NETWORK

Field of the Invention

The present invention relates to data communication. In particular, this invention relates to controlling Digital Subscriber Line, DSL, connections.

Background to the Invention

Digital Subscriber Lines, DSL, are a common technology for providing digital communication over existing twisted copper pair subscriber lines. The subscriber line extends between two DSL modems. A first DSL modem is typically located in the customer's premises, and the second modem may be located at the local exchange (known as the 'central office' in US terminology), a street cabinet, or distribution point (sometimes known as 'drop point'). Typically, the local exchange, street cabinet or distribution point includes a DSL Access Multiplexer, DSLAM (a form of aggregation transceiver device) comprising several DSL modems (one for each subscriber line).

The DSLAM (at the exchange, cabinet or distribution point) connects the first DSL modem at the customer's premises to the Core Network.

There are various Network Management Systems within the Core Network to manage and control the various subscriber lines. One form of Network Management System is a Dynamic Line Management, DLM, engine for controlling DSL connections. In current DSL deployments, the DLM engine in the Core Network receives data periodically from a DSLAM relating to all DSL connections to that DSLAM. This data is usually sent at intervals (e.g. once a day) as a bulk data file and includes, for example, the number of retrains or number of errors per 15 minute bin since the last data upload. The DLM engine receives this data from all its connected DSLAMs, and analyses the data to determine if any actions need to be taken. For example, if one subscriber line has retrained many times, a line profile with a more conservative rate cap may be applied. The present inventors have identified an issue with the current DLM technique. That is, the DLM engine requires a bulk data file to be uploaded (typically to the Operations Support System, OSS) and processed periodically. As more DSL connections are established in the network, and/or more data points for each line are added, the bulk data file becomes too large to upload and process in a small time frame. This makes the DLM engine slow to react to any issues on the network. Furthermore, trying to process bulk data for an ever increasing number of lines and/or data points is leading to overload of the systems in the OSS stack. Also, the integrity of the data collection process varies day-to-day as many lines may have incomplete aggregated data due to failures in data collection and processing.

International Patent Application Publication Number 2012/015817 discloses a DLM engine which responds to a disruption on the line (e.g. an increase in errors) by determining and applying an appropriate control setting at the PHY layer which will achieve a known desired effect on a higher-layer application.

It is therefore desirable to alleviate some or all of the above problems.

Summary of the Invention

According to a first aspect of the invention, there is provided a method of controlling an

Access Network including a plurality of Digital Subscriber Line, DSL, connections, the method comprising the steps of: receiving data relating to an event on a DSL connection, the data indicating a factor contributing to the event; and a Dynamic Line Management, DLM, engine, residing in the Access Network, responding to the event, wherein the DLM engine response is dependent on the factor contributing to the event.

The present invention provides an improved DLM architecture which allows the DLM engine to react to events on a DSL connection by taking into account a factor contributing to the event. By reacting to the factor contributing to the event, the DLM engine may respond in a more appropriate manner and forestall such events occurring in the future.

Furthermore, the DLM engine resides in the Access Network, contrary to the Core Network deployment of the prior art. This allows the DLM engine to receive the data relating to an event (including the factor contributing to the event) and react to it in realtime. The centralized Core Network DLM engine of the prior art relies on data being uploaded in bulk by multiple DSLAMs, which would be too inefficient if modified to contain the factor or factors contributing to the event. Accordingly, a distributed architecture is used, which, whilst a departure from the prior art technique, allows the DLM engine to focus on a smaller number of DSL connections and react to events occurring on them in real-time.

Distributing the DLM engine throughout the Access Network has several other benefits. Firstly, there is no longer a single point of failure as experienced by the Core Network deployment of the prior art. Secondly, distribution of multiple DLM engines means there are fewer lines associated with each DLM engine and new DLM engine algorithms can be tested and implemented in a controlled rollout. There are further advantages. Currently, when the firmware of a DSLAM is updated

(e.g. as a standard is updated), the DLM engine in the Core Network deployment needs to know which DSLAMs have been updated and which have not. This is burdensome for the Network Operator. However, with the distributed DLM architecture of the present invention, the DLM engine can be updated in conjunction with the associated DSLAM's firmware update.

