Dynamic delay compensation of two-way time transfer

Abstract

There is provided a method for enabling time comparison and dynamic delay compensation of timing between a first node (A) and a second node (B) in a two-way time transfer network (N), which provides a delay compensation with improved ability to compensate for dynamic delays during for real-world network conditions The method comprises monitoring time information sent between the nodes and observations associated with delays caused by environmental effects associated with the transmission medium in the network that affects the timing, to derive at least two delay compensation factors for reducing time errors in the two way time transfer to compensate for static and dynamic environmental delays in the timing.

Description

[0001] DYNAMIC DELAY COMPENSATION OF TWO-WAY TIME TRANSFER TECHNICAL FIELD

[0002] The present invention relates to the field of signal transmission, and more particularly a method for delay compensation of timing between a first node and a second node in a two-way time transfer network.

[0003] BACKGROUND OF THE INVENTION

[0004] Time transfer is a scheme where multiple sites share a precise reference time. Multiple techniques have been developed, which may include transferring reference clock synchronization from one point to another, often over long distances. Time transfer may be used for time synchronization between different entities or nodes in a network, which is essential for the function of the network. Synchronization of timing in a network may be performed using different kind of synchronization protocols depending on the network and application. Some examples are, e.g., Precision Time Protocol (PTP), also known as IEEE 1588, NTP, and Dynamic Synchronous Transfer Mode (DTM) time transfer. Examples of different networks are for example Local Area Network (LAN), and Wide Area Network (WAN).

[0005] Distribution of digital terrestrial television in Over-the-top (OTT) networks and mobile digital television (MDTV) frequently utilizes single frequency networks (SFN). In an SFN, several transmitters simultaneously send the same signal over the same frequency channel. The transmitters in an SFN must be synchronized to send their signals at the same time to avoid interference at the receiving antennas. This is commonly achieved by installing global positioning system (GPS) receivers at all transmitter sites. GPS receivers, however, may be easily intentionally or unintentionally jammed, or fail for other reasons such as equipment failure, and represent an additional cost in the network in terms of equipment and supervision. Further, the military control of the GPS may be an issue. Techniques for time synchronization of network nodes without utilization of GPS also exist. For instance, the network time protocol (NTP) may be used to synchronize the clocks of network nodes to a master node or a reference clock using time stamps. However, the accuracy of NTP, at least in non-dedicated networks, is far too limited for the purpose of time synchronization in digital television (DTV) distribution networks.

[0006] The PTP is a protocol used to synchronize clocks throughout a computer network, also known as IEEE 1588. IEEE 1588-2019 includes a profile concept defining PTP operating parameters and options. Several profiles have been defined for applications including telecommunications, electric power distribution and audiovisual. IEEE 1588-2019 introduces a clock associated with network equipment used to convey PTP messages. The transparent clock modifies PTP messages as they pass through the device. Timestamps in the messages are corrected for time spent traversing the network equipment. This scheme improves distribution accuracy by compensating for delivery variability across the network. PTP messages may use the User Datagram Protocol over Internet Protocol (UDP / IP) for transport.

[0007] Two-way time transfer and time transfer measurements

[0008] To illustrate the main principles of two-way time transfer in a network N (prior art), refer now to Figures 1 and 2. Source nodes A and C and a slave B are connected via a network N. Time is to be distributed from the source node A, with a local time scale tA, to the slave node B, with a local time scale ts. The source node A may retrieve its time scale from a reference clock, e.g. a GPS receiver as illustrated in Figure 1, or it may be synchronized to a master node of the network. The synchronization may be performed using different kinds of synchronization protocols depending on the network and application. Some examples are, as mentioned above, e.g., PTP, NTP, and DTM time transfer.

[0009] Through the timing / synchronization protocol, node B is configured for receiving time information from node A. It may be further realized that time stamp interchange may be overlapped (tB1 < tB3 < tB2) or reversed order (tB3 < tB1 < tB2) without changing the functionality as long as the exchange is relatively close in time. Further, node A may insert its local time tAi into a stream which is transmitted to node B and reaches node B at local time tB2, as illustrated in Figure 2..

[0010] A pseudo-range observation pAB=tB2-tAi is formed in the receiver at node B. The local clock of node A is then tA2. In the same way, node B may send a time stamp to node A at local times tB3 and tA3, respectively, which is received at node A at local times tA4 and tB4, respectively. A pseudo-range observation pBA=tA4 - tB3 is formed in the receiver at node A. Further, the following relations apply:

[0011] AT= tA- tBEq. 1 tA4 = tB4 – AT Eq. 2 tA2 = tB2 – AT Eq. 3

[0012]

[0013] = tAl + dABJink Eq. 4 tB4 = tB3 + dBA. Iink Eq. 5 where AT is the clock difference between node A and node B, dAB,linkand dBA,linkare the transmission delays over the link from node A to node B, and vice versa, respectively.

