Methods for analyzing wireless LAN

Abstract

The present invention provides a method for analyzing a wireless local area network, wherein the wireless local area network includes multiple devices, wherein at least one device among the multiple devices is configured to change its media access control (MAC) address, the method comprising the following steps: obtaining data including multiple records, wherein each record among the multiple records relates to a device among the multiple devices and includes: a device tag, a MAC address and a performance metric set; for each device tag among the multiple records, determining a uniqueness value of the device tag based on a count of records having the device tag among the multiple records; processing each record among the multiple records to assign a unique identifier to the record by determining that the MAC address of the record is a local MAC address and that the uniqueness value of the device tag meets a uniqueness threshold; and analyzing the performance metric set of multiple records having the same unique identifier.

Description

Technical Field

[0001] The present invention relates to telecommunication networks. Background Art

[0002] Customer premises equipment (CPE) in a telecommunications network is a device that is typically located in a customer's home or business and connects the customer to the operator's network (such as a digital subscriber line (DSL) network, a fiber to the premises (FTTP) network, and / or a cellular network). CPE can also provide a local area network (including a wireless local area network) to the customer to connect one or more devices to the network operator's network.

[0003] CPE can be managed remotely by the operator. Remote management can be achieved by an automatic configuration server (ACS) using the TR-69 protocol (standardized by the Broadband Forum). The TR-69 protocol defines a data set that is periodically collected from the CPE for analysis by the ACS. This data set identifies all devices connected to the CPE. The ACS analyzes this data as part of a management service (e.g., to identify faults).

[0004] The data set collected according to the TR-69 protocol identifies each device connected to the hub by its Media Access Control (MAC) address. However, some devices now change their MAC addresses based on a MAC randomization process. MAC randomization is the process by which a device uses a randomly generated MAC address instead of its actual MAC address when communicating with other devices in the network. One of the first uses of MAC randomization was that a device sent a randomly generated MAC address as part of a probe request message when scanning for network access points, but used its actual MAC address when connecting to one of those network access points. The process was subsequently developed so that the device would connect to a network access point using a MAC address randomly generated for that network (so that the device uses a different randomly generated MAC address for each network it connects to). Thereafter, each network will identify the device by its network-specific randomly generated MAC address. The process has recently been developed so that a device periodically (e.g., once a day) updates this identifier in the network with a new randomly generated MAC address after connecting to the network. This can be achieved by reconnecting to the network with a new randomly generated MAC address.

[0005] A technical problem arising from this periodic MAC randomization process is that, over a period of time greater than the periodicity of the MAC randomization, the MAC address is no longer used as a permanent identifier for a device. For example, if a first record in a network-specific data set is collected before an instance of a periodic MAC randomization process, and a second record is collected in the same network-specific data set after that instance of MAC randomization, it cannot be determined whether these records are related to two separate devices or to the same device for which MAC randomization has been performed. This problem applies to data sets collected using the TR-69 protocol, but also to any other data set that identifies individual devices connected to a specific network by their MAC addresses. When applied to TR-69 data sets, operator management processes (such as fault identification and / or network optimization) may be impaired as a result. Summary of the invention

[0006] According to a first aspect of the present invention, there is provided a method for analyzing a wireless local area network, the wireless local area network comprising a plurality of devices, wherein at least one of the plurality of devices is configured to change its media access control (MAC) address, the method comprising the following steps: obtaining data comprising a plurality of records, wherein each of the plurality of records relates to a device of the plurality of devices and comprises: a device tag, a media access control (MAC) address, and a set of performance metrics; for each device tag of the plurality of records, determining a uniqueness value for the device tag based on a count of records in a plurality of records having the device tag; processing each of the plurality of records to assign a unique identifier to the record by determining that the MAC address of the record is a local MAC address and that the uniqueness value of the device tag satisfies a uniqueness threshold; and analyzing the set of performance metrics for a plurality of records having the same unique identifier.

[0007] The uniqueness value may be based on a ratio of a count of records in the plurality of records that use the device tag to a count of different timestamps in the plurality of records that use the device tag.

[0008] The step of analyzing the set of performance metrics may include combining the set of performance metrics for all records using the same unique identifier.

[0009] The method may further comprise the step of causing configuration of a device of the plurality of devices based on the analysis.