After the step of receiving data relating to an event on the DSL connection, the method may further comprise the steps of: storing the received data in memory in a proxy management unit; and forwarding the data stored in said memory in the proxy management unit to the DLM engine. Thus, the data may be received and stored in a proxy management unit (also known as a Persistent Management Agent). In G.fast deployments, this allows the data to be stored on a unit with a more reliable power supply, which may be used as a backup or proxy in the event a Distribution Point Unit, DPU (which are reverse powered) loses power. The present invention may therefore co-operate with such a proxy management unit in a dedicated node in the Access

Network, and react to the data once forwarded to the DLM engine.

After the step of storing the received data in said memory, the method may further comprise the step of: classifying the data stored in said memory as either an event related to the stability of the DSL connection or an event related to the speed of the

DSL connection. Furthermore, the data stored in the data buffer may only be forwarded to the DLM engine if it is classified as an event related to the stability of the DSL connection. The DLM engine may therefore only react to the event if it is classified as one related to the stability of the DSL connection. Alternatively, the event may be speed related, and the event may be forwarded to an ANCP relay such that an Access Network Control Protocol, ANCP, message may be sent to a Communications Provider Broadband Remote Access Server, CP BRAS, associated with the DSL connection. The DLM engine may respond to the event after the event has occurred a predetermined number of times. This response may be after the event has occurred a predetermined number of times in a predetermined time period, and/or may be after the event has occurred a predetermined number of times for the same contributing factor. The DLM engine may respond to the event using an On Line Reconfiguration, OLR, process.

The method may further comprise the initial step of gathering data on the DSL connection, wherein the data is gathered using the data gathering function defined in ITU-T Recommendation G.993.2 (201 1 ) Amendment 3.

The DSL connection may be a VDSL, VDSL2 or G.fast connection. The DLM engine may reside in a dedicated node in the Access Network or collocated with the DSLAM (e.g. in the street cabinet).

According to a second aspect of the invention, there is provided an Access Network node, the Access Network including a plurality of Digital Subscriber Line, DSL, connections, the node comprising a receiver adapted to receive data relating to an event on a DSL connection, the data indicating a factor contributing to the event; and a Dynamic Line Management, DLM, engine adapted to respond to the event, wherein the

DLM engine response is dependent on the factor contributing to the event.

The node may further comprise a proxy management unit including a receiver adapted to receive the data relating to the event; memory adapted to store the received data; and a transmitter adapted to transmit the data relating to the event to the DLM engine.

The node may further comprise a processor adapted to classify the data stored in said memory, wherein the data stored in said memory is only forwarded to the DLM engine by the transmitter if it is classified as an event related to the stability of the DSL connection. The node may further comprise a processor adapted to classify the data stored in said memory, wherein the data stored in said memory is classified as either an event related to the stability of the DSL connection or an event related to the speed of the DSL connection.

The node may further comprise a counter adapted to count events on the DSL connection, wherein the DLM engine may be adapted to respond to the event when the counter reaches a predetermined threshold. The counter may be adapted to count events on the DSL connection having the same contributing factor, and/or the DSL engine may be adapted to respond to the event when the counter reaches a predetermined threshold in a predetermined time period.

The DLM engine may respond to the event using an On Line Reconfiguration, OLR, process.

The DSL connection may be a G.fast connection and the receiver may be adapted to receive data relating to the event on a DSL connection from a Distribution Point Unit, DPU.

The DSL connection may be a VDSL connection and the receiver may be adapted to receive data relating to the event on the DSL connection from a street cabinet. Alternatively, the DSL connection may be a VDSL connection and the node may reside in a street cabinet.

Brief Description of the Figures

In order that the present invention may be better understood, embodiments thereof will now be described, by way of example only, with reference to the accompanying drawings in which:

Figure 1 is a schematic diagram illustrating a first embodiment of the present invention, including a first and second Customer Premises Equipment, a Distribution Point Unit, and a head-end node;

Figure 2 is a schematic diagram illustrating the elements of the CPE, DPU and headend node of Figure 1 ; Figure 3 is a flow diagram illustrating a first embodiment of a method of the present invention;

Figure 4 is a schematic diagram illustrating a second embodiment of the present invention, including a first and second Customer Premises Equipment, a street cabinet, and a head-end node; and

Figure 5 is a schematic diagram illustrating the CPE, street cabinet and head-end node of Figure 4.

