EP2807770A1

EP2807770A1 - Time synchronization in dsl environment

Info

Publication number: EP2807770A1
Application number: EP12701876.0A
Authority: EP
Inventors: Martin Kuipers; Antti Olavi PIETILAINEN
Original assignee: Nokia Solutions and Networks Oy
Current assignee: Nokia Solutions and Networks Oy
Priority date: 2012-01-26
Filing date: 2012-01-26
Publication date: 2014-12-03
Also published as: WO2013110332A1

Abstract

A method and a device for time synchronization of network components in a digital subscriber line environment is provided, wherein at least one time instance point is conveyed via at least one synchronization sequence, and wherein the at least one synchronization sequence is correlated with a local function for determining the moment of arrival of the at least one time instant point. Furthermore, Also, an according transmitter, receiver and transceiver as well as a system comprising at least any of those are suggested.

Description

Time Synchronization in DSL Environment The invention relates to a method for time synchronization of network components in a digital subscriber line

environment. Also, an according transmitter, receiver and transceiver as well as a system comprising at least any of those are suggested.

It is a general motivation to provide time synchronization to network elements in need of a synchronized clock signal. In case of cellular technologies, base (transceiver) stations need to be time synchronized. The accuracy of such synchronization is rather high, varying typically between

+/- 500 ns and +/- 3 ys . However, stricter requirements are already envisioned, e.g., the E911 location accuracy target may require an accuracy amounting to +/- 150 ns over the air interface .

Known approaches use time synchronization via GPS satellite receivers. This solution is rather expensive due to the costs involved for GPS antenna installation. In addition, the synchronization is dependent on GPS signals and does not work without such signals .

Another prior art approach is the precision time protocol (PTP), which provides time synchronization over wires.

Fig.l visualizes a time delivery model according to the PTP between a source time running on a first network component 101 and a slave time running on a second network component 102. At a time tl, the network component 101 sends a synchronization message 103 to the network component 102, which arrives at the network component 102 at a time t2 ' . At a time t3 ' , the network component 102 sends a delay request 104 to the network component 101, which arrives at the network component 101 at a time t4. The t4 timestamp is sent to the slave separately in a Delay_Resp message.

The time stamps tl and t4 are created in the network component 101 and the time stamps t2 ' and t3 ' are created in the network component 102.

The time error of the clock of the network component 102 is At' = t2'-t2 (1) although t2 is unknown but can be calculated from the round trip delay (RTT) , assuming symmetric delay to both

directions . i2 = il +—— (2) yielding

RTT

Δί' = ί2'-ί1 - as the time error. The RTT, on the other hand, can be calculated from the four time stamps as follows : RTT = [(ί2' - 11) + (U - t3')} . ( 4 )

By inserting equation (4) into equation (3), the time error can be calculated using the known time stamps as Δί' = ^ · [(ί2' - il) + (£3' - £4)] . (5)

For further illustration of the PTP, reference is made to http : / /en . wikipedia . org/wiki/Precision_Time_Protocol . Ethernet switches supporting PTP are known. However, the PTP standard does not describe an appropriate physical layer for copper loops used in telecommunication

applications .

The problem to be solved is to overcome the disadvantages mentioned above and in particular to provide efficient time synchronization with high accuracy for telecommunication and data networks .

This problem is solved according to the features of the independent claims . Further embodiments result from the depending claims . In order to overcome this problem, a method for time synchronization of network components in a digital

subscriber line environment is provided,

— wherein at least one time instance point is

conveyed via at least one synchronization sequence, — wherein the at least one synchronization sequence is correlated with a local function, in particular a local sine wave function for determining the moment of arrival of the at least one time instant point .

Advantageously, as the spectral width of a sine wave is indefinitely small, the shape of the sine wave does not change in a linear transport medium even when dispersion is present. As the shape of the wave is preserved, it can be ( cross-) correlated with the local (sine wave) function for determining the moment of arrival of the time instance point with high accuracy.

