided in the appended claims.

Preferably, the present invention provides an adaptive overload control system for controlling the amount of traffic processed by a network access controller configured to control a plurality of network access points. Each network access point provides traffic to the communication network and the system determines whether an overload condition exists in the network access controller, and if so, the network access point grants access to the traffic to the communication network. Create at least one global constraint to limit the rate. Examples of global constraints include the requested rate per access line managed by the access point. The global constraint is thus determined for all network access points by the access controller determining if the rate at which traffic is being provided to the access controller for processing has exceeded a predetermined limit. The global constraint created by the controller may specifically be for traffic whose destination includes a particular address. Alternatively, the global constraints generated by the network access controller may instead include basic global constraints applied to the traffic independently of the destination address of the traffic. The controller then passes the at least one global constraint to the one or more network access points. It is preferable if the communication technology used is a multicasting (or broadcasting) type of technology.

Each network access point receiving the global constraint then processes the global traffic constraint to determine a plurality of local constraints. The receiving network access point performs the following steps to determine the local constraints: determining a predetermined local gap interval imposed on the traffic; And determining an initial gap interval that is different between each of the plurality of network access points and is different from subsequent predetermined local gap intervals. The initial gap interval is determined in a random or pseudo-random manner to ensure that the synchronization effects at the network access controller that would otherwise occur in high call rate scenarios are eliminated.

The initial gap interval is applied immediately without the need to be activated by call arrival. If an initial gap interval of 0 is applied, the next call received by the media gateway will prompt the imposition of a standard gap interval.

Other aspects of the invention are provided by the appended claims and preferred features are recited by the dependent claims. The above preferred features may be suitably combined with any aspect of the present invention, where such suitable combinations will be apparent to those skilled in the art.

As such, the present invention provides a method for controlling the rate at which traffic to access a communication network is handled by an access controller, and more particularly, a method for controlling a call rate provided at a media gateway controller. Advantageously, the grant rate at each access point within the control area of the access controller can be changed dynamically and almost instantaneously in response to the restriction requested by the access controller.

1 is a schematic diagram of a communication network,

2 shows overload results of an access controller in a communication network,

3 shows a fixed gap spacing overload control scheme,

4 shows the number of calls provided to the access controller as a function of time for a fixed gap interval scheme implementing the Crawford algorithm,

5 shows the steps in the overload control process according to the invention,

6 illustrates an embodiment of the invention in which the initial gap interval is varied between different network access points in a communication network,

7 shows the results of applying an embodiment of the present invention to the number of calls provided to a network access controller,

FIG. 8 is an enlarged part of the configuration shown in FIG. 7;

9 is a result of applying another embodiment of the present invention to the number of calls provided to a network access controller,

FIG. 10 is an enlarged view of a part of the configuration shown in FIG. 9;

11 shows a message flow between a media gateway and a media gateway controller in an embodiment of the invention,

12 schematically illustrates how overload control can be implemented based on a telephone number in an embodiment of the invention,

13 illustrates a locally defined constraint update process, and

14 illustrates a globally defined constraint update process.

Best Mode for Carrying Out the Invention The best mode of the present invention contemplated by the inventor will now be described with reference to the drawings. Those skilled in the art deem that the embodiments described in detail below are merely exemplary and alternatively those features and steps that are clearly equivalent to those described herein are implied to be included as an alternative to the features described herein. It will be appreciated. Therefore, the scope of the present invention is defined by the appended claims. In particular, when “call” is referred to, this term should not be construed to include only voice calls, it can be supported by the network and is equivalent to the manner in which conventional calls are handled and And / or any other form of communication known to those skilled in the art that can be converted to a form that can be processed by an access controller (eg, a short message service call and / or an emergency message call and / or an email). Currency). Similarly, the term "traffic" refers to a "call" consisting of voice traffic or other forms of communication traffic, such as short message text message traffic, e-mail traffic, emergency message traffic, etc., within a VoIP network. (calls) ". Thus, as mentioned above, the present invention is primarily described with respect to voice calls, but those skilled in the art will appreciate that the term "call" includes traffic that includes other media such as e-mail or sms text messages. It will be appreciated that "callers" may cause concentrated agitation within the network when communicating with more than one address.

In FIG. 1, a communication network 1 is shown in which access is requested by a number of network access points A ₁ ... A _N , B ₁ ... B _M , C ₁ ... C _L. Access to the communication network 1 via an access point A ₁ ... A _N , B ₁ ... B _M , C ₁ ... C _{L is} provided by one or more access controllers such as access controllers X ₁ , X ₂ , X ₃ ). As shown in Fig. 1, network access via A ₁ ... A _N is controlled by X ₁ , and network access via B ₁ ... B _M is controlled by X ₂ and the like. In general, the number of access points controlled by individual access controllers is quite high, for example hundreds of access points (and more) can be controlled by a single access controller. Each access point may be connected to a different number of incoming lines, such that access point A ₁ may be connected to only one line, access point A ₂ may be connected to 25 lines, and so on. Not shown in 1. A single access controller such as X ₁ may therefore have a number of different types of access points to control within its control area, which network access points differ not only in the number of lines providing access to the network, but also in their capacity and the like. .