[0014] The estimated time error TE between the nodes A and B can then be expressed as:

[0015] TE = (pBAJmk - PAB,link) / 2 = AT + ( dsAjink ’ dAB,link) / Eq. 6 while a Round Trip Time (RTT) can be expressed as:

[0016] RTT = (dBA,link + dAB,link) = (pBA,link + pAB,link) Eq. 7 The two-way time transfer is based on bidirectional exchange of time information between a pair of interfaces. In a basic mode of operation, the propagation delays over the link, dAB,link and dBA,link, respectively, may be assumed to be symmetric and may be calculated from the measured round trip time, which is the sum of the transmission delay of the link connections nodes A and B, dAB,link and dBA,link, according to:

[0017] dAB,link = dBA,link = RTT / 2. Eq. 8

[0018] In case of asymmetric transmission delays, i.e., dAB,link ≠ dBA,link, an asymmetric transmission delay may be corrected using a time correction factor, TEcorr = TE + tcorr. In case of asymmetric transmission delays, i.e., dAB,link ≠ dBA,link, a calibration constant Casym may be used to take the measured asymmetry into account:

[0019] dAB,link = RTT*casym Eq. 9 and dBA,link = RTT*(1 - casym) Eq. 10 where 0 < caSym < 1, and caSym = 0.5 for symmetric transmission delays. The determination of the calibration constant casymrequires knowledge of the round trip time and an asymmetry error A£, which is known, e.g., when the link is operating and both nodes receive correct time through other time sources than the link to be calibrated. The asymmetry error AE is formed from the TE expression (Eq. 6) by compensating for the time error AT between the nodes:

[0020] ΔE = TE – ΔT = TE – (tA - tB) Eq. 11 Under the assumption that tA = tB further simplification gives:

[0021] ΔE = (pBA,link - pAB,link) / 2 = (dBA,link - dAB,link) / 2 Eq. 12 Since the sum of dAB,link and dBA,link is known as RTT, dAB,link and dBA,link can be calculated as:

[0022] dAB,link= RTT / 2 - ΔE Eq. 13 and dBA,link= RTT / 2 + ΔE Eq. 14 Given this value, the casymvalue is easy to calculate from either of the dAB,link or dBA,link values to become:

[0023] casym = dAB,link / RTT = 1 – dBA,link / RTT = -1 / 2 – ΔE / RTT Eq. 15 The input and output delays of the interface, dA,out and ds n, respectively, are used in expressing transmission delays as: dAB = dA,out + dAB,link + dB,in, where dAB,link is the transmission delay of the link connections nodes A and B. A corresponding relation applies to dBA- The compensated values dAB,iink and deAMcan be calculated from dAe and d dBA as:

[0024] dAB,link = dAB - dA,out - dB,in Eq. 16 and dBA,link = dBA – dB,out – dA,in . Eq. 17 The observed pseudo-ranges thus becomes after compensation:

[0025] pAB,link = pAB - dA,out - dB,in Eq. 18 and pBA,link = pBA – dB,out – dA,in . Eq. 19 US 7,535,931 B discloses a two-way time transfer protocol for estimating a time error between the clocks of two network nodes. US 9,277,256 B discloses a node for facilitating time distribution in communication networks using a time-locked loop. As none of these methods are ideal for real-world network conditions, improved time transfer methods are needed. SUMMARY OF THE INVENTION

[0026] An object of the present disclosure is to provide an improved method for enabling time comparison between a first node and a second node in a two-way time transfer network which provides delay compensation with improved ability to compensate delays during for real-world network conditions.

[0027] According to a first aspect of the inventive concept, there is provided a method for delay compensation of timing between a first node and a second node in a two-way time transfer network, the method comprising: sending, from the first node, a first signal comprising local time information to the second node, receiving, in the first node, remote time information from the second node in response to the first signal, establishing two or more compensation factors of a delay compensation model by: determining a set of the local and remote time information, determining information from one or more observations associated with a transmission medium in a subnetwork including the first node and the second node, wherein the observations are made simultaneously with or are deducted from the set of local and remote time information, deriving the two or more compensation factors using the set of local and remote time information and the observations; and delay compensating timing using the compensation factors in the delay compensation model. The method is advantageous in that the corrected time error, i.e., corrected time difference, between the clocks of the respective nodes is provided, which is unbiased from estimated environmental effects that causes dynamic delays in the transmission in the subnetwork.

[0028] According to an embodiment of the method, the two or more compensation factors comprises at least one static delay compensation factor and at least one dynamic delay compensation factor, or at least two dynamic delay compensation factors.

[0029] According to an embodiment of the method, the static delay compensation factor is determined by extracting a static portion of a set of estimated time transfer time error, TE, values estimated using the set of local and remote time information and / or determined as a static portion derived from the one or more observations. According to an embodiment of the method, the at least one dynamic delay compensation factor is determined based on variations of round trip time, RTT, values estimated using the set of local and remote time information.