The synchronization sequence may comprise a sine wave sequence. Preferably, the correlation function has the same shape as is expected from the incoming synchronization sequence. However, if the synchronization sequence is not a sine wave sequence, the local function used for correlation purposes may be of different shape as the incoming

synchronization sequence. It is noted that the cycle or period mentioned herein could be understood, e.g., in the meaning of a sine wave period. It is noted that a sine wave could have many periods . In addition, the terms cycle and period can be understood to be associated with other than sine wave purposes.

The correlation can be conducted by changing frequency and/or time of a correlator function .

In an embodiment, the at least one time instance point conveyed is received at one component and correlated with the local (sine wave) function for determining the moment of arrival of the at least one time instant point.

It is yet an embodiment that the frequency of the local function, in particular said local sine wave function, is varied .

Hence, by modifying its frequency or by adjusting

(stretching or compressing) the local function in time it can fit into the incoming signal.

In another embodiment, the at least one time instance point is conveyed from a transmitter to a receiver via at least one synchronization sequence.

In a further embodiment, the at least one synchronization sequence is conveyed via a frequency carrier in a DMT modulation scheme. It is noted that the transmission of data in a DMT based system is not done by continuously transmitting subcarriers or tones . Instead, DMT symbols are transmitted that are processed separately.

Hence, in a next embodiment, the at least one

synchronization sequence is conveyed via DMT symbols, wherein coefficients Z of the IDFT samples f ■

x = Vexplj-2-π —\-Z. for n = 0 to 2N -1 are set constant with a defined phase for a used tone m and a symbol /, i.e.

Z_MJ =-exp(;(2^ ·(/·*-[/·*])+ <P_CP )) and

Ζ_{ΐΝ ,}ι =jexp(- j{2 ^■ {l^■ k - \l^■ kJ)+ φ_£Ρ )) with wherein L_Cp is a number of samples dedicated to the cyclic prefix.

This results in a real-valued sine signal which is continuous in phase and starts with the phase <p₀=0 at the beginning of the first sample of the first symbol. The amplitude of the samples could be changed (e.g., lowered) for samples corresponding to one sine wave for indicating a location in the sequence . A lowered amplitude cannot indicate the exact time stamp point, because of distortion. In case no DMT modulation (e.g., DSL modulation signal) is present, the amplitude could be changed to indicate the exact time stamp point, which is a certain number of cycles away (previously or upcoming) .

Pursuant to another embodiment, the synchronization sequence is started at a phase amounting to <p₀= at certain instances in time.

This could be done at every synchronization symbol which is inserted after, e.g., 256 data symbols to synchronize both VTUs and to signal online reconfiguration. As an

alternative, every y-th (e.g., tenth) synchronization symbol can be used.

According to an embodiment, the synchronization sequence is started in accordance with a predefined reference.

Hence, the start of a new sequence can be determined in accordance with a global reference. The sine signal may thus begin with the first DMT symbol. It is possible to count and refer to the number of periods of the sine signal. Even if the system loses synchronization, it will still be possible to recover the correct number of periods, because the count of DMT symbols is known when the

communication is re-established. As an option, the reference period can then be signaled via an embedded channel (e.g., an embedded operations channel - eoc) . The eoc conveys management and/or control information similar to a control plane. Pursuant to an embodiment, the synchronization sequence is conveyed at least partially in the absence of any other modulation, in particular any DSL modulation. According to another embodiment,

- a local sampling clock is used to generate the DMT signal,

- a network timing reference is used to determine the timing of the sampling clock.

As the DMT signal with all its tones is generated using a sampling clock that is not synchronized to any other reference clock, the frequency and stability of the generated sine signal may not fulfill the requirements for precise synchronization.

In VDSL2, the sub-carrier spacing is proportional to the sampling clock. Nevertheless, VDSL2 has a mechanism to transport an 8 kHz network timing reference (NTR) from the VTU-0 to the VTU-R. The VTU-0 derives a local 8 kHz timing reference (LTR) by dividing its sampling clock by an appropriate number. The VTU-0 may estimate the change of the phase offset between the NTR and the LTR and provides this estimate in cycles of the sampling clock running at the frequency 8192 Δ/, where Af is the subcarrier spacing.

Hence, the NTR information can be used to determine the exact frequency of the sine wave used for synchronization.