In one embodiment of the present invention, the communication network consists of an Internet Protocol (IP) network, and the network access point includes media gateways (MGs) capable of supporting the conversion of voice traffic into IP traffic, and thus VoIP (Voice) over IP) service to be supported. The rate at which traffic (typically a call) is authorized to the IP communication network is regulated by one or more network access controllers. In this embodiment, each network access controller includes media gateway controllers (MGCs).

More generally, the term “network access point” refers to a point in the area of a communication network that functions to provide access to the communication network from another network, and the term “network access controller” refers to the network access point. It refers to a point within the communication network area that provides control.

Media Gateway Control Protocol Architecture and Requirements Standard Document RFC 2805, MEGACO Protocol RFC 2885 (now obsolete by RFC 3015), MEGACO Modified RFC 2886 (now obsolete by RFC 3015), now obsolete by RFC 3525 MEGACO protocol (with modifications) RFC 3015, and Megaco IP Telephony Media Gateway Application Profile (RFC 3054) and Gateway Control Protocol Version 1 (RFC 3525) jointly form part of the MEGACO standard literature relating to the implementation of media gateway control. The term describes the formal definition of the media gateway controller and media gateway. All of the above standard documents are available through the MEGACO charter, which is accessible from the IETF Standards Forum website (url: www.ietf.org). Corresponding standards supporting the media gateway control protocol are provided by ITU-T H.248.1.

These standards define the functionality of the media gateway (MG) as providing media mapping and / or code conversion functions between potentially dissimilar networks, which may be packet, frame or cell networks. For example, MG terminates Switched Circuit Network (SCN) facilities (trunks, loops), packetizes media streams (not yet packetized), and delivers packetized traffic to the packet network. The MG may also be configured to connect a packet (eg VoIP) network to a two-wire wired analog copper transmission line or even another packet (eg ATM) network including multiple access lines by "loop emulation". The MG performs these functions in reverse order for the media stream flowing from the packet network to the SCN. MG is not limited to SCN packet / frame / cell functions: a conference bridge with all packet interfaces as well as a voice recognition system with an Interactive Voice Recognition (IVR) unit, an audio resource function, or a cell interface May also be MG. MGC may be defined as providing a control function for the MG.

In Figure 3, a fixed gap interval call gapping technique is outlined to illustrate the effect that fixed gap interval applied at the network access point may have on the number of calls provided to the network access controller. Is shown. If this form of gaping is applied to the access controller shown in FIG. 1 to regulate traffic entering a communication network, the fixed gap interval call mapping process (e.g., by each of the network access points may be normal Crawford). Implemented using an algorithm may result in the access controller experiencing repetitive call congestion. These repetitive congestions occur in a synchronous fashion and result from the Crawford algorithm, which accepts calls from any MG during the gap interval after accepting the first call from all MGs in its basic form. Synchronization damages the operation of the overload control servo system because the MG cannot deliver the rate provided by the MGC until sufficient time has elapsed for the synchronization to be removed by the randomness of call arrival. This is an unacceptably long time for stable control if the arrival time between calls (for calls provided to MG) is much smaller than the gap interval. The synchronization effect occurs only in the overload response if the call rate is high enough and the gap interval is long enough. Synchronization depends on the ratio of the time of arrival between calls to the call provided to the MG relative to the gap interval that each MG imposes on the arrival call and the distribution of the number of lines in the individual MGs. If all MGs have the same number of lines, the synchronization effect is more obvious, but if the number of lines attached to each MG has a wide and continuous distribution, the response of the arrival rate in the MGC to the change in gap spacing still affects The synchronization impact is less severe even if

Referring to FIG. 3 in detail, the upper portion of the figure schematically shows the number N of access points applying a regular gap interval by implementing the Crawford algorithm, which is well known to those skilled in the art. In the lower part of the figure, a schematic illustration of the number of calls received by a network access controller (eg MGC) when a constant gap interval is imposed is shown as a function of time.

As is well known to those skilled in the art, the Crawford algorithm allows a gap access that allows each network access point to block all subsequent calls for the first time that is accepted but then for a predetermined time (gap interval) g. Intervals can be applied to the call (such blocked calls are indicated by arrows in hatched portions in FIG. 3-unblocked calls are indicated by arrows in unhatched portions). After the first gap interval g has expired, the next call requesting network access is accepted but subsequent call gap intervals apply.

As shown in FIG. 3, initially a high call rate results in a high degree of overlap of the call gap intervals imposed by each of the access points, resulting in a total number of calls processed by the access controller to almost zero. Each network access point (MG # 1 ... N shown by way of example in FIG. 3) forwards the initial, unblocked call to a network access controller (also referred to as a network access controller). Then, for a given gap interval g = Δt, the network access controller (e.g., MGC) experiences a sharp decrease in the total number of calls received. However, at a timescale approximately equivalent to the period of the gap interval g, the number of calls offered to the network access controller by all the access points again increases rapidly. Only as time progresses, the maximum number of calls offered to the access controller decreases and spreads over a wider interval as the gap spacing correlation decreases and the occurrence of overlap decreases. This is illustrated by the wider offset between subsequent gap intervals in FIG. 3. This "synchronization" of the call pattern eventually disappears if the gap is properly maintained for an extended period of time.