[0010] According to a second aspect of the present invention, a computer program is provided, the computer program comprising instructions, when the program is executed by a computer, the instructions cause the computer to perform the steps according to the first aspect of the present invention. The computer program may be stored on a computer readable carrier medium.

[0011] According to a third aspect of the present invention, there is provided a data processing apparatus comprising a processor adapted to execute the steps of the method according to the first aspect of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order that the present invention may be better understood, embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

[0013] Figure 1 is a schematic diagram of a telecommunication network according to a first embodiment and a second embodiment of the present invention;

[0014] Figure 2 Is Figure 1 A flowchart of the steps of the method of the first embodiment and the second embodiment of the present invention implemented by a user terminal equipment (CPE) of a network;

[0015] Figure 3 Is Figure 1 A flowchart of the steps of the method of the first embodiment of the present invention implemented by a network management system (NMS) of a network;

[0016] Figure 4 is an example data extraction of a key performance indicator table during implementation of a first embodiment of the method of the present invention;

[0017] Figure 5 is an example data extraction of a daily hostname count table during implementation of a first embodiment of the method of the present invention;

[0018] Figure 6 is an example data extraction of a hostname uniqueness table during implementation of a first embodiment of the method of the present invention;

[0019] Figure 7 is a first example data extraction of a key performance indicator and a consolidated table during implementation of a first embodiment of the method of the present invention;

[0020] Figure 8 is a second example data extraction of a key performance indicator and a consolidated table during implementation of the first embodiment of the method of the present invention;

[0021] Fig. 9 is a flow chart of the steps of the method of the second embodiment of the present invention;

[0022] Fig.10 is a flow chart of the steps of the method of the second embodiment of the present invention; and

[0023] Fig.11 is a representation of the input vector used in the method of the second embodiment of the present invention. DETAILED DESCRIPTION

[0024] Now refer to Figure 1 A first embodiment of a telecommunications network 100 is described. The telecommunications network 100 includes an operator's network 200 and a customer's network 300. The operator's network 200 and the customer's network 300 are connected via an access connection 400, which in this embodiment is a digital subscriber line (DSL). The customer's network 300 includes a customer premises equipment (CPE) 310 and a plurality of user devices 320.

[0025] CPE 310 includes an access network communication interface 311, a processor 313, a memory 315, a wired communication interface 317, and a wireless communication interface 319. The access network communication interface 311 enables CPE 310 to communicate with the operator's network 200 via the access connection 400. The wired communication interface 317 and the wireless communication interface 319 enable CPE 310 to provide a wired local area network and a wireless local area network, respectively.

[0026] In this embodiment, each of the plurality of user equipments 320 includes a wireless communication interface 321 (for communicating with the wireless local area network of the CPE 310 ).

[0027] CPE 310 may be identified by the MAC address of a network interface controller (NIC) associated with its access network communication interface 311. Each of the plurality of user devices 320 may be identified by the MAC address of the NIC associated with its respective communication interface 321.

[0028] The operator's network 200 includes an access network communication node 210 and a network management system (NMS) 220. The access network communication node 210 enables the NMS 220 (and any other node of the operator's network, or any node in an external network connected to the operator's network) to communicate with the CPE 310 via an access connection 400. The NMS 220 includes a communication interface 221, a processor 223, and a memory 225.

[0029] Now refer to Figures 2 to 8 A first embodiment of the method of the present invention is described. Figure 2 310 and the steps performed by the processing module 313 and the storage module 315 of the CPE 310 are shown. Figure 3 The steps performed by the processing module 223 and the storage module 225 of the NMS 220 are shown.

[0030] In the first step (S101) of the first process, the processor 313 of the CPE implements a diagnostic function to collect data on each of the multiple devices 320. The data is collected periodically, and the periodicity can be remotely configured by the network operator (the periodicity can range from one hundred milliseconds to every ten seconds). Each record in the data includes a MAC address of the user device, a timestamp indicating the time when the data was collected from the user device, a host name of the user device, and parameters of the user device. The parameters include a received signal strength indication (RSSI), a downlink physical layer (PHY) rate, and an uplink PHY rate. The parameters of the user device may also include:

[0031] Wireless LAN channel number;

[0032] The name of the access point (may be different from the name of the CPE when the CPE is the root access point in a mesh wireless LAN);

[0033] Downlink packet counter value;

[0034] Uplink packet counter value;

[0035] retransmission counter value; and

[0036] Total active time.