[0030] According to an embodiment of the method, the set of local and remote time information is collected during a time period. The predetermined time period is activated at one or more of at startup, by a network event, during updating of the delay compensation model, and at a predetermined periodicity, or is a continuously moving time window or time period. The length of the time period may be preset or conditioned by quality conditions or threshold conditions on the measurements.

[0031] According to an embodiment of the method, the steps of determining a set of local and remote time information, performing observations and / or establishing the delay compensation model is performed in the first node, the second node, and / or in a third (central) node, i.e., the method steps may be performed in a distributed manner and may be orchestrated by an orchestrator node or may be performed locally in the node.

[0032] According to an embodiment of the method, the delay compensation model is provided as: T

[0033]

[0034] TEcorr = TE – [tcorr + Σnf=1ΣnfO=1ρcorr(f,O)(Mf – AMf)O], wherein TEcorr is the corrected time error, TE is an estimated time error or time difference between the first and the second node, tCOrr is one or more static correction factors, Pcorr is one or more dynamic correction factors, n is a number of observations, f is the specific observation, nf is a maximum polynomial order of specific observations, O is a polynomial order of a specific observation, M is a measurement of a specific observation, AM is an average measurement (e.g., RTT0) of a specific observation, and pcorr (f,o) is a dynamic compensation factor for specific observations of a certain polynomial order.

[0035] According to an embodiment of the method, the delay compensation model is provided as: TEcorr= TE – [tcorr+ pcorr* (RTT - RTT0)], wherein RTT is the estimated round trip time between the first and the second node, and RTT0is an average value of RTT derived from the set of local and remote time information. According to an embodiment of the method, the delay compensation model is provided as: TEcorr= TE – (tcorr+ pcorr* RTT), wherein RTT is the estimated round trip time between the first and the second node.

[0036] According to an embodiment of the method, the observations are associated with delay in the transmission medium. The observations may be directly observed, e.g., by monitoring temperature and / or observe nominal wavelength of optomodules, such as Small Factor Packages (SFP), by means of external or built in temperature sensors, thermal imaging, or indirectly observed, e.g., by analyzing the local and remote time information over time.

[0037] According to an embodiment of the method, the network is an IP-network, Wide Area Network, WAN, Local Area Network, LAN, Digital Subscriber Line, DSL, Synchronous Digital Hierarchy, SDH, Optical Transport Network, OTN, a Plesiochronous Digital Hierarchy, PDH, network, wireless network, microwave link network, satellite link, satellite network, or a mobile network.

[0038] According to an embodiment of the method, the step of deriving the two or more compensation factors using the set of local and remote time information and the observations comprises using best fit calculations, such as e.g., least squares methods, median-median line methods, least absolute deviation methods, nonlinear curve fitting, robust regression methods. Using automated data analysis or optimization techniques using Al, i.e., trained neural networks, to analyze large datasets and identifying patterns and relationships in the collected local and remote time information and observations is conceivable.

[0039] According to an aspect of the inventive concept, there is provided a node in a network comprising means for performing the inventive method described herein.

[0040] According to an aspect of the inventive concept, there is provided a software module adapted to perform the inventive method described herein when executed by a computer processor.

[0041] Embodiments of the present inventive method are preferably implemented in a network system by means of software modules for signaling and providing data transport in form of software, a Field-Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC) or other suitable device or programmable unit, adapted to perform the method of the present invention, an implementation in a cloud service or virtualized machine (not shown in diagrams). The software module and / or data-transport module may be integrated in a node comprising suitable processing means / processor and memory means / memory, or may be implemented in an external device comprising suitable processing means and memory means, and which is arranged for interconnection with an existing node.

[0042] Further objectives of, features of, and advantages with, the present invention will become apparent when studying the following detailed disclosure, the drawings and the appended claims. Those skilled in the art realize that different features of the present invention can be combined to create embodiments other than those described in the following.

[0043] BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The above, as well as additional objects, features and advantages of the present invention, will be better understood through the following illustrative and non-limiting detailed description of preferred embodiments of the present invention, with reference to the appended drawings, where the same reference numerals will be used for similar elements, wherein:

[0045] Fig. 1 is a schematic block diagram illustrating an exemplifying two-way time transfer network in which embodiments of the inventive concept are applicable;

[0046] Fig. 2 is a schematic illustration of two-way time transfer between two nodes; Fig. 3 is a flow chart illustrating embodiments of a method according to the present inventive concept; and

[0047] Fig 4 is a flow chart illustrating two-way time transfer according to embodiments of a method according to the present inventive concept. DETAILED DESCRIPTION

[0048] Aspects of the present disclosure will be described more fully hereinafter with reference to the accompanying drawings.