This aspect is more clearly seen in the simulation example shown in FIG. 4, in which case the normal Crawford algorithm was used to impose a gapping process on multiple access points including both 25 line MG and single line MG. In FIG. 4, after the access point receives each call, the local gap interval is determined according to the type of network access point. For example, if the MG receives a call over 25 lines, while the other call is blocked (and the caller receives some signal indicating that the call is blocked, such as during a line call or disconnected or another dial tone) ), A 60 second gap is given. However, each single line MG applies a gap spacing of 25 times the gap spacing imparted by the 25 line MG. Thus in this embodiment a gap interval of 1500 seconds is applied by each single line MG. In each type of MG, once the local gap interval expires, the next call received by the network access point is delivered to the access controller and the next gap interval is given while further calls are terminated.

In Fig. 4, the normal Crawford algorithm was applied to the 250,000 lines mixed in the example shown, where 125,000 lines were distributed to 25 line access points and 125,000 lines to a single line access point. The call rate provided from the access point to the access controller is 1600 calls per second (cps), and the target rate of the access controller is 160 cps. This 10-fold overload is thought to be the typical type of intensive congestion that telecom networks can expect to experience as televoting mass-call services become popular.

The overall aspect shown in FIG. 4 shows a synchronized admission profile from the 25 line MG superimposed on the tendency to decrease as the single line MG gradually applies their first gap. The simulation does not represent the synchronization of traffic approved by the single line MG, as this occurs over much longer time than shown in the figure of FIG.

The admission traffic rate shown in FIG. 4 varies for a time constant that approximately corresponds to the call gap interval per access point (ie gateway). However, this is because the first repetitive congestion itself can overload the access controller, and secondly, the access controller can make control update decisions at most once per gap period (i.e. every 60 seconds). , To provide a problem. This is too slow to adapt to an overload occurrence that requires a control update every five seconds.

In order to cope with a very high call rate and several times the time rate during the call (for example 20 times or more), the access controller according to the present invention implements high speed overload control. Specifically, overload control is externally imposed on the access controller by having each access point function as an external controller that implements local access constraints to limit traffic rates approved for the network. The present invention allows an access controller to determine an overload condition based on the number of calls that the access controller receives from a plurality of access points (generally not all access points in the area, but not necessarily all). The access controller then determines the global constraint (eg, per line) that is imposed to reduce the traffic rate admitted to the network. Each access point then changes the global constraints to determine the local constraints (eg, per access point) that it uses to regulate the rate of traffic that is granted to the network. In a preferred embodiment of the present invention, the access controller no longer actively responds to off-hook indications received from calls blocked by the access point granting the call gapping process. In another preferred embodiment, the access point does not need to send an off-hook to the access controller for a call that is not authorized by local constraints applied at the access point.

Advantageously, the granting of an initial randomized local gap interval applied by the access point without requiring the initiating call to be received by the access point allows the access controller to provide more sensitive overload control and to update the global traffic rate constraints. Allow the gun to be provided. Changing the global traffic rate by each access point is particularly advantageous in networks where the access point has different capacity (eg, configured to receive traffic along different line numbers). For example, the number of access points in an IP network where each access point includes a media gateway ranges from a single line MG (for example, there may be thousands of them) to a much larger capacity (eg, 16,000 lines). Ability to handle up to a single MG can vary greatly in configuration. In such IP networks, it is not possible to apply global constraints across the entire network because the gap gaps given may not be suitable for all of the various types of MGs.

As such, the present invention provides an adaptive overload control system that controls the amount of traffic handled by a network access controller, where the network access controller is configured to control a plurality of access points, each network access point being associated with the received traffic. Provide access to the communication network.

The steps performed to implement the overload control system are schematically illustrated in FIG. 5. In FIG. 5, the overload control system includes a controller that monitors and analyzes traffic requests received from all (or some) network access points in the control area (step 100), to determine if an overload condition exists (step 102). do. By determining if an overload condition is likely to occur from the level of traffic provided to the processing access controller, the central overload system can be implemented by an access controller that provides a consistent and faster response to the overload condition in the network. This central overload response is provided by the network access controller that multicasts one or more control messages to each network access point that functions as a source of traffic provided to the access controller (step 104). The at least one control message includes at least one globally determined traffic rate constraint. In some embodiments of the invention, one or more global constraints are determined by the access controller, for example if per-number constraints are imposed. In this embodiment of the present invention, general, basic, and constraints are imposed to prevent other traffic from occupying the network access controller with a processing request, and prevents calls to a specific address from occupying the network access controller. One or more other global constraints may be imposed for this purpose.

A plurality of network access points in the control area of the network access controller each receive the one or more multicast messages that include at least one global constraint. Depending on the type of multicast scheme employed, this may be all or a subset of the network access points in the control area. The received global constraint information is then processed by the receiving network access point to determine one or more local constraints (step 106). The set of at least one local constraint placed on the access point functioning as a traffic source for the communication network is determined by changing the global constraint according to one or more characteristics of the network access point. The local constraints are at least i) from the global constraint (s) duration of the first gap interval to grant (step 108), and ii) subsequent traffic requests (ie, subsequent access to the communication network). Determining a duration (step 110) of a subsequent gap interval traffic request to grant to the calls).