Detailed Description of Embodiments

A first embodiment of the present invention will now be described with reference to

Figures 1 to 2. Figure 1 shows a Fibre to the Distribution Point (FTTDP) optical network 1 . In this embodiment, the optical network is a Passive Optical Network (PON). A distribution drop point unit 20 (DPU) is provided which is located on a pole proximal to a plurality of subscriber premises. The DPU 20 is connected to a first and second Customer Premises Equipment (CPE) 10, 1 1 via first and second overhead drop wires 12, 13 respectively. The DPU 20 is further connected to a head-end node 30 via an optical connection. The DPU 20 therefore contains an Optical Line Terminal (OLT) for converting between the electrical signals which may be transmitted over the overhead drop wires 12, 13 and optical signals which may be transmitted to/from the optical head-end unit 30. The DPU 10 also acts as a Digital Subscriber Line Access

Multiplexer (DSLAM) to aggregate data for multiple subscriber lines (in this example, for CPEs 10, 1 1 ) onto a single high-capacity uplink to the head-end node 30.

In this embodiment of the invention, the FTTDP deployment is a form of G.fast connection (however, the skilled person will understand that the DPU may also support a VDSL connection). In other forms, the DPU may be mounted on a pedestal, in a footway box, or an underground chamber connected to the CPE via either overhead or underground wires. The DPU 20 is reverse powered via the overhead drop wires 12, 13, for example by a reverse powering scheme as disclosed in Applicant's European patent application number 13250085.1 (Agent's Ref: A32534).

The head-end node 30, which will be described in more detail below, resides in the Access Network and connects the DPU 20 to the Internet 3, Network Management System (NMS) 5, Operation Support System (OSS) 7 and Communications Provider Broadband Remote Access Server (CP BRAS) 9. Figure 2 is a schematic diagram illustrating the various features of the first CPE 10, DPU 20 and head-end node 30 (the skilled person will understand that a similar architecture exists for the second CPE 1 1 ). The CPE 10 and DPU 20 both include a G.fast transceiver 1 1 , 21 for sending and receiving signals over drop wire 12 (thus forming a G.fast connection). Both G.fast transceivers 1 1 , 21 cooperate with a processing module 13, 23. Both processing modules 13, 23 include a performance statistics module 13a, 23a configured to monitor the transceivers 1 1 , 21 and record data relating to the connection (e.g. in memory or in a data buffer). The two performance statistics modules 13a, 23a communicate via an EOC channel (the logical connection is shown in dotted lines in Figure 2), which is physically connected over the G.fast connection.

The CPE 10 also includes a customer port/Layer 2 switch.

The processing modules 13, 23 for the CPE 10 and DPU 20 are both configured to implement control plane functions, which will be discussed in more detail below.

The DPU 20 processing module 23 further includes a data gathering module 23b. In general terms, the data gathering module is configured to monitor the G.fast connection (e.g. using data from the performance statistics module 23a) and record data relating to any event occurring on the connection (e.g. in memory or in a data buffer). As explained in more detail below, the data not only specifies what event occurred (e.g. a retrain), but also specifies what factor contributed to the event (e.g. a retrain due to a loss of margin).

The DPU 20 transceiver 21 and processing module 23 are both connected to a Layer 2 switch/backhaul module 25, which is connected via an optical fibre connection 21 to a corresponding Layer 2/backhaul module 35 in the head-end node 30. The head-end node 30 is therefore configured to receive backhaul data from the DPU 20 via the optical fibre connection 21 , and in addition is configured to receive data from the data gathering module 23b. This data is routed to a processing module 33 (the logical connection is shown in dotted lines), which is generally configured to classify the data and instruct an appropriate entity depending on its classification. The head-end node 30 further includes a DLM engine 31 , an Access Node Control Protocol (ANCP) Relay 37 and Persistent Management Agent (PMA) 39. The skilled person will understand that moving the DLM engine 31 into the Access Network is a departure from the prior art methods of implementing the DLM engine in the Core Network as part of the Network Management System. The benefits of distributing the DLM engine 31 throughout the Access Network will become clear upon review of the following description.

To aid understanding of these features and their interaction, a method of the present invention will now be described with reference to Figure 3.