In order to complete the synchronization procedure between the VTU-0 and the VTU-R, the mechanism described above may be used also in the opposite direction. In upstream direction, the VTU-R generates the sine signal in an appropriate upstream band. Basically, this is identical to the procedure for the downstream.

In yet another embodiment, at least one synchronization sequence is conveyed via at least one frequency carrier in uplink direction. According to a next embodiment, at least one synchronization sequence is conveyed via at least one frequency carrier in downlink direction. Downlink (downstream) corresponds to a direction towards an end user equipment (e.g., CPE), whereas uplink (upstream) corresponds to a direction towards a central unit (e.g., ONU, DSLAM, etc.) and/or the core network.

Pursuant to yet an embodiment,

- at least one synchronization sequence is conveyed via at least one frequency carrier in uplink direction,

- at least one synchronization sequence is conveyed via at least one frequency carrier in downlink direction,

- the frequency carrier in uplink direction and the frequency carrier in downlink direction are close to each other on the frequency band.

The frequency for the upstream sine signal can be selected close (in particular adjacent) to the downstream frequency for the sine signal in order not to introduce any asymmetry when calculating the propagation delays. For example, the highest upstream carrier and the lowest downstream carrier (or vice versa) of two band next to each other can be used for that purpose .

There may be delay asymmetries between the forward and reverse directions causing time error. If the asymmetry is known and invariable, the asymmetry can be compensated when the time error is determined. Further, the asymmetry may be variable as a function of the round-trip-time or modulation used in the DSL signal. If the function is an invariable characteristic of the method, the asymmetry could at least partially be compensated in the time error calculation. According to yet an embodiment, the at least one

synchronization sequence is correlated with the local (sine wave) function via a curve fitting mechanism. According to another embodiment, points of the local (sine wave) function are used for correlation purposes which coincide with the at least one synchronization sequence received . The problem stated above is also solved by a transmitter in a digital subscriber line environment comprising or being associated with a processing unit that is arranged

- for conveying at least one time instance point via at least one synchronization sequence, wherein the at least one synchronization sequence is conveyed via a frequency carrier in a DMT modulation scheme.

The problem stated above is further solved by a receiver in a digital subscriber line environment comprising or being associated with a processing unit that is arranged

- for receiving at least one time instance point via at least one synchronization sequence,

- for correlating the at least one synchronization sequence with a local function, in particular a local sine wave function, for determining a moment of arrival of the at least one time instant point.

It is noted that the steps of the method stated herein may be executable on any such processing unit as well.

It is further noted that said processing unit can comprise at least one, in particular several means that are arranged to execute the steps of the method described herein. The means may be logically or physically separated; in

particular several logically separate means could be combined in at least one physical unit. Said processing unit may comprise at least one of the following: a processor, a microcontroller, a hard-wired circuit, an ASIC, an FPGA, a logic device.

The problem stated above is in addition solved by a transceiver comprising a receiver and a transmitter as described herein.

The solution provided herein further comprises a computer program product directly loadable into a memory of a digital computer, comprising software code portions for performing the steps of the method as described herein.

In addition, the problem stated above is solved by a computer-readable medium, e.g., storage of any kind, having computer-executable instructions adapted to cause a computer system to perform the method as described herein.

Embodiments of the invention are shown and illustrated in the following figures : shows a frequency usage scheme and DMT modulation as applied in ADSL2plus; shows the structure of the physical media dependent (PMD) sub-layer of a VDSL2 system; visualizes cyclic extension, windowing and overlap of DMT symbols according to ITU Recommendation G.993.2;

Fig.5 shows an exemplary diagram comprising a sequence of sine waves, wherein the time stamp point n=K is a particular point in a sequence of sine wave cycles;

Fig.6 shows a schematic diagram visualizing the

correlation between the measured (received) signal and the locally created sine wave at the receiving component ;

Fig.7 shows a diagram visualizing the output of the

correlation process .

In DSL applications, symbols are used for transporting information, which have a spectral width of several kilohertz . The shape of the waveform changes due to dispersion which leads to inaccuracy regarding the location of the transmitted time stamp signal.

The approach presented solves the dispersion problem by using synchronization sequences for transporting time instance points .