[0049] As previously described with reference to Fig. 1 and equations 1- 19, in a two-way time transfer system, two nodes will both transmit, and receive each other's messages, thus performing two one-way time transfers to determine the difference between the remote clock and the local clock. The sum of these time differences is the Round Trip Time between the two nodes. It is often assumed that this delay is evenly distributed between the directions between the nodes. Under this assumption, half the Round Trip Time is the propagation delay to be compensated when comparing time, or performing time estimates and clock synchronization, between the nodes, etc., see e.g. Eq. 6 and Eq. 7. A drawback is that the two-way propagation delay must be measured and used to calculate a delay correction. To calculate delays and determine delay compensation, information such as time stamps, time difference measurements, correction factors, and various statistics between nodes involved in the two-way time transfer need to be available.

[0050] Referring again to Fig. 1 of node A and node B which may communicate over a network N. In the time-transfer system between node A to node B, signals are exchanged over the network over some transmission medium. The transmission medium may be guided media such as copper wires (twisted pair cables), coaxial cables, and fiber optics, or wireless media such as Wi-Fi, radio waves, microwaves, and infrared. It can also be a mix of these media, depending on the pathway along which the signals are transmitted.

[0051] In an ideal system, the delay for the signals from node A to node B (dAB.itnk) is the same as the delay for the signals from node B to node A (dBA,imk), see Eq. 8. Each node transmits signals that express the nodes time (timestamp at a given moment), tA and tg, respectively, and as it receives the signal it measures the received time stamp in relation to its own local time. As the time-differences on both sides are collected, one may conclude on the sum of the delay between the two nodes in the value of Round Trip Time (RTT) from Eq. 7 and the Time Error (TE) from Eq. 6. We can then express RTT as:

[0052] RTT = dBA,link + dAB,link (Eq. 7) TE=(tA-tB) - (dBA,link - dAB,link) / 2 (Eq. 6) RTT comprises the sum of the respective time-differences calculated between the nodes in both directions, and it can be shown to be the sum of the delay from node A to node B and the delay from node B to node A (Eq. 7). RTT is thus unbiased by any asymmetry in the link delay and it is further unbiased by any time-difference between the clocks of the node A and B.

[0053] TE comprises the difference of the two time-differences, and it can be shown to be the difference of the time of node A and the time of node B, and half the difference in delay depending on direction (AB, BA) between the nodes (Eq. 6). TE is thus biased by any asymmetry in delay in the different directions. As the time difference between the clocks in node A and B is used to steer the clock in one of the nodes in order to null the difference, if there is an asymmetry driving the TE to zero, it will embed the asymmetry into the time offset of the steered clock, which is thus an unwanted property.

[0054] One known approach to reduce the bias from the asymmetry in the delay is to estimate or measure the asymmetry and compensate it. For instance, the PTP technology compensate for input and output delays of the equipment. Further, PTP estimates the transition delay of timing packets through a Transparent Clock, and update a time-correction field.

[0055] Static compensation in two-way time transfer systems refers to corrections applied to account for fixed asymmetries in the links, e.g., between a node A and a node B. Static two-way time transfer involves making simultaneous time difference measurements between two fixed points, or nodes, where using multiple time transfer measurements between the nodes allow for offset correction of time errors caused by static asymmetry in accordance with Eq. 20, where TECOrr is the corrected time error, TE is the measured time error and tstatis a static time correction factor. Hence, a common approach to compensate for asymmetries in delays, as observed in asymmetry error AE of Eq. 11, is to use a static offset in relation to measured or estimated delays to correct the TE.

[0056] TEcorr = TE+ t stat (prior art) Eq. 20 The static compensation can also be expressed in terms of actual delays for both directions or delay corrections (based on assumed RTT / 2 delay). These approaches are variations of the same method, which are mathematically equivalent, and thus do not require separate discussion. The static compensation approach disregards the possible influence of environmental factors of asymmetry error AE of Eq. 11 which may be dynamic, such as temperature variations, traffic load variations, etc..

[0057] Dynamic compensation in in two-way time transfer systems refers to corrections applied to account for dynamic asymmetries in the links, as observed in asymmetry error AE of Eq. 11. A common approach to compensate for dynamic asymmetries in delays is to use a dynamic calibration factor in relation to measured or estimated delays or round trip time to correct the TE. A corrected value TEcorrof the time error which includes dynamic factors affecting a subnetwork can be expressed as Eq. 21, where TEcorris the corrected time error, TE is the estimated time error, RTT is the estimated round trip time, and dstatis a dynamic time correction factor.

[0058] TEcorr = TE + dstat * RTT (prior art) Eq. 21 A problem with the above described approaches with static or dynamic compensation is that choosing either of these two methods, a static TE offset compensation or a dynamic RTT compensation, involves a compromise. You either optimize offset of asymmetry error AE, with little ability to reduce any dynamic influence on the subnet, e.g., temperature variations in the subnet, or you optimize the timing with respect to dynamic asymmetry of asymmetry error AE in the subnet, e.g., temperature variations, with little ability to reduce the offset.