In an embodiment of the invention in which the communication network supports IP (or VoIP), global traffic rate constraints may be determined by the central destination MGC functioning as a controller for a set of MGs in the network. The MGC multicasts control messages using layer-2 and layer-3 mechanisms so that the MGC controller sends only one global constraint message that is copied on a network basis whenever a gap interval needs to be given or updated. Enable multicasting. The global constraint message is received by virtually no delay by all traffic sources (MGs) subscribing to the multi-cast group to quickly respond to an overload condition in the processing MGC implemented by all MGs receiving the global constraint message. To make it possible.

Each access point receiving a multicast global constraint message first performs a modified Crawford algorithm to modify the received global constraint information to receive traffic that the access point wants to access the network at a totally determined gap interval. Adjust the interval to better suit the traffic rate and number of lines. The initial local gap is then randomized in length to eliminate any synchronization effects that may occur in high volume scenarios with other network access points that may determine similar local gap intervals. More details on randomizing the gap spacing are described later.

Any suitable technique can be used to determine the duration of the initial gap interval, and the access point will operate effectively as if the gap of the subsequent (fixed period) time interval was given at some point in the past. This is shown schematically in FIG. 6. Figure 6 shows multiple network access points # 1-#N of the same type all implementing an initial gap that changes in a random manner, followed by subsequent gaps that all have the same spacing. The term gap interval here means the length of time after the approval of the call, during which no further calls are accepted. In fact, this simply delays the start of the next complete gap interval by varying amounts for all access points subscribing to the multicast issued by the access controller. The duration of the first gap interval may be determined using any suitable technique, such as, for example, a random or pseudo random technique, as long as the result removes synchronization of call blocking experienced at the access controller (e.g., at the MGC). Can be.

To determine global constraints, in one embodiment of the present invention, the access controller provides an estimate of the gap per line g interval and current rate r per line to each access point via multicast message (s). do. In the steady state, there is a probability (1 / (1 + r * g)) (for traffic with steady state and Poisson distribution) that the access point will not have an active gap interval. For an access point with an active gap interval, the initial gap is said to persist for an initial time period corresponding to the gap in the standard interval period reaching the previous point in time between zero and the gap interval G applied by the particular MG. Think of it (see Figure 6). Determines whether each access point should have an active gap interval first using a random number, and then how much of the gap interval remains for execution (evenly distributed in the range from 0 to G = g / L, where L Is the number of lines managed by the gateway), otherwise the synchronization effect that may occur on the provided call rate at the access controller will no longer occur. This can be seen in FIGS. 7 and 8. In Figures 7 and 8, global constraints are provided to 5,000 access points (e.g. MGs) by an access controller (e.g. MGCs). Each access point is configured to receive traffic along 25 lines passed through a communication network (e.g., VoIP).

As shown by the embodiment of the present invention in FIG. 8, it is assumed that the number of calls provided to one MG by a plurality of MGs is initially 800 calls per second, while the MGC has a target rate of about 100 calls per second. lets do it. Such call volume is typical in remote voting type scenarios. Thus, the MGC applies external overload control by multicasting the global call rate limit including the global gap interval parameter to several MGs in the control region. The MGs then apply the global constraint to their capacity to determine a local gap interval. The MG then determines the initial gap spacing in a random or pseudo-random fashion, respectively, and the initial gap spacing can be a range between zero and the maximum local gap spacing. Each MG then gives the initial gap interval (if not zero) without waiting for a call to receive, which substantially means that the initial gap interval is granted as soon as the MG determines its duration. This in turn results in a reduction in the number of calls requested by the MG to the MGC, which does not indicate the synchronization impact known to occur in such a scenario when conventional call mapping techniques are used. In Figure 7, MGC requests a limit at t = 100 seconds. That is, at t = 100, the global constraint is generated and passed to each of the MGs. The traffic authorized through each MG is then affected by local constraints almost immediately upon receipt of global constraint information from the MGC.

FIG. 8 shows the area indicated by FIG. 7 in more detail. In FIG. 8, the relative proportions of the types of calls carried by the MG to the MGC following the MG implementing the local gap condition are schematically illustrated. In short, the call subsequently delivered to the MG is i) the call delivered unaffected by the local gap condition when the initial gap interval 0 was previously applied; ii) a call transferred after the initial shortened local gap interval has been applied; iii) The normal constant length local gap spacing may be a call delivered after being applied by the MG. The monetary group in (iii) clearly constitutes a substantial proportion of the currency much later than the currencies of (i) or (ii).

In FIG. 8, the sum of all transferred currencies approved for MG is shown, indicated by a line with X representing each marked point. Initially, the number of calls authorized by the MG that did not actually give an initial gap interval (ie, the initial gap interval was zero) occupies most of the received calls. This is illustrated by the line indicated by <> with each marked point. However, over time, the line marked by 점 shows that the initial (shortened) local gap interval begins to account for the majority of the approved calls after expiration. As a result, as indicated by the line with Δ, the number of authorized calls after expiration of the normal (regular) local gap occupies most of the amount of calls delivered to the MGC by the MG.