As the spectral width of a sine wave is indefinitely small, the shape of the sine wave does not change in a linear transport medium even when dispersion is present. As the shape of the wave is preserved, it can be cross-correlated with a function, e.g., a local sine wave function, for determining the moment of arrival of the time instance point with high accuracy. Hereinafter, time synchronization is illustrated for a DMT based DSL system, wherein DSL may refer to, e.g., ADSL, ADSL2plus or VDSL2.

At first, clocks of CPEs (VTU-Rs) are frequency

synchronized with a clock of a DSLAM (VTU-O) via generating appropriate sine waves. DMT based systems modulate

information via large set of tones (frequencies, i.e.

around carrier frequencies) . Each tone is used to carry a specific amount of information by using modulation schemes. Fig.2 shows a frequency usage scheme and DMT modulation as applied in ADSL2plus . One of the tones can be used per direction to generate sine waves required. A pure sine wave may in this regard correspond to an un-modulated tone. However, the

transmission of data in a DMT based system is not done by continuously transmitting subcarriers or tones . Instead, DMT symbols are transmitted that are processed separately.

The structure of the physical media dependent (PMD) sublayer of a VDSL2 system is depicted in Fig.3 for

illustration purposes (see also specification according ITU recommendation G.993.2) . The data stream is modulated using an IDFT adding cyclic extensions and applying a windowing technique between two successive DMT symbols. The last L_Cp samples of the IDFT output are prepended to the 2N output IDFT samples x_n as the cyclic prefix (CP) . The first L_Cs samples of x_n are appended to the block of x_n + L_Cp samples as cyclic suffix (CS) . The first β samples of the cyclic prefix and the last β samples of the cyclic suffix are used for shaping the envelope of the transmitted signal, wherein the windowed parts of consecutive symbols may overlap.

Fig.4 visualizes cyclic extension, windowing and overlap of DMT symbols according to ITU Recommendation G.993.2.

In order to send a continuous sine wave, a frequency amounting to can be generated with Δ being the sub-carrier or tone- spacing width amounting to Δ/ = 4.3125kHz or Af = 8.625kHz . The period of a DMT symbol can be set to i.e. the number of sine cycles during one DMT symbol corresponds to

If it is started with a phase <p₀=0 at the beginning of the first /=0 symbol, the generated sine waves will provide a phase of the signal at the end of the DMT symbol according to

<p, =2τ·((/ + ΐ)·£-[(/ + ΐ)·£]) In order to be phase-continuous, the DMT symbol / has to start with a phase amounting to p_M =2;r ·(/·£- [/·*_!)

This can be achieved in case the coefficients used in the IDFT

Y n^■ i

exp j-2-π — · Z. for n = 0 to 2N^■ are set constant but with a defined phase for the used tone m and the symbol /, i.e.

Z_m} =-exp(]{2π^■ (/ -k - \l -kJ)+ <p_cp))

J and =j^eXP(^" A²^^■ (^{l■ k} - l^{l ■ k}J)+ <PCP )) with This results in a real-valued sine signal which is

continuous in phase and starts with the phase <p₀=0 at the beginning of the first sample of the first symbol.

There are several possibilities to signal the start of a new sequence. Due the IDFT processing applied, it may not be possible to alter the coefficients Z_ml(n) and Z_2N__ml(n) as a function of n, to lower the amplitude for the samples corresponding to one sine wave. Therefore, another mechanism is required to provide a suitable reference for the start of a new sequence. The first possibility is to start the new sequence at a phase amounting to <p₀= at certain instances in time. This could be done at every synchronization symbol which is inserted after 256 data symbols to signal online reconfiguration.

Nevertheless, this means that the correlation time for the synchronization process is limited since the PLL to track the sine signal needs to be reinitialized every 64.25 ms in case the mandatory value for the cyclic extension is used. It could be an alternative to use every y-th

synchronization symbol, wherein y could be in a range, e.g., between 1 and 100.

Another possibility to signal the start of a new sequence is to use a global reference. The synchronization sequence begins at the first DMT symbol. It is possible to count and refer to the number of periods of the sine signal. Even if the system loses synchronization, it will still be possible to recover the correct number of periods from the equations given above, because the count of DMT symbols is known when the communication is re-established. The reference period can then be signaled via an embedded channel (e.g., an embedded operations channel - eoc) .