[0037] This data is referred to hereinafter as detailed diagnostic data.The detailed diagnostic data is stored in the memory 315 of the CPE.

[0038] In step S103, the processor 313 of the CPE performs a compression function to convert the detailed diagnostic data into the summary diagnostic data. The detailed diagnostic data is periodically converted into the summary diagnostic data, and the periodicity can be remotely configured by the network operator (the periodicity can be in the range of 1 minute to 10 minutes). The conversion can be achieved by applying one or more statistical functions to the detailed diagnostic data, such as the sum, average, maximum or minimum of various parameters of the user equipment. The summary diagnostic data is also stored in the memory 315 of the CPE.

[0039] In step S105, the CPE sends the summary diagnostic data to the NMS (ie, via access connection 400). Figure 3 In step S201, the NMS 220 receives the summary diagnostic data and stores it in the memory 225. Steps S101 to S105 and step S201 are performed periodically, so that the NMS 220 collects additional records and adds them to the summary diagnostic data already stored in the memory 225. In this embodiment, the summary diagnostic data stored in the memory 225 spans a time period that is greater than the MAC randomization period implemented by one or more of the plurality of user devices 320.

[0040] The NMS 220 stores a first table and a second table based on the summary diagnostic data. The first table, the key performance indicator table, includes values ​​of one or more key performance indicators for each user device in the plurality of user devices 320. The key performance indicators include, for example, total active time, total poor RSSI time, and total poor coverage time (which can be calculated from the summary diagnostic data). The key performance indicator table includes a MAC address and a host name as identifiers of the user devices. When additional records are added to the summary diagnostic data, the key performance indicator table is updated with the new records. In this embodiment, the key performance indicator table is updated with new records from the past 28 days (thereby removing any data older than 28 days).

[0041] Figure 4 An example key performance indicator table is shown in , with a single key performance indicator for total active time. It should be noted that placeholder MAC addresses (MAC1, MAC2, MAC3) are used for clarity and are not in the form of real MAC addresses.

[0042] The second table, the daily hostname count table, is populated with records as described below in step S203. This first embodiment allows the NMS 220 to determine whether multiple records of the KPI table with different MAC addresses relate to the same user device, and if so, merge those records of the KPI table.

[0043] In step S203, the NMS processor 223 obtains the data for the current day in the summary diagnostic data table and processes the obtained data to calculate, for each different host name in the obtained data, a first count, named a sample count, which is a count of all records with the host name in the obtained data; and a second count, named a unique timestamp count, which is a count of all different timestamps for a subset of records with the host name in the obtained data. The daily host name count table is updated with an entry for each different host name with its sample count value and unique timestamp count value. Each record contains an additional field, namely, a calculation date, which identifies the date on which the sample count value and the unique timestamp count value were calculated.

[0044] The daily host name count table includes records for each host name for each day of a time period, which in this embodiment is 28 days (i.e., the above process is performed on each of the past 28 days to record the sampled count value and unique timestamp count value for each different host name on each of the past 28 days).

[0045] The example daily hostname count table is in Figure 5On the eleventh day of this example, the sample count value is 25 and the unique timestamp count is 24. This can be calculated if there is a MAC randomization event on the eleventh day, so that for the same host name, there are two records using the same timestamp but different MAC addresses.

[0046] In step S205, the NMS processor 223 obtains the data of the past 28 days in the daily host name count table. The NMS processor 223 processes the obtained data to identify each different host name and calculates a uniqueness ratio for each different host name. The uniqueness ratio is calculated as the sum of the sample counts of the host name within 28 days divided by the sum of the unique timestamp counts of the host name within 28 days. The NMS processor 223 creates a third table, namely a host name uniqueness table, which includes a record for each different host name and also indicates a binary IsUnique field for each different host name to indicate whether the uniqueness ratio of the host name is below a uniqueness threshold (e.g., 1.05 or 1.1). Each record contains another field, namely the calculation date, which identifies the date when the record was created.

[0047] Figure 6 An example host name uniqueness table is shown in .

[0048] In step S207, the NMS processor 223 adds the KPI table and the hostname uniqueness table to the hostname field to create a new table, namely the KPI merged table. The KPI merged table includes all fields from the KPI table and, for the hostname, the IsUnique field and the calculation date field from the hostname uniqueness table.