[0059] This is further worsened by the fact that in a practical network, intermediary equipment may add static delays which are asymmetric, fiber may be setup in such a way that different physical paths are selected between nodes depending on e.g., varying network traffic or routing changes, at which point the fiber length and environmental impact are very different for signals transmitted in different directions. This results in that any common mode effects in the subnet are reduced, and thus larger differential mode effect properties show up with very different static and dynamic properties.

[0060] In view of the above identified problem, one object of the present invention is to provide a new method for providing enhanced delay compensation of timing in two-way time transfer, which allows comparison of time between nodes in a two-way time transfer system and optionally compensation of the timing therein based on multiple parameters, including static offset and dynamic environmental factors in the system.

[0061] The new compensation method is based on the realization that one needs to provide a model that joins the effects of static offset and static and dynamic environmental factors, while allowing estimation and reduction of each of these effects independently, to reduce the offset and variations of produced timing. Compensation of environmental factors

[0062] Environmental factors may be dynamic in nature which in turn may require dynamic compensation. Dynamic compensation in two-way time transfer systems refers to attending to any alternating, dynamic asymmetries, or changing asymmetries, which may occur in the links between the nodes.

[0063] As an example, let us consider using fiber as transmission medium between the nodes. It is known that temperature shifts in fiber may cause an increase in attenuation as the refractive index if the glass can alter, which leads to higher signal loss over distances and which can shift the transmission spectrum of the fiber.

[0064] Further, temperature shifts may also affect the transmission delay in fiber. One way to address the effects of the temperature shifts is to co-locate unidirectional fibers to form a bi-directional pair. Temperature induced variations of the fiber may then be assumed to be the same in both directions, which allows for using a first-degree compensation when compensating TE, i.e., an offset compensation. This may be referred to as a common mode effect. When an external disturbance affects e.g., signals in both directions in a similar way, differences between the signals in both directions can be analyzed and common mode noise or disturbances can be filtered out and compensated for as a static compensation.

[0065] In contrast to common mode effects, differential mode effects occur when external disturbances affect signals in different directions / paths differently, requiring different measures to be taken to compensate for the disturbance.

[0066] For instance, if one assumes that the delay difference is due to difference in laser wavelength in the fiber at different temperatures of the fiber it makes sense to, instead of compensating asymmetry in respect to TE using an offset compensation / first-degree offset compensation as in static compensation / common mode compensation, compensate asymmetry in respect the measured or estimated RTT, when aiming to reduce the temperature effects of the fiber.

[0067] According to an embodiment of the method presented herein, a delay compensation model is established having a corrected value TEcorrof the time error which includes environmental factors affecting a subnetwork can be expressed as Eq. 22.

[0068] TEcorr= TE-(tcorr+ pcorr* RTT) Eq. 22, where TE is the estimated time error at a given moment, tcorr is a static delay compensation factor which may include environmental factors and other static delays in the subnet, pcorris a dynamic delay compensation factor which may include dynamic environmental factors and other dynamic delays in the subnet, and RTT is the measured or estimated RTT at a given moment.

[0069] According to another embodiment of the inventive concept, a delay compensation model is established which considers the well known fact that RTT has a large static delay, while the delay compensation model according to Eq. 22 aims to include the RTT variations. In order to better model the RTT variations, an RTT model is introduced as Eq. 23, from which we can extract the RTT variations RTTv(t) according to Eq. 24.

[0070] RTT(t)=RTT0+ RTTv(t) Eq. 23 RTTV(t)=RTT(v)-RTT0Eq. 24 The improved delay compensation model utilizes RTTv(t) instead of RTT(t), and the corrected time error TEcorris then expressed as the Eq. 25:

[0071] TEcorr= TE- (tcorr+ pcorr* RTTV(t)) = TE – (tcorr+ pcorr* (RTT(t) – RTT0)) Eq. 25, wherein TECOrr is the corrected time error at a given moment, TE is an estimated value of the time error at a given moment, tcorris a static correction factor, pcorris a dynamic correction factor, RTT(t) is a measured or estimated RTT(t) at a given moment, and RTT0is an average or mean value of RTT. The RTT0value may be established at a calibration period and represents an average (mean) value of RTT monitored during calibration / a predetermined time period. RTT0may be provided as a simple, weighted or exponential moving average.

[0072] The benefit of this model is that the dynamic variation of RTT measurements, RTTV(t), around the calibrated measurement average RTT0is used, and thus the static offset of RTT, RTT0, does not shift the static correction of the delay compensation model as in Eq. 22.