As a first step, an event occurs on the G.fast connection. In one example, the event occurs at the first CPE 10 (step S1 .1 ), such that the performance statistics module 13a of the first CPE 10 signals the occurrence of this event and sends its related data to the DPU 20 performance statistics module 23a via the EOC channel of the G.fast connection. In another example, an event occurs at the DPU 20 (step S1 .2). In both examples, the event is recorded via the data gathering module 23b of the DPU 20 (step S2). The data gathering module 23b records the occurrence of the event and a factor (or factors) contributing to the event. In this embodiment, the data gathering module 23b implements the "Data gathering function" as defined in ITU-T Recommendation G.993.2 Amendment 3, which creates a record for each occurrence of an event, with each record entered in a data buffer with an event timestamp, event identifier, and additional event data defining a factor or factors contributing to the event. The data gathering module 23b forwards this data to the head-end node 30 via the Layer 2 switch/backhaul connection (step S3).

The PMA 39 is configured to receive the data from the data gathering function and store it in a data buffer. In this embodiment, and as explained in more detail in Applicant's UK patent application number 1316053.6 (Agent's Ref: A32585), the PMA 39 stores metric data relating to the DPU 20. As the head-end node 30 and PMA 49 have a more reliable power supply than the DPU 20 (which is reverse powered), the

PMA 39 may act as a back-up or proxy for critical metric data in the event the DPU 20 loses power.

As shown in step S4, the processing module 33 is further configured to poll the data buffer of the PMA 39 to retrieve the data created by the data gathering module 23b. In this embodiment, the processing module 33 and PMA 39 therefore include a Q management interface, as described in ITU-T Recommendation G.997.1 Amendment 2, such that the processing module 33 may poll the PMA's 39 data buffer every N seconds using the "EVENT TRACE BUFFER O" and "EVENT TRACE BUFFER R" functions (for events occurring at the DPU and CPE respectively). The value of N may be between 1 second (the minimum granularity of reporting) and the maximum buffer length related to "LOGGING_DEPTH_0"/"LOGGING_DEPTH_R" parameters from the above ITU-T Recommendation. The processing module 33 therefore receives the data created by the data gathering module 23b. Generally, this data relates to events occurring on the G.fast connection between the transceivers 21 , 1 1 at the DPU 20 and CPE 10, and includes at least one factor contributing to each event. In the next step of the method of this embodiment (step S5), the processing module 33 categorizes the event as either 'speed' or 'stability' related. This classification allows the processor 33 to determine which entity should act in response to the event.

In step S6.1 , the processing module 33 classifies the event as stability related and thus determines that the DLM engine 31 should act in response to the event. In this embodiment, the DLM engine 31 resides in the head-end node 30 and thus in the

Access Network, allowing it to act in real-time in response to any events reported by the data gathering module 23b (contrary to the prior art technique, which only allowed periodic updates of bulk data files to a DLM engine in the Core Network). The event data is therefore forwarded to the DLM engine 31 .

The DLM engine 31 is configured to act on the G.fast connection in response to the event, wherein the action takes into account the factor or factors contributing to the event. In this embodiment, the DLM engine 31 is configured to act on the G.fast connection via the control plane functionality implemented by the PMA 39 and the processing modules 13, 23 of the DPU 20 and CPE 10.

The skilled person will understand that the DLM engine 31 is highly configurable and each type of event and factor contributing to the occurrence of the event may have one of several appropriate responses. An example of the DLM engine 31 reacting to an 'End of Showtime' event will now be described. In this example, the DLM engine 31 receives data relating to an 'End of Showtime' event from the processing module 33. The data indicates the event identifier classifying it as 'End of Showtime', a timestamp for the event, and an indication of a factor contributing to the event. In this example, the factor contributing to the event is persistent severely errored seconds (i.e. several consecutive severely errored seconds).

In response, the DLM engine 31 responds to the event by acting on the line. The action taken is in dependence on the factor contributing to the event. Accordingly, in this example, the DLM engine 31 responds to the 'End of Showtime' event caused by persistent severely errored seconds by increasing the level of error correction on the G.fast connection. This action is taken via the PMA 39 and control plane interface to the DPU 20 and CPE 10. For example, the action may utilize the On Line Reconfiguration (OLR) functionality to change configuration without causing a retrain.

If a retrain is required, the PMA 39 could wait for the G.fast line to enter a low power state or non-traffic state before executing the change.

Thus, the skilled person will understand that the DLM engine 31 of the present invention is able to react to the event in a real-time manner, and the action taken in response to the event is specific to the cause of the event and thus will help forestall such events occurring in the future.

In another example, the DLM engine 31 receives data relating to an 'End of Showtime' event from the processing module 33, but the data indicates that the factor contributing to the event was due to a loss of margin. In response, the DLM engine 31 moves the G.fast connection to a more conservative line profile having a reduced line rate, which is applied to the connection via the PMA 39 and control plane interface to the DPU 20 and CPE 10.