[0049] In step S209, the NMS processor 223 processes all records in the KPI merge table to create two new fields, unique identifier and unique host name. The unique identifier field is determined based on the following logic:

[0050] 1. If the MAC address of the record in the KPI merge table is a global (also called universal) MAC address (which can be determined if the second character is not equal to one of 2, 4, A or E), then the unique identifier is set equal to that MAC address;

[0051] 2. If the MAC address of the record in the KPI merge table is a local MAC address (which can be determined if the second character is equal to one of 2, 4, A, or E), and the IsUnique value for the host name in the KPI merge table is "yes"), then the unique identifier is set equal to the latest MAC address for the host name (i.e., the MAC address of the record in the KPI merge table that uses the same host name and has the latest timestamp value); and

[0052] 3. If the MAC address of the record in the key performance indicator merge table is a local MAC address, and the IsUnique value for the host name of the record in the key performance indicator merge table is "No", or there is no host name value in the record in the key performance indicator merge table, then the unique identifier is set to NULL.

[0053] A unique hostname is determined based on the following logic:

[0054] 1. If the MAC address of a record in the KPI merge table is a global MAC address, set the unique host name equal to the latest host name for that MAC address (i.e., the host name of the record of the KPI table that uses the same MAC address and has the latest timestamp value);

[0055] 2. If the MAC address of the record in the key performance indicator merge table is a local MAC address, and the IsUnique value for the host name of the record in the key performance indicator merge table is "yes", then set the unique host name equal to the host name of the record in the key performance indicator merge table; and

[0056] 3. If the MAC address of the record in the key performance indicator merge table is a local MAC address, and the IsUnique value for the host name of the record in the key performance indicator merge table is "No", or there is no host name value in the record in the key performance indicator merge table, then the unique host name is set to NULL.

[0057] exist Figure 7 An example of the key performance indicator merge table after step S209 is shown in FIG.

[0058] In step S211 , the NMS processor 223 groups records in the KPI merge table that have the same (non-null) unique identifier and the same (non-null) unique host name, so that the KPI values ​​are combined (eg, summed).

[0059] exist Figure 8 An example of the key performance indicator merge table after step S211 is shown in FIG.

[0060] In step S213, the NMS processor 223 identifies one or more configurations of the network by analyzing the data in the key performance indicator merge table. These reconfigurations can be, for example, updating the quality of service (QoS) parameters for the most active device group (i.e., those devices with the largest "total active time" value), updating the steering policy for the most active device group, and / or reconfiguring the transmission properties of the access point (or access point group in the mesh network) based on the coverage KPI indicator.

[0061] The process may then loop back to step S201 for subsequent execution the next day. On the next day, the NMS 220 receives additional data to be added to the summary diagnostic data. The key performance indicator table and the daily host name count table are then updated based on the latest 28 days of summary diagnostic data. The remaining steps of the first embodiment described above may then be performed on the updated records.

[0062] The first embodiment described above enables the NMS 220 to determine when multiple records with different MAC addresses relate to the same user device. This is achieved by determining that the host name associated with those multiple records is sufficiently unique and using the latest MAC address associated with that host name as a unique identifier. All records in the key performance indicator table using that unique host name can then be analyzed as being related to a single user device.

[0063] In the above embodiment, the data set includes records for a single CPE. However, those skilled in the art will appreciate that the NMS 220 may collect data from multiple CPEs and may apply the same method to the multi-CPE data set (e.g., by isolating data for a specific CPE based on a CPE identifier such as a serial number and applying the same method to the data set for the specific CPE).

[0064] Those skilled in the art will appreciate that the first embodiment can be applied when the user device is identified by another form of persistent device tag (such as a server name or fully qualified domain name). In addition, the first embodiment can be applied to any situation where the data set for the CPE network identifies the user device by its MAC address and device tag and the MAC address may change.

[0065] Now refer to Figure 2 , Fig. 9 , Fig.10 and Fig.11 A second embodiment of the method of the present invention is described. This second embodiment utilizes Figure 1 In this second embodiment, CPE 310 performs the same Figure 2 The same process illustrated causes it to send the summary diagnostic data to the NMS 220. In this embodiment, the NMS 220 performs the following machine learning process to determine that multiple records with different MAC addresses relate to the same user device.