[0073] When using a delay compensation model of Eq. 22, the static component of RTT(t), the RTT0, will create a static time correction of pcorr*RTT0, which needs to be compensated in the tcorrcomponent in excess to a time offset tofs. Using the improved model of Eq. 25 will provide a tcorrthat only compensates for the static offset tofs. This establishes uncorrelated variables, thus an offset component which is unbiased with respect to RTT.

[0074] According to an embodiment of the present invention, a dynamic compensation factor for a specific observation factor of a certain polynomial order is derived to be utilized in a delay compensation model. The corrected time error can then be expressed as:

[0075] T

[0076]

[0077] TEcorr= TE - [tcorr+ ∑f=1n∑O=1npcorr(f, O)(Mf- AMf)O] Eq. 26, where n is the number of observation factors, f is the specific observation factor, nf is a maximum polynomial order of specific observation factors, O is a polynomial order of a specific observation factor, M is a measurement of a specific observation factor, AM is an average measurement (e.g., RTT0) of a specific observation, and Pcorr(f, O) is a dynamic compensation factor for specific observation factors of a certain polynomial order.

[0078] The benefit of using this generalized delay compensation model is that different contributions to asymmetry error AE, direct or indirect, can be compensated for in a linear and higher order as needed for performance and dominant parameters, in order to reduce the remaining offset and variation in the asymmetry error, and thus provide improved compensated performance.

[0079] As known in such dynamic systems, some parameters can have cross dependence, such that the multiplication of two dynamic factors can have a significant correlation to the variation, so that the model can be extended by considering any such combination as a new dynamic factor with its own dynamic compensation factor (s).

[0080] Fig. 3 is a flow chart illustrating establishing of delay compensation factors for use in a delay compensation model for use when delay compensating timing in the network, steps S1-S6, are illustrated.

[0081] Fig. 4 is a flow chart illustrating delay compensation in a two way time transfer network, in which delay compensation in normal operation, steps S10-S40.

[0082] Referring now to Fig. 4, according to the inventive concept, observations, i.e., measurements of observation factors, such as environmental effects on a subnet are performed by monitoring two-way time transfer measurements over the subnet, between a first node and a second node in the network, over a predetermined or selected period of time. The two-way time transfer is performed in the subnet, by sending (S10), from the first node, a first signal comprising local time information to the second node, and receiving (S20), in the first node, remote time information from the second node in response to the first signal. The normal two-way time transfer operation of two-way time transfer between the first and the second node comprises exchanging the local and remote time information in steps S10 and S20 and performing two-way time transfer calculations including equations 6 and 7 (S30), and then time compensating the nodes to derive a corrected time error, TECOrr, by using a selected delay compensation model (S40), a delay compensation model according to the invention e.g., any of equations 22, 25 or 26. During normal operation in a two-way time transfer network, time comparison between the first node and the second node is performed continuously.

[0083] According to embodiments of the inventive concept, referring now to Fig. 3, to provide the two or more parameters in a selected delay compensation model, two-way time transfer is performed in the subnet, by sending (SI), from the first node, a first signal comprising local time information to the second node, and receiving (S2), in the first node, remote time information from the second node in response to the first signal. This exchange of local time information of the two-way time transfer operation is monitored during a period of time, to determine (S3) a set of said local and remote time information. The set of local and remote time information may include monitored two-way time transfer measurements of the local and remote time information and time transfer estimates derived therefrom, and preferably includes RTT, the estimated round trip time of a subnetwork between the first node and the second node, and the time error TE.

[0084] In addition, observations of environmental effects on the subnet are performed. This includes making one or more observations associated with a transmission medium of the subnetwork which are performed in parallel with collecting the set of local and remote time information and / or are deducted from the set of local and remote time information (S4). The monitored TE may contain offsets and variations within the network, and may be affected by differential mode variations, e.g., different linkways in the subnetwork between the nodes.

[0085] Based on the set of local and remote time information and the information derived in steps S3 and S4, two or more compensation factors are derived (S5). The monitored RTT may contain offsets and variations within the network, and may be affected by common mode variation, e.g., temperature variations). Based on the observations, the collected RTT data is analyzed (or optionally the monitored two-way time transfer measurements are analyzed and RTT data is estimated over time.

[0086] Static delay compensation factor is determined by extracting a static portion of a set of estimated time transfer time error, TE, values estimated using said set of local and remote time information and / or determined as a static portion derived from said one or more observations. To isolate and estimate any dynamics or variations in TE to compensate for dynamics in the link, one may use, e.g., best fit according to least square method to derive pcorr. The monitored / estimated RTT data is compared to asymmetry errors AE (estimated from the collected two-way time transfer measurements) occurring during the same predetermined time, see Eqs. 11- 14.

[0087] The step of comparing the monitored / estimated RTT and the asymmetry errors AE over time may be performed using e.g., best fit according to the least squares method, to isolate and estimate any dynamics or variations in the RTT to compensate for it, see e.g. Box, Jenkins and Reinsei, " Time Series Analysis -Forecasting and Control", 4th edition.