9 and 10 illustrate another embodiment of the present invention, illustrating the effect of global constraints imposed by the MGC when two different groups of MGs are included in the control region. 9 and 10 show more complex aspects. Here 125,000 lines are connected to a 25 line gateway and 125,000 lines are connected to a single line gateway. The single line gateway doubles the rate of calls to 25 times compared to the result obtained for 25 line MG.

However, as Figures 7 to 10 clearly show, otherwise the synchronization effect that allows an access controller (e.g., MGC) to experience repetitive processing demand congestion has a gap gap that varies in a random manner between the access points. It is eliminated by the introduction of initial gaps. Apart from their differing periods, the initial gaps perform conventional gap functions differently. All calls arriving before expiration of the interval timer are rejected; If a call arrives after the interval timer, the call is accepted and a new gap timer is started.

When the gap interval update arrives after the start of local overload control implemented by the network access point, the control given almost immediately can be provided by ensuring that the above-described limiting process is automatically repeated when a second control message is received. Can be. One possible scheme for implementing an update is that the access controller periodically determines if the overload condition is still exceeded, and one or more global constraints from the aggregated traffic rate that is then being received from all network access points providing traffic to the access controller. It is to determine if an update in the condition (ie a new adjustable global gap interval) is needed.

If new global gap information is sent to each access point that contributes to the call rate measured by the access controller, the access point then determines a local gap interval that is less than the currently given gap interval, and any currently running interval timer. Is updated to reflect this change and the currently running interval is appropriately reduced (or set to zero). If the local gap interval is updated to a longer period as a result of the new global gap interval, the local interval timer is extended to increase any gap already granted. Or, in either case, the current running interval is stopped and a new initial local gap interval is defined between zero and the new gap interval period.

Another advantage of the present invention is that the implemented access control may allow a network access point (e.g., MG) to prioritize a call for a particular call, for example an emergency service (999, 911, etc.) , As well as to enable proper termination of the blocked call (eg to ensure that all relevant state devices reflect all changes in the on-hook condition). The present invention also allows the called address to be determined before the wpd is imposed by properly selecting the location of a global constraint message in the message flow between MGC and MG.

For example, referring to FIG. 11, an example of a message flow between MGC and MG is shown in accordance with one embodiment of the present invention. Although in the following embodiments explicit references are made to standards and MGCs relevant to the context of MG, those skilled in the art will note that alternative standards support similar message flow between different types of network access points and network access controllers. I can understand that.

Referring to FIG. 11, initially, a MG line-state-machine (LSM) is in an off-hook state when a call is initiated, and the off-hook then notifies the MGC. This notifies the MG of the message. If the H.248 protocol is supported, the off-hook status should be immediately notified by MGC to the MGC call processing application (CpApp), but it is not necessary to notify the off-hook status, If strict compliance is not required, this may be postponed or not implemented at all. If the MGC CpApp is notified, a message indicating the type of dial-tone to use will be returned to the MG via MGC with the details of any dial-plan supported. An alternative use of H.248 is that the MGC will send the dial-tone and dial-plan (for application to a specific physical termination, ie copper wire) to the MG prior to the call (eg when the required dial tone changes). To help. This eliminates the need for MGC to return something in response to the off-hook and grants the possibility of not sending the off-hook.

When a call proceeds, the MG LSM passes the called digits to the MG, which then implements the dial plan and the required gap according to the control constraints imposed by MGC / MGC CpAppl. We will perform the analysis. If the gap is being granted, the MG would like to convey this by, for example, generating an Equipment Engaged Tone to the call originator. If no gap is given, the number 1-n may be passed to the MGC.

In an alternative embodiment of the present invention, the MG does not receive confirmation for off-hook messages to reduce the per-call processing burden of the MGC. Instead, the MGC only provides details of the dial tone type (ie DT (n)) and details of whether the dial-plan should be implemented.

The dial plan means that after a certain number of numbers have been sent to the MG, the MG analyzes the dialed numbers to determine if the call complies with the criteria imposed by a dial-plan that implements a gap constraint. do. Alternatively, basic constraints may be imposed on certain numbers not supported by the dial-plan. If the call is determined to be exempt from the gap constraint, it is sent to the MGC, but otherwise the call is approved, i.e. if already assigned, so that no gap is granted until after the next call is received. If the interval has expired or the initial gap interval is zero, it is only sent to the MGC. This is shown schematically in FIG. 12 and shows the point at which the access controller notified three different types of calls.

In Fig. 12, the first call relates to the caller dialing 0800654321. This number is not recognized as related to the priority call after the first three digits (which prompts MCC early notice). However, the other two numbers shown in FIG. 12 relate to priority numbers 112 and 999 which prompt the MGC for early notification. If another number is dialed, for example 012345 67890, which is not supported by the dial-plan, the default (or any) constraint may be applied by the MG.