The next issue to be solved is the fact that the DMT signal with all its tones may be generated using a sampling clock that is not synchronized to any other reference clock.

Therefore, the frequency and stability of the generated sine signal may not fulfill the requirements for precise synchronization. While the frequency of the sine signal does not have to be linked to a precise reference, it is still may be useful to know the frequency accurately to estimate the time of a time stamp point in future.

In VDSL2, the sub-carrier spacing is proportional to the sampling clock. Nevertheless, VDSL2 has a mechanism to transport an 8 kHz network timing reference (NTR) from the VTU-0 to the VTU-R. The VTU-0 derives a local 8 kHz timing reference (LTR) by dividing its sampling clock by an appropriate number. The VTU-0 may estimate the change of the phase offset between the NTR and the LTR and provides this estimate in cycles of the sampling clock running at the frequency 8192 Δ/ . Hence, the NTR information can be used to determine the exact frequency of the sine wave used for synchronization.

The VTU-R then needs to derive the exact time stamp point which is K periods after the start of a synchronization sequence. The exact point in time when the sine wave has, for example, a zero crossing can be determined (e.g., estimated) by a PLL running at the VTU-R that is linked to the sine signal.

In order to complete the synchronization procedure between the VTU-0 and the VTU-R, the mechanism described above has also to be applied in the opposite direction. In upstream direction, the VTU-R generates the sine signal in an appropriate upstream band. Basically, this is identical to the procedure for the downstream. However, the frequency for the upstream sine signal can be selected close to the downstream frequency for the sine signal in order not to introduce any asymmetry when calculating the propagation delays .

Advantageously, an upstream band adjacent to a downstream band can be utilized to keep asymmetry and the effect of dispersion low. For example, the highest upstream carrier and the lowest downstream carrier (or vice versa) of two band next to each other can be used for that purpose.

Hence, the approach presented bears the advantage of precise time synchronization between a DSLAM and an associated CPE. This is a significant improvement in accuracy compared to mere time stamps related to particular samples of a DMT symbol. The solution presented also is much less affected by dispersion or frequency dependent group delay of the signal compared to other solutions . This approach thus allows for a precise synchronization of systems with rather low sampling rates (e.g., ADSL2/2plus) .

Correlation of time stamp point The component generating the sine wave signal used for delivering time knows the phase of the signal compared to its local clock. Thus, this component also knows the time when the time stamp point is sent out. The time of the time stamp point is communicated via a data channel .

The incoming DSL signal comprising said sine wave signal is sampled at a receiving component. The time instant of each sample (measured by the receiving component's local clock) is known. The correlation process determines the phase of the sine wave compared with the time instance of the samples . Fig.5 shows an exemplary diagram comprising a sequence of sine wave cycles . The time stamp point n=K is a particular point in the sequence of sine wave cycles . The synchronization sequence begins with a starting symbol at n=0. It can be signaled in various ways . According to one example, the amplitude of the sine wave at the

beginning of a sequence can be lowered thereby indicating the beginning of the sequence. The start of a sequence can also communicated by other means; in such scenario, no starting point indication has to be part of the sine wave signal (the amplitude of the sine wave signal may thus be constant) . In this case the other means could also refer directly to the time stamp point at n=K.

The time stamp point is a certain part of the K-th cycle of the sine wave signal, for example the zero crossing shown in Fig.5; as an alternative, a peak of the sine wave signal could be used. The timestamp point could be in the middle of the sequence. The receiver may determine the location of the time stamp point by fitting a locally generated sine wave with the received sine wave (curve-fitting mechanism) .

In order to determine the correct frequency for the local sine wave, the receiver could be frequency synchronized to the signal received from the DSLAM. The PLL filter time constant should be large enough, e.g., one second, for minimizing jitter and wander. DSL recommendation uses operation channels (e.g., eoc or roc) for conveying all kind of control information. It is suggested herein, to code the time stamp value within an eoc or a roc message. When the synchronization sequence reaches the receiving component, an algorithm at the receiving component may provide an initial guess of the location of the time stamp point by locating the K-th zero crossing (according to the example shown in Fig.5) . The receiving component's local time at the point of receiving this sample is stored. The number of samples per cycle may be relatively low

especially if the timing signal is analog-digital converted by a sampler of the DSL chip. This is not a problem since the local correlation sine wave may have a higher density of points. For correlating purposes, only those points of the local sine function are used which coincide with the measured points at a given time.