[0066] In the first step (S301), Fig. 9As illustrated, the NMS 220 receives the summary diagnostic data and stores it in the memory 225. In step S303, the NMS processor 223 obtains the MAC address, RSSI, downlink PHY rate, and uplink PHY rate of each record for a first time period (this may be limited to those records whose unique identifier and unique host name fields of the key performance indicator table are empty in the first embodiment described above). The first time period is selected so that it is less than the MAC randomization period of any user device. For example, if the MAC randomization period causes one or more user devices to update their MAC addresses once every 24 hours, the time period of the obtained data is less than 24 hours (e.g., 12 hours, 8 hours, or 6 hours).

[0067] In step S305, the NMS processor 223 divides the acquired data into a training data set and a validation data set. In step S307, the NMS processor 223 determines the minimum and maximum values ​​of each parameter for all records in the training data set, and independently scales the value of each parameter of each record between 0 and 1 based on its original value and the minimum and maximum values.

[0068] In step S309, the NMS processor 223 processes the training data set to determine the counts of each parameter value in consecutive non-overlapping intervals of 0.1 (i.e., 0 to 0.1, 0.1 to 0.2, ..., 0.9 to 1) for each MAC address for multiple time intervals in the time period (i.e., the first hour, the second hour, etc.).

[0069] The NMS processor 223 therefore determines the count of each parameter value between 0 and 0.1, the count of each parameter value between 0.1 and 0.2, and so on for the first MAC address for the first hour in the time period; and determines the count of each parameter value between 0 and 0.1, the count of each parameter value between 0.1 and 0.2, and so on for the first MAC address for the second hour in the time period. For each hour in the time period, a 1-dimensional vector is created for each MAC address in the training data set, which contains the count of the downlink PHY rate in each interval from 0 to 0.1 to 0.9 to 1, the count of the RSSI in each interval from 0 to 0.1 to 0.9 to 1, and the count of the uplink PHY rate in each interval from 0 to 0.1 to 0.9 to 1. Fig.11 An example of this vector constructed from these counts is shown in . In step S311, these vectors for the training data set are stored and each vector is associated with an identifier that is unique to the MAC address (and therefore the user device) from which the vector was derived in step S309.

[0070] In step S313, a neural network is trained to map between the vectors of the training data set and the associated identifiers. In this second embodiment, the "neuralnet" library of the R programming language is used, where the input layer consists of the vectors, the output layer consists of the identifiers associated with the vectors, and the hidden layer has a dimension greater than the count of different identifiers and less than the count of elements in the vectors (i.e., 30 in this example, where each vector contains 10 intervals for each of the three parameters). In more detail, the neuralnet parameters are selected as follows:

[0071] Sigmoid / Activation function: Logistic (although ReLu and tanh can be used instead);

[0072] Repetitions: 1 (although this can range from 1 to 5);

[0073] Minimum threshold: 1e-4 (although this could be in the range of 1e-3 to 1e-5);

[0074] Learning rate: 0.01 (although this can be in the range of 0.005 to 0.03)

[0075] Error function: Sum of squared errors (SSE) (although cross entropy can be used instead); and

[0076] Number of intervals: 24 (although 5 to 15 can also be used)

[0077] It should be noted that the above process is performed on the summary diagnostic data of a particular CPE, so that the neural network is trained for that particular CPE. If the NMS 220 receives data from multiple CPEs, the above process is performed on a subset of data related to a particular CPE (i.e., the data related to the particular CPE is retrieved in step S301 by using a unique identifier for the particular CPE, such as its serial number), so that the remaining steps train the neural network for the particular CPE. The neural network can also be retrained periodically (e.g., every hour) based on the latest data in the summary diagnostic data.

[0078] In step S315, the NMS processor 223 processes the validation data set in the same manner as in steps S307 to S311 described above (i.e., the NMS processor 223 scales the values ​​of the various parameters of each record in the validation data set, creates a vector for each MAC address, and stores each vector in memory together with an associated identifier that is unique to the MAC address). In step S317, the trained neural network is tested using the vectors of the validation data set as input. If the performance of the trained neural network meets the threshold (based on a comparison of the output of the trained neural network with the identifier associated with each vector in the validation data set), the trained neural network passes validation and can be used in subsequent steps of this second embodiment. If the performance of the trained neural network does not meet the threshold, the neural network can be retrained (e.g., by using different training data and / or different neural network parameters).