[0088] According to an embodiment of the inventive concept, when at least one of the nodes is calibrated, it is sufficient to monitor the time differences between the nodes, e.g., in relation to a reference signal which is directly applied to the node or by measuring a reference signal via the node, to derive the asymmetry error in the subnetwork between the nodes. Thus only the asymmetry error in the link, including any dynamics in the asymmetry, needs to be compensated for (using any of equations 22, 25 or 26).

[0089] Subsequent to being derived or updated, the delay compensation factors are applied in a selected delay compensation model, e.g., according to the delay compensation models presented herein, in equations 22, 25, and 26, (S6). The resulting delay compensation model may be used to delay compensate timing between the nodes during normal operation (S40).

[0090] The delay compensation models presented herein, in equations 22, 25 and 26, enable studying of how different environmental factors affect delay in two-way time transfer timing in a network in relation to environmental states and changes over a predetermined time period, such as in an example studying temperature variations affecting parts or the whole of the network system. The observations performed and derived during set up of the compensation factors are associated with delay in the transfer medium in a subnet. The temperature variations may be cyclic, e.g. due to the rising and setting of the sun or an air conditioning (AC) going on and off. By observing the long term behavior over a period of time, by any means of comparing timing in the two-way time transfer system, environmental factors that affect delay in the system may be indirectly or directly monitored and compensated for, to, e.g., cover cycles of temperature such that variations of the cycles can be separated from the static offset, the optimum compensation for environmental based RTT variations can be achieved while also providing simultaneous or separate offset compensation.

[0091] Further, other environmental effects that affect the timing may be similarly observed and compensated for. For example, the temperature of the fiber affects the wavelength of the laser light travelling in the fiber, which produces different delay through the fiber that may thus be estimated and compensated for. Different temperature variations and cycles may be handled separately, such as variations due to the sun or a local air conditioner (AC) running next to the fiber. For instance, considering that the AC of the operating room is cycling in its own rate, then the temperature's effect on the laser wavelength can be distinguished from the variations caused by the sun on the fiber. Multiple such environmental factors can exist in each direction.

[0092] The term observations, as used herein, refers to monitoring different timing parameters in the two-way time transfer system, but may also refer to monitoring the actual aspects that affect delays in the system which can be used as a basis for correcting delays in a network, parameters such as observing different wavelengths, i.e., measurement of actual wavelength chromatic dispersion, measuring length variations in fiber, cables, etc., temperature measurements, detecting new channel numbers along the path etc. Thus, multiple direct or indirect observations of environmental variations and operational conditions can be made and be compensated for by identifying tcorrand pcorr. Such estimations may be done locally to the node, or be aided by network knowledge. Sensitivity estimations can use traditional methods such as least square methods. According to embodiments of the present inventive concept the delay compensation factors RTT0, tCOrr, and pCOrr are extracted or derived by monitoring RTT, TE, and TEcalduring a calibration period or some other predetermined time period, and performing analysis of collected two-way time transfer data. TEcalis derived as the difference in clock time in the respective node at a calibration time, i.e., the actual time difference between the nodes measured during calibration. The calibration value is then used as the node time difference AT= TEcalwhen estimating TE in the normal time transfer, Eq. 11.

[0093] A static portion of the environmental observations may be derived, e.g., by using optimal fit and least squares estimation or incremental least squares (arXiv:1604.01004 [physics. data-an], M. Danielsson, 1989) on estimated RTT values, or alternatively the mean value RTT0of the RTT values, observed over the predetermined time. The predetermined time period can refer to either a calibration period, a designated update interval, or a continuously moving period, i.e. a rolling time window.

[0094] Thus, the new method allows for both static error offset compensations and dynamic compensations due to environmental influence based on dynamic RTT. The environmental factors may comprise the temperature of the fiber in its location, which temperature may be different in different directions if the fibers are not colocated or are of different lengths, the wavelength of the laser which is temperature dependent, etc. The temperature variations may relate to different temperature cycles, such as cyclic variations of the sun or a local AC, for example.

[0095] According to an embodiment of the present inventive concept, RTT0, tCOrr, and Pcorr are extracted or derived from data collected by monitoring RTT, TE, and TEcalduring calibration / over a predetermined time period / designated update interval, and performing analysis of data. This may be performed for a specific node or for multiple of nodes in a network. Data monitoring may thus be performed for a large number of nodes, and RTT, RTT0, tcorroch pcorrmay be derived for a large number of nodes. By analyzing different nodes and comparing data between nodes, the parameters may be utilized as a starting point for compensating different paths in the network. Data handling and analysis may be performed in a dedicated node. Data from observations may be optionally distributed within the network for validation or other reasons.

[0096] The apparatus and method disclosed herein can, however, be realized in many different forms and should not be construed as being limited to the aspects set forth herein. Like numbers in the drawings refer to like elements throughout.