A measured point refers to the measured signal averaged over a short integration period, common to analog-to- digital converters . The time instant of the measured point may refer effectively to the time at the middle of the integration period. If the measured point refers to some other time, a correction or compensation scheme can be used . The cross-correlation is determined between the received sine wave and the locally created sine wave. Fig.6 shows a schematic diagram visualizing the correlation between the measured (received) signal and the locally created sine wave at the receiving component . The correlation function is calculated at different phase offsets between the sine waves, i.e. one sine wave is slid over the other.

As the correct cycle containing the time stamp point is already determined, the sliding length of the measured and locally created sine function could be rather short, for example, a little less than a cycle of the sine wave or several cycles. However, the whole or almost the whole synchronization sequence could be used by the correlation for noise filtering purposes .

Fig.7 shows a diagram visualizing the output of the correlation process. The output of the correlation process is a function of the location of the correlator time stamp point sliding along the measured samples . As the output is periodic, the cycle of the sine wave has to be known, e.g., determined in advance .

The x-axis value at the peak of the correlation function tells the time of the local clock when the sample nearest to the time stamp point of the received sine wave function was measured.

The actual peak of the correlator output may occur between two data points. Therefore, a correlation peak function could be fitted to the correlation data points for

targeting sub-sample time accuracy.

The frequency of the sine wave may not be accurately known. In this case, the frequency of the local correlation sine wave could be scanned to find the accurate frequency of the transmitted sine wave . When the frequency is exactly the same as that of the transmitter, the correlation is maximal. After this procedure, the cross-correlation operation should be repeated for increasing the accuracy.

If the transmitting clock at the DSLAM is free running, then in the transmitting component a similar cross- correlation function as in the receiving component can be used for finding out the exact time of the time stamp point. This information can be sent to the receiver afterwards .

A high frequency may be easier for locating the time stamp point. However, the curve fitting may be quite efficient in finding the time stamp location, at for example 1/20 of sine wave duration even when affected by noise.

There may be delay asymmetries between the forward and reverse directions causing time error. If the asymmetry is known and invariable, the asymmetry can be compensated when the time error is determined. Further, the asymmetry may be variable as a function of the round-trip-time or modulation used in the DSL signal. If the function is an invariable characteristic of the method, the asymmetry could at least partially be compensated in the time error calculation.

Hence, synchronizing the network components can be

summarized considering the following steps, whereas least one of those steps i s deemed optional :

1) It is assumed that a time at a head end (e.g., of the VTU-O) is derived from a global time reference (e.g., a master clock) .

[2a) The sine wave to be conveyed downstream is generated at the head end. The frequency of this sine wave is derived from the time available at the head end.

(2b) As an alternative to (2a) : A free-running oscillator is used to create the frequency of the sine wave.

J3) Determine the time of the time stamp point (compared to the time stamp point tl shown in Fig.l) of the sine wave using the time at the head end.

J4) The time of this time stamp point (tl) is transmitted to the receiver (remote end, e.g., a VTU-R) . 5) The receiver determines the time when the time stamp point is received (corresponding to the time t2 ' shown in Fig.l) in the synchronization sequence using its own time base.

(6) The receiver locks the frequency of the sine wave to be transmitted to its time base. Optional: A free- running oscillator may be used to create the frequency of the sine wave to be transmitted in upstream direction .

(7) The receiver determines the time t3 ' (e.g., via

correlation) of the time stamp point when the upstream sine wave is transmitted to the head end using its time base for further processing by itself.

(8) The head-end determines the time of the received time stamp point (which corresponds to the time stamp point t4 of Fig.l) in the upstream synchronization sequence.