[0079] Go to Fig.10 In step S319, the NMS processor 223 obtains data from the summary diagnostic data, which covers the RSSI, downlink PHY rate and uplink PHY rate of each record for the second time period. The second time period is now greater than the MAC randomization period of one or more of the user devices, so that multiple records related to the same device can have different MAC addresses, so that it cannot be determined whether these multiple records are related to the same user device based solely on these records. To solve this problem, in step S321, the NMS processor 223 processes the data obtained in step S319 in the same manner as in steps S307 to S309 above (that is, the NMS processor 223 scales the values ​​of each parameter of each record in the obtained data and creates a vector for each MAC address). In step S323, the NMS processor 223 applies the verified neural network to the data (obtained in step S319 and processed in step S321), which outputs an identifier for each input vector. Therefore, the identifier identifies all records in the data related to the same device, even if they use different MAC addresses. In step S325, the data may be analyzed with the identifier (and the analysis may involve other parameters that do not form part of the vector) to identify one or more reconfigurations of the telecommunications network. The analysis may be, for example:

[0080] Creating a coverage indicator for the CPE based on the RSSI values ​​of each user equipment, which can be used to reconfigure the transmission parameters of the CPE;

[0081] • Creating utilization patterns, roaming behavior patterns, and other key performance indicators for individual user devices, which can be used to identify high utilization patterns and / or problematic devices and assign new QoS parameters and / or new steering policies to these devices.

[0082] Those skilled in the art will appreciate that the use of the neuralnet library of the R programming language is non-trivial, and any suitable machine learning process (including support vector machines, random forests, or other forms of neural networks, such as convolutional neural networks or recurrent neural networks) may be used instead. When using these other machine learning processes, the data preparation phase of steps S303 to S311 may still be applied.

[0083] In the above embodiment, the CPE 310 and the access network communication node 210 are connected via DSL. However, this is not critical and other forms of access connection may also be used, such as fixed wireless access or FTTP.

[0084] In the first embodiment, the hostname uniqueness threshold is in the range of 1.05 to 1.1. However, this is not critical. The threshold is determined by the MAC randomization period. The shorter the randomization period (relative to the time window used to determine the hostname uniqueness value (28 days in the first embodiment)), the higher the threshold will be. In other words, the hostname uniqueness threshold has an inverse relationship with the MAC randomization period.

[0085] In a second embodiment, RSSI, downlink PHY rate, and uplink PHY rate are used to identify records associated with the same user device. The inventors have determined that using three parameters provides the right balance of accuracy (when identifying records associated with the same device) without unnecessary calculations (because more than three parameters will increase the resources required to train / run the machine learning model). In addition, RSSI, downlink PHY rate, and uplink PHY rate are examples only, and other parameters may be used.

[0086] A person skilled in the art will appreciate that any combination of features is possible within the scope of the claimed invention.

Claims

1. A method for analyzing a wireless local area network, wherein the wireless local area network comprises a plurality of devices, wherein: At least one device among the plurality of devices is configured to randomly change its media access control (MAC) address, and the method comprises the following steps: Data comprising a plurality of records is obtained, wherein each record in the plurality of records relates to a device in the plurality of devices and comprises: Device tags, Media Access Control MAC address, and Performance measurement set; For each device tag in the plurality of records, determining a uniqueness value for the device tag based on a count of records in the plurality of records with the device tag and a ratio of the count of records in the plurality of records with the device tag to a count of different timestamps in the plurality of records with the device tag; processing each record of the plurality of records to assign a unique identifier to the record by determining that the MAC address of the record is a local MAC address and the uniqueness value of the device tag satisfies a uniqueness threshold; and The set of performance metrics for multiple records having the same unique identifier is analyzed.

2. The method according to claim 1, wherein: The step of analyzing the set of performance metrics includes combining the set of performance metrics for all records using the same unique identifier.

3. The method according to claim 1 or 2, wherein: The step of analyzing the performance metric set comprises: By analyzing the set of performance metrics for a plurality of records having the same unique identifier, a configuration for a device among the plurality of devices is identified.

4. A computer-readable carrier medium storing a computer program which, when executed by a data processing device, causes the data processing device to perform the method according to any one of claims 1 to 3.

5. A data processing apparatus comprising a processor adapted to perform the steps of the method according to any one of claims 1 to 3.

Images