[0097] The terminology used herein is for the purpose of describing particular aspects of the disclosure only, and is not intended to limit the disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0098] In some embodiments a non-limiting term "node" or "network node" is used. It should be understood that this term refers to any type of node that may send and / or receive information, such as data and control information, over a network. A physical node is typically an electronic device that is attached to a network, and is capable of creating, receiving, or transmitting information over a communications channel. The node may be a data communication equipment (DCE) such as a modem, hub, bridge or switch; or data terminal equipment (DTE) such as a digital telephone handset, a printer or a host computer. In an example, the node may be a computer terminal connected to a network, such as a local area network (LAN), wide area network (WAN) or the Internet. The node typically comprises

[0099] processing circuitry, interface, a clock, memory,.etc.

[0100] The term "network" refers to any type of network over which a network node may communicate, such as a local area network (LAN), wide area network (WAN) or the Internet. The network may be referred to as an Internet Protocol (IP) network, a communication network that uses IP to send and receive messages between one or more computers, which may be implemented in Internet networks, LAN and enterprise networks, for example. The network may also be a single frequency networks (SFN) or a digital television (DTV) distribution network, which may, e.g. be based on a Dynamic Synchronous Transfer Mode (DTM) network as standardized by the European Telecommunications Standards Institute (ETSI).

Claims

CLAIMS1. A method for delay compensation of timing between a first node and a second node in a two-way time transfer network, the method comprising:sending (SI), from the first node, a first signal comprising local time information to the second node;receiving (S2), in the first node, remote time information from the second node in response to the first signal;establishing two or more compensation factors of a delay compensation model by:determining (S3) a set of said local and remote time information; determining (S4) information from one or more observations associated with a transmission medium in a subnetwork including the first node and the second node, wherein said observations are made simultaneously with or are deducted from said set of local and remote time information;deriving (S5) said two or more compensation factors using said set of local and remote time information and said observations; and delay compensating timing using (S6) said compensation factors in said delay compensation model.

2. A method according to claim 1, wherein said two or more compensation factors comprises at least one static delay compensation factor and at least one dynamic delay compensation factor, or at least two dynamic delay compensation factors.

3. A method according to claim 2, wherein said static delay compensation factor is determined by extracting a static portion of a set of estimated time transfer time error, TE, values estimated using said set of local and remote time information and / or determined as a static portion derived from said one or more observations.

4. A method according to any preceding claim, wherein said at least one dynamic delay compensation factor is determined based on variations of round trip time, RTT, values estimated using said set of local and remote time information.

5. A method according to any preceding claim, wherein said set of local and remote time information is collected during a time period, and wherein said predetermined time period is activated at one or more of at startup, by a network event, during updating of said delay compensation model, and at a predetermined periodicity, or is a continuously moving time window.

6. A method according to any preceding claim, wherein determining a set of local and remote time information, performing observations and / or establishing the delay compensation model is performed in said first node, said second node, and / or in a third node.

7. A method according to any preceding claim, wherein said delay compensation model is provided as:nnf T 1 LFt. = TE - tCOrr " I" IS PcorrCf > O)(Mf ~ AMf)°f=l 0 = 1, wherein TECOrr is the corrected time error, TE is an estimated time error, tCOrr is one or more static compensation factors, pCOrr is one or more dynamic compensation factors, n is the number of observations, f is the specific observation, nf is a maximum polynomial order of specific observations, O is a polynomial order of a specific observation, M is a measurement of a specific observation, AM is an average measurement of a specific observation, and pCOrr (f,o) is a dynamic compensation factor for specific observations of a certain polynomial order.

8. A method according to claim 7, wherein said delay compensation model is provided as: TEcorr= TE – [tcorr+ pcorr* (RTT - RTT0)], wherein RTT is the estimated round trip time between said first and said second node, and RTT0is an average value of RTT derived from said set of local and remote time information.

9. A method according to claim 7, wherein said delay compensation model is provided as:TEcorr= TE − (tcorr+ pcorr* RTT), wherein RTT is the estimated round trip time between said first and said second node.

10. A method according to any preceding claim, wherein said observations are associated with delay in said transmission medium.

11. A method according to any preceding claim, wherein said network is an IP-network, Wide Area Network, WAN, Local Area Network, LAN, Digital Subscriber Line, DSL, Synchronous Digital Hierarchy, SDH, Optical Transport Network, OTN, a Plesiochronous Digital Hierarchy, PDH, network, wireless network, microwave link network, satellite link, satellite network, or a mobile network.

12. A method according to any preceding claim, wherein said step of deriving (S5) said two or more compensation factors using said set of local and remote time information and said observations comprises using best fit calculations.

13. A node in a network comprising means for performing a method according to any of claims 1-12.

14. A software module adapted to perform a method according to any of claims 1-12 when executed by a computer processor.

Images