(9) The head end conveys the time information regarding to this time stamp point (t4) to the remote end. 10) The remote end determines the time error

List of Abbreviations :

ADSL Asymmetric Digital Subscriber Line

CPE Customer Premises Equipment

DMT Discrete Multi-Tone

DS Downstream

DSL Digital Subscriber Line

DSLAM DSL Access Multiplexer

eoc Embedded Operations Channel

IDFT Inverse Discrete Fourier Transform

LTR Local Timing Reference

NTR Network Timing Reference

ONU Optical Network Unit

PLL Phase-Locked Loop

PMD Physical Media Dependent

PMS-TC Physical Media Specific Transmission Convergence roc Robust Operations Channel

US Upstream

VDSL Very High Speed Digital Subscriber Line

VTU VDSL2 Transceiver Unit

VTU-0 VTU at the ONU (or central office, exchange, cabinet, etc., i.e., operator end of the loop)

VTU-R VTU at the remote site (i.e., subscriber end of the loop)

Claims

Claims :

A method for time synchronization of network

components in a digital subscriber line environment,

- wherein at least one time instance point is

conveyed via at least one synchronization sequence,

- wherein the at least one synchronization sequence is correlated with a local function, in particular a local sine wave function, for determining the moment of arrival of the at least one time instant point .

The method according to claim 1, wherein the at least one time instance point conveyed is received at one component and correlated with the local function for determining the moment of arrival of the at least one time instant point.

The method according to any of the preceding claims, wherein the frequency of the local function, in particular said local sine wave function, is varied.

The method according to any of the preceding claims, wherein the local function is varied by adjusting, in particular by stretching or compressing, the local function in time such that it substantially fits into the incoming signal.

The method according to any of the preceding claims, wherein the at least one time instance point is conveyed from a transmitter to a receiver via at least one synchronization sequence.

The method according to any of the preceding claims, wherein the at least one synchronization sequence is conveyed via a frequency carrier in a DMT modulation scheme . The method according to claim 6, wherein the at least one synchronization sequence is conveyed via DMT symbols, wherein coefficients Z of the IDFT samples

Y n^■ i

exp j-2-π — · Z. for n = 0 to 2N^■ are set constant with a defined phase for a used tone m and a symbol /, i.e.

Z_mJ =-exp(]{2π^■ (/ · k - \l^■ kJ)+ <p_CP))

J and

^Z _2N-_m,i =J ^exP("ή^π^■ (/ · k - \l^■ kJ)+ ⁽p_cp )) with r wherein L_Cp is the cyclic prefix.

The method according to claim 7, wherein a start of the synchronization sequence is signaled by changing the coefficients Z_ml(n) and Z_2N__ml(n) as a function of n

The method according to any of claims 7 or 8, wherein the synchronization sequence is started at a phase amounting to #¾=0 at certain instances in time.

The method according to any of claims 7 to 9, wherein the synchronization sequence is started in accordance with a predefined reference . 11. The method according to any of the preceding claims, wherein the synchronization sequence is conveyed at least partially in the absence of any other

modulation, in particular any DSL modulation.

The method according to any of the preceding claims,

- wherein a local sampling clock is used to generate the DMT signal,

- wherein a network timing reference is used to

determine the timing of the sampling clock.

The method according to any of the preceding claims, wherein at least one synchronization sequence is conveyed via at least one frequency carrier in uplink direction .

The method according to any of the preceding claims, wherein at least one synchronization sequence is conveyed via at least one frequency carrier in downlink direction.

The method according to any of claims 1 to 12, wherein

The method according to any of the preceding claims, wherein the at least one synchronization sequence is correlated with the local function via a curve fitting mechanism.

The method according to claim 16, wherein points of the local function are used for correlation purposes which coincide with the at least one synchronization sequence received.

18. A transmitter in a digital subscriber line environment comprising or being associated with a processing unit that is arranged

19. A receiver in a digital subscriber line environment comprising or being associated with a processing unit that is arranged,

20. A transceiver component comprising a transmitter

according to claim 18 and a receiver according to claim 19. 21. A communication system comprising at least one

transmitter according to claim 18, at least one receiver according to claim 19 and/or at least one transceiver according to claim 20.

22. A computer program product directly loadable into a memory of a digital computer, comprising software code portions for performing the steps of the method according to any of claims 1 to 17.