One way to determine the global gapping interval at the access controller is to determine the rate at which the controller is rejecting the off-hook provided by the signal source. The controller can multicast a single global gap per user, each traffic source then scales the global gap from a global gap size to a local gap size in proportion to its capacity, and then from zero to the local The initial gap period is arbitrarily given to the gap size. Each traffic source then applies the local gap to all non-priority calls, and the gap only affects the services provided for the normal call.

However, in another embodiment of the present invention, the controller performs similar adaptive internal control and measures the rate of rejecting off-hook messages received from the traffic source, but the controller depends on the called identities of the caller. Analysis of the reject rate is performed to determine which caller identity needs to be mapped. In this sense, call congestion includes network addresses and the like.

The controller then determines global constraints and per-number global constraints and multicasts these global constraints to each traffic source. The traffic source then adjusts the global constraints from the global gap information provided by the controller to determine which gap interval size should be applied according to the capacity of the traffic source. A random initial gap interval size is also determined. When a call is made, the dialed number is analyzed by the network traffic source to determine the gap interval most suitable for use for the particular dialed number scheme. As such, in FIG. 12, if a particular digit is recognized as being associated with a priority number, no gap is assigned (eg, 112 or 999 numbers), but the 0800 6543231 number is subject to normal blocking. However, in another embodiment of the present invention, the dial plan indicates that only televoting / bulk dialing numbers are likely to be given a gap, which allows the access controller to avoid overloading traffic traffic without overloading traffic processing. You will be able to make normal calls.

In this way, a VoIP network with a 16,666 cps in-time capacity can effectively handle non-bulk-dial related traffic, for example when more than 100,000 cps are transferred from a digital line exchange in a large dialing situation. Can be. In such an embodiment of the present invention, the control request of the MG-MGC simply maximizes the effective call handling capability in the overloaded MGC that depends on the MG response time (usually 200 ms or less), enabling both priority calls. And optionally to protect the ordinary (non-priority currency) from experiencing the congestion caused by televoting.

13 and 14 contrast the overall structure of the overload control with respect to the arrangement of the functions. In Figures 13 and 14, A is the acknowledgment function of MGC, and R is the MG local overload constraint. Restriction, U is the update function.

13 illustrates a MG-MGC scenario in a VoIP network where restriction updates are implemented. In Figures 13 and 14, the off-hook signal is sent to the MGC by the MG, the MGC applies admission control, and the MG applies appropriate constraints to limit the overload of the MGC. However, in FIG. 13, the MG updates the overload locally without any control by the MGC. This update may be based on one or more criteria. For example, the update may be based on an explicit rejection of a new call attempt by the MG to the MGC, or a long delay seen by the MG between sending a message to the MG and an acknowledgment of the MGC indicating an excessive burden on the MGC. Can be based on. The expert in the art uses only the MGC's response to the call to determine if the MGC is overloaded and applies its local constraints, so that the structure of FIG. 13 has only a single-line MG or several lines. It is understood that it is not appropriate for MG, and inevitably sends a lower rate call to MGC than a larger MG. If only low rate call events are sent, the MG has limited information based on estimating the presence and severity of overload in the MGC, and constraint application is slow. However, for large gateways, this embodiment of the present invention has the advantage of not requiring any change to the H.248 standard recommendation.

14 schematically shows MGC determining the updated conditions in contrast. This allows more sensitive overload control to be implemented. In Figure 14, when MG receives an off-hook condition, MG passes it to MGC and MG updates its global constraint. The MGC then determines if a new updated global constraint is created and returns the new global constraint (i.e. the new global gap interval) to the selection (or all MGs) using the appropriate communication technique (multicasting) so that all MGs Provide one or more control messages to receive. Each MG then processes the call using a modified local constraint that determines from the updated global constraint information provided by the MGC to determine if the call is to be rejected.

The MGC accesses the aggregated traffic flows it receives from each MG, receiving higher traffic rates than each MG receives individually, which is an accurate estimate of the traffic rate (overall and for a particular dialed number). To be obtained. Since the traffic rate received by the MGC is higher than the traffic rate in the individual MG, than the MG determines the local traffic rate and generates local constraints without being dependent on the MGC (as shown in FIG. 13). There may be more frequent updates to external overload control conditions implemented by each MG (received updated global constraints from the MGC).

This second control option has the advantage that a set of overloads can be avoided because there is an instantaneously valid gap with a randomized initial interval (multicasted by MCC). However, the H.248 standard recommendation would require modifications taking into account the global gap constraints returned by multicast messages from the MGC, and the present invention delivers global gap interval information to each MG that contributes to the aggregated traffic rate defined by the MGC. Can be implemented by other suitable techniques. It is up to each MG (although they can be determined centrally by the MGC) to modify the global constraints they receive to satisfy their characteristics, and implement local gap logic and appropriate dial-tones.

In a high volume call scenario, each network access point will receive multiple calls within a short time, all of which have the same target address (eg, the same number is very short and can be dialed by multiple calls in a nearly instantaneous time). has exist). Control for both access points can be implemented by the present invention relatively quickly to prevent the response time from rising to an unacceptable level as shown in FIG.