The above two examples illustrate two ways the DLM engine 31 may respond to an 'End of Showtime' event in different manners depending on the factor contributing to the event. However, the skilled person will understand that there are many different types of event which may occur on a G.fast connection which may then be forwarded to the DLM engine 31 for action. In this embodiment, the data gathering module 23b classifies events as per ITU-T Recommendation G993.2 Amendment 3 (Table 1 1 -43), which includes the following events which may be forwarded to the DLM engine 31 : End of Showtime, Previous End of Showtime, Failed Initialization, Successful Initialization, Line Failure, CRC-8 Anomalies, OLR, Bitswap, Retransmission Event or Defect. The above ITU-T Recommendation further specifies the factors contributing to the event which may be recorded by the data gathering module 23b, and, by using the method of the present invention, may ultimately be acted upon in response by the DLM engine 31 . In the above examples, the DLM engine 31 acts in response to a single event.

However, as shown in Figure 3 and steps S6.2 to 6.3, the DLM engine 31 includes a counter which is configured to count the number of events caused by a particular factor in a predetermined time period, and react to the event (wherein the reaction takes into account the factor contributing to the event) if the number of events caused by a particular factor in a predetermined time period exceeds a threshold.

As noted above, the processing module 33 may instead classify the event as 'speed' related. In this case, and as shown in steps S7.1 to S7.3 of Figure 3, the processing module 33 forwards the event to the ANCP Relay 37. The function of the ANCP Relay 37 (as described in more detail in Applicant's UK patent application number 1316054.4,

Agent's Ref: A32571 ) is to identify the Communications Provider serving the G.fast connection and route the ANCP notification to the corresponding CP BRAS 9. This allows the DPU 20 to maintain a single connection with the head-end node 30 rather than multiple connections between the DPU 20 and the corresponding BRAS for all Communications Providers serving each subscriber line connected to the DPU 20.

The ANCP Relay 37 is therefore configured to act in response to the event by identifying the Communications Provider serving the G.fast connection, generating an ANCP message, and sending the ANCP message to the identified Communications Provider.

The events that may be classified as a 'speed' event and are therefore forwarded to the ANCP Relay are also covered in ITU-T Recommendation G993.2 Amendment 3 (Table 1 1 -43), and include: Downstream Initialization Net Data Rate, Upstream Initialization Net Data Rate, Downstream Net Data Rate After Successful SRA, Upstream Net Data Rate After Successful SRA, Downstream Net Data Rate After Successful SOS, Upstream Net Data Rate After Successful SOS.

In an example, the processing module 33 may receive a Downstream Initialization Net Data Rate event indicating that the G.fast connection has initialized at a low data rate.

The processing module 33 classifies this as a 'speed' event, and forwards the event to the ANCP Relay 37. The ANCP Relay 37 identifies the Communications Provider serving the G.fast connection, and generates and forwards to the Communications Provider an ANCP message indicating the new downstream data rate. This allows the CP BRAS 9 to make the necessary changes in response to the new downstream data rate (e.g. throttle back any traffic for that G.fast connection according to the new rate).

A second embodiment of the present invention will now be described with reference to Figures 4 and 5. Figure 4 illustrates a Fibre To The Cabinet (FTTC) deployment of the second embodiment. A street cabinet 120 is provided, which interconnects a first and second Customer Premises Equipment (CPE) 1 10, 1 1 1 to a head-end node 130. The street cabinet 120 connects to the CPEs 1 10, 1 1 1 via twisted copper pair wires 1 12, 1 13 and to the head-end node 130 via an optical fibre connection 121 . The street cabinet 120 therefore contains an Optical Line Terminal (OLT) for converting between the electrical signals which may be transmitted over the twisted copper pair wires 1 12,

1 13 and optical signals which may be transmitted to/from the optical head-end unit 130 over the optical fibre connection 121 . The street cabinet 120 also acts as a Digital Subscriber Line Access Multiplexer (DSLAM) to aggregate data for multiple subscriber lines (in this example, for CPEs 1 10, 1 1 1 ) onto a single high-capacity uplink to the head-end node 130.

In this embodiment, the DSL deployment is a VDSL connection. However, the skilled person will understand that VDSL may also be deployed at either the distribution drop point or exchange. Furthermore, the skilled person will understand that the street cabinet setup of this embodiment may use other DSL technologies, such as ADSL,

ADSL 2, ADSL 2+, etc.