Thus, those skilled in the art will understand that the access controller determines the aggregated traffic rate for the traffic provided from each contributing network access point. This integration rate is higher than the rate at the individual access point, which means that even when external overload control has already been implemented, the access controller can more reliably determine whether the external overload control requires an update from the actually provided traffic rate. . Each network that receives control messages of local constraints (ie, local gap intervals) from the access controller derived from information generated by the access controller (which may include, for example, the aggregated traffic rate and the traffic rate per line). The grant by the access point may then be implemented more quickly than if the update conditions are simply determined by the network access point itself. By immediately giving an initial gap period that each access point determines itself (i.e., independent of the initial gap interval imparted by each of the other access points), the access points do not need to communicate with each other.

Although the means by which the control message is communicated to each network access point by the network access controller is preferably performed by a multicasting technique specific to the access points contributing to the traffic rate provided in the network access controller, Other suitable communication means known in the art may be used.

Where multicasting communication means are used by the access controller, the network may employ a communication multicasting process that enables the access controller to multicast global traffic access rate constraint (s) to the network access point via the communication network. Need to support Multicasting is desirable because it allows faster control to be given to the MG of the IP network, makes the MGC more sensitive, and updates the call constraints in less time than is known in the art. Adaptation of multicast call constraints globally determined by the access controller to satisfy an individual MG allows for more effective call restriction by each MG within the control area of the MGC.

One skilled in the art will note that if the underlying transport network has enough bandwidth to ensure that it is in place at most of the access points quickly enough for effective control of the overload, then the constraint message for each network access point is determined. It will also be appreciated that unicast transmission may be used instead of multicast. This removes any dependence on slow unicast deployments to randomize the onset of gap intervals, either because traffic constraints are dull (if constraint deployment is slow) or synchronous (if constraint deployment is fast). do.

The following summary is included herein.

An adaptive overload system that controls the amount of traffic handled by a network access controller is described with respect to a network access controller configured to control a plurality of network access points. Each network access point provides traffic to the communication network and the system determines at the network access controller whether there is an overload, and if so, at least to limit the rate at which the network access point grants the traffic to the communication network. Creating a global constraint. The controller then multicasts at least one global traffic constraint to one or more of the plurality of network access points. Each network access point receiving the global constraint then processes the global traffic constraint to determine a plurality of local constraints. The receiving network access point to determine the local constraint: determining a predetermined local gap interval to be granted to the traffic; And determining an initial gap interval that is different between each of the plurality of network access points and that is different from a subsequent predetermined local gap interval. The initial gap interval is determined in a random or pseudo-random manner to ensure that the synchronization effects at the network access controller that would otherwise occur in high call rate scenarios are eliminated.

While the foregoing embodiments have been described with reference to changing the Crawford algorithm to determine a local gap interval, those skilled in the art will recognize that other techniques that may be applied by the network access point to limit the traffic rate provided to the controller. It will be appreciated that the constraint algorithms may be modified as appropriate to implement alternative constraints in other embodiments of the present invention.

Claims (38)