A schematic diagram of the first CPE 1 10, street cabinet 120 and head-end node 130 is shown in Figure 5 (and the skilled person will understand that the second CPE 1 1 1 uses a similar architecture). As shown in the diagram, the first CPE 1 10 and street cabinet 120 include a VDSL transceiver 1 1 1 , 121 maintaining a VDSL connection over wire 1 12. In a similar fashion to the first embodiment described above, the CPE 1 10 also includes a processing module 1 13 having a performance statistics module 1 13a. Similarly, the street cabinet 120 includes corresponding elements to the DPU 20 of the first embodiment, and therefore includes a processing module 123 having a performance statistics module 123a and a data gathering module 123b, and a Layer 2 switch/backhaul 125 terminating an optical fibre connection 121 to a corresponding Layer 2 switch/backhaul 135 in the head-end node 130. The head-end node 130 is similar to that of the first embodiment, and thus also includes a DLM engine 131 , processing module 133 and ANCP Relay 137. The headend node 130 also connects the street cabinet 120 to the Internet 103, NMS 105, OSS 107 and CP BRAS 109. In this example, the head-end node 130 also includes a PMA 139. However, the skilled person will understand that the PMA 139 is not essential in this VDSL deployment as there would be local or forward power. Thus, the DLM engine 131 may instead act directly on the DSLAM (e.g. via the Q interface).

The skilled person will therefore realise that the VDSL deployment of the second embodiment may be controlled in the same manner as the G.fast deployment of the first embodiment. Accordingly, an event occurring on the first or second CPE 1 10, 120 or the street cabinet may be recorded by the data gathering module 123b, together with a factor or factors contributing to the event, which is then forwarded to the PMA 139 of the head-end node 130. The processing module 133 of the head-end node 130 may then poll the data buffers of the PMA 139 (or DSLAM processor 123) every N seconds to determine the recent events occurring on the VDSL connection, and classify these events as 'speed' or 'stability' related. All stability related events may then be forwarded to the DLM engine 131 , which may then respond to the event in dependence on the factor or factors contributing to the event (immediately or if the number of events in a given period of time reaches a threshold). Alternatively, all speed related events may be forwarded to the ANCP Relay 137, which forwards an ANCP message to the appropriate CP BRAS 109 in response.

The skilled person will therefore understand that the present invention provides an improved Dynamic Line Management system which is distributed in or about the Access Network such that it may react to events occurring on DSL connections in a real time manner. In the prior art, the events were recorded and a bulk file was uploaded periodically to a DLM engine residing in the Core Network, which would then analyse the data and make any necessary changes. However, these bulk data files would only include basic information regarding the event (rather than indicating why a particular event occurred) as the bulk data files would be too large to upload and process at the DLM engine in the Core Network. However, by distributing the DLM engine as in the present invention, the DLM engine is able to act on the events in real time, and react according to the cause of the event.

The skilled person will also understand that it is not essential that the data gathering modules of the embodiments above implement the "Data gathering function" of ITU-T Recommendation 993.2 Amendment 3. That is, the data gathering module may implement any suitable function to record data relating to an event and a factor or factors contributing to the event.

The skilled person will also understand that it is not essential for the elements of the head-end node to be placed in a dedicated entity in the Access Network. That is, these elements may instead be collocated with the DSLAMs of the DPUs or street cabinet. However, especially in the case of the DPU which is only reverse powered, it is beneficial to separate these functional elements and place them in a dedicated module which may have a more reliable power supply. Furthermore, the elements may be distributed about one or several entities in the Access Network, and each process step mentioned above may be implemented by one or more processing modules. The skilled person will also understand that the ANCP Relay is a non-essential part of the present invention. However, the head-end node provides a suitable entity to perform the ANCP Relay function, such that all speed related events can be forwarded to it for generating the appropriate ANCP messages. The above two embodiments relate to a G.fast and VDSL deployment of the present invention. However, the skilled person will understand that the present invention is applicable to any form of DSL connection by implementing the functions of the headend node into the Access Network (either in a dedicated node or distributed over several nodes in the Access Network). Accordingly, the present invention is also applicable to ADSL, ADSL2, ADSL2+, VDSL2 and any other current and future version of xDSL technology.

The skilled person will understand that any combination of features is possible within the scope of the invention, as claimed.