An adaptive overload control method for controlling the amount of traffic provided to a network access controller by a plurality of network access points for network admission processing, the method comprising: The plurality of network access points provide traffic to a communication network under control of the network access controller, The method allows the network access controller to externally control the amount of traffic handled by the network access controller by regulating the traffic rate provided by the plurality of network access points, The method executed in the network access controller, Determining whether an overload condition exists in the network access controller, If the overload condition exists, Creating one or more global traffic constraints including a global gap interval to limit the rate at which a network access point grants the traffic to the communication network, and Passing the one or more global traffic constraints to all or a portion of the plurality of network access points, The method executed at each network access point receiving the one or more global traffic constraints, Processing the received global traffic constraints to determine a plurality of local gap interval constraint conditions for each network access point, The determining of the plurality of local gap gap constraint conditions may include: Determining a local gap interval [Delta] t imposed on the traffic received by each network access point, and Determining an initial local gap interval Δt _{0 that} is different from the determined local gap interval Δt, The local gap spacing DELTA t is determined by scaling the global gap spacing depending on the capacity of each network access point, Each of the initial local gap intervals Δt ₀ is independently determined by each network access point receiving the one or more global traffic constraints, The duration of the initial local gap interval Δt ₀ has a range from 0 to a local gap interval Δt determined for each network access point, The method executed at each of the network access points further includes: Instead of waiting for traffic to be received at each network access point, and in response to each network access point receiving the global traffic constraint, each network access point has a duration of the initial local gap interval Δt ₀ . And upon determining, determining the respective initial local gap interval DELTA t ₀ at each of the plurality of network access points receiving the one or more global traffic constraints. . The method of claim 1, Passing the one or more global traffic constraints to all or a portion of the plurality of network access points, wherein the one or more global traffic constraints are multicast to all or a portion of the plurality of network access points. Adaptive overload control method. 3. The method according to claim 1 or 2, And the initial local gap spacing (Δt ₀ ) is determined using a random or pseudo-random technique at each network access point. 3. The method according to claim 1 or 2, The communication network is a VoIP network, Adaptive traffic control method, characterized in that the call-related traffic (call-related) traffic. 3. The method according to claim 1 or 2, The network access controller is a media gateway controller (MGC) and each of the plurality of network access points comprises a media gateway (MG). 3. The method according to claim 1 or 2, And a global traffic rate constraint is determined for said address by said network access controller. 3. The method according to claim 1 or 2, Each local gap interval DELTA t is the number of lines on which a network access point receives traffic for transmission over the communication network, and the network access by all network access points providing traffic to the network access controller. Adaptive overload control method, characterized by the adjustable gap interval determined by the network access controller based on integrated traffic provided to the controller. The method of claim 2, And said multicasting communicates with all of said network access points controlled by said network access controller. 3. The method according to claim 1 or 2, And wherein said network access controller determines said at least one global traffic constraint by analyzing a rate at which an off-hook message is rejected by said network access controller. 3. The method according to claim 1 or 2, And wherein said network access controller analyzes a rate at which traffic is provided to said network access controller by said plurality of network access points to determine said one or more global traffic constraints. 3. The method according to claim 1 or 2, And said network access controller determines said one or more global traffic constraints by analyzing a rate at which traffic is rejected by said network access controller. 3. The method according to claim 1 or 2, A dial-plan is implemented by the network access point to make it unnecessary to send an off-hook status message to the network access controller when a local gap interval (Δt) constraint is imposed. Adaptive overload control method. 3. The method according to claim 1 or 2, And the local gap spacing Δt has a size in minutes. 3. The method according to claim 1 or 2, Wherein each network access point determines the initial local gap interval Δt ₀ using a probabilistic method. 3. The method according to claim 1 or 2, The to the initial local gap interval (△ t ₀₎ if a non-zero initial local gap interval (△ t ₀₎ is 0, the local gap interval (△ t) which is determined by each network access point from all of the network's access point Adaptive overload control method, characterized in that determined by each network access point to be uniformly distributed in a range. An adaptive overload control system for controlling the amount of traffic provided to a network access controller by a plurality of network access points for network admission processing, The plurality of network access points provide access to a communication network for the traffic under the control of the network access controller, The system allows the network access controller to externally control the amount of traffic handled by the network access controller by regulating the traffic rate provided by the plurality of network access points, The network access controller, Means for determining an overload condition of the network access controller, One or more global traffic constraints, including a global gap interval, to limit the rate at which a network access point grants the traffic to the communication network when the overload condition of the network access controller is determined by the determining means. Means for generating a condition, and Means for passing the one or more global traffic constraints to all or a portion of the plurality of network access points, Each of the network access points receiving the one or more global traffic constraints, Means for processing the received global traffic constraints to determine a plurality of local gap interval constraint conditions for each network access point; Means for determining a local gap interval [Delta] t imposed on the traffic received by each network access point, and Means for determining an initial local gap interval Δt _{0 that} is different from the determined local gap interval Δt, The local gap spacing DELTA t is determined by adjusting the global gap spacing depending on the capacity of each network access point, Each of the initial local gap intervals Δt ₀ is independently determined by each of the plurality of network access points, and the duration of the initial local gap interval Δt ₀ is from 0 to the respective network accesses. Has a range up to the local gap spacing (Δt) determined for the point, Each of the network access points further comprises: Instead of waiting for traffic to be received at each network access point, and in response to each network access point receiving the global traffic constraint, each network access point has a duration of the initial local gap interval Δt ₀ . And upon determining, means for imposing a respective initial local gap interval DELTA t ₀ at each of the plurality of network access points receiving the one or more global traffic constraints. A network access controller that receives traffic by a plurality of network access points for network admission processing and controls the amount of traffic provided, The plurality of network access points provide, under the control of the network access controller, access to the communication network for the traffic, and the network access controller regulates the traffic rates provided by the plurality of network access points. Externally control the amount of traffic handled by the controller, The network access controller, Means for determining an overload condition of the network access controller, One or more global traffic constraints, including a global gap interval, to limit the rate at which a network access point grants the traffic to the communication network when the overload condition of the network access controller is determined by the determining means. Means for generating a condition, and Means for passing the one or more global traffic constraints to all or a portion of the plurality of network access points, And wherein the global gap interval is adjusted depending on the capacity of the forwarded network access point to generate a local gap gap (Δt) at the forwarded network access point. A network access point that, under the control of a network access controller, provides access to a communication network for traffic and controls the amount of traffic provided to the network access controller for network authorization processing, The network access controller forwards one or more global traffic constraints to the network access point under overload to limit the rate at which the network access point grants the traffic to the communication network, thereby processing by the network access controller. Externally control the amount of traffic Means for receiving the one or more global traffic constraints, Means for processing the received global traffic constraint to determine one or more local gap interval constraint conditions for the network access point; Means for determining a local gap interval [Delta] t imposed on the traffic received by the network access point, and Means for determining an initial local gap interval Δt _{0 that} is different from the determined local gap interval Δt, The local gap spacing DELTA t is determined by adjusting the global gap spacing depending on the capacity of the network access point, The duration of the initial local gap interval Δt ₀ has a range from 0 to a local gap interval Δt determined for the network access point, The network access point may further include In response to receiving the global traffic constraint without waiting for traffic to be received, the initial local gap interval Δt ₀ as soon as the network access point determines the duration of the initial local gap interval Δt _0. Means for imposing a network access point. delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete

Images