System and Method for Supporting Phase Alignment on DOCSIS

Abstract

A method for adjusting a phase error includes a first device detecting a phase error at a timing interface of the first device, the first device modifying DOCSIS timing protocol parameters based on the detected phase error, and the first device modifying the DOCSIS timing protocol parameters to cause a modification of a timestamp output of a second device in a manner based on the detected phase error.

Description

Technical Field

[0001] Cross - reference to Related Applications This application claims the benefit of U.S. Provisional Patent Application No. 63 / 150,966, filed on February 18, 2021.

Background Art

[0002] The subject matter of this application generally relates to the synchronization of clocks in a DOCSIS - supported transmission network that uses a mobile network as an origin - destination point for transmitting data.

[0003] Originally, cable television (CATV) networks used dedicated coaxial cables to transmit television content to subscribers over long distances. However, current CATV transmission systems have replaced many of the coaxial cables with more efficient optical networks, constructing a hybrid transmission system in which cable content is transmitted over most of the distance from the headend using optical fibers and terminated at the subscriber over coaxial cables.

[0004] Historically, the headend also included a Cable Modem Termination System (CMTS), used to provide high-speed data services such as video, cable internet, and voice over internet protocols. A CMTS typically includes both an Ethernet interface (or other high-speed data interface) and an RF interface, allowing internet-originating traffic to be routed (or bridged) through the Ethernet interface, through the CMTS, and then to an optical RF interface connected to the cable company's hybrid fiber coaxial (HFC) system. Downstream traffic is transmitted from the CMTS to the subscriber's cable modem, while upstream traffic is transmitted from the subscriber's cable modem back to the CMTS. Many modern HFC CATV systems have combined the functionality of a CMTS with a Television Transmission System (EdgeQAM) in a single platform called a Converged Cable Access Platform (CCAP), using the Data over Cable Service Interface Specification (DOCSIS).

[0005] As networks expand and headends become increasingly congested with equipment, many content providers have in recent years been using distributed architectures to extend CMTS / CCAP capabilities across the entire network. This distributed architecture keeps data and video signals in digital format for as long as possible and extends the digital signals beyond the CMTS / CCAP into deeper parts of the network before converting them to RF. This distributed architecture works by replacing analog links between the headend and the access network with digital fiber (Ethernet / PON) connections.

[0006] One such distributed architecture is the Remote PHY (R-PHY) Distributed Access Architecture, in which the physical layer (PHY) of a conventional CMTS or CCAP is relocated by being pushed into the network's fiber nodes. Thus, while the CMTS / CCAP core performs higher-layer processing, the R-PHY devices in the nodes receive downstream data transmitted by the core, perform downstream PHY functions, convert the data from digital to analog RF signals, and distribute it to cable modems. In the reverse direction, the R-PHY devices within the nodes receive upstream RF data transmitted by the cable modems, convert the data from analog to digital, perform upstream PHY functions, and then optically transmit the digital data to the core. The CMTS and R-PHY devices use DOCSIS.

[0007] Furthermore, the DOCSIS architecture is employed as a mechanism for exchanging data in cellular communications between the core network of a cellular system (typically the global internet) and local networks communicating with, for example, cell towers. This data exchange is typically referred to as the "x-hall" portion of cellular communications, including backhaul, midhaul, and fronthaul. However, cellular networks also need to synchronize their backhaul, midhaul, and fronthaul portions with mobile base stations due to their own unique timing requirements. The DOCSIS network can also be used to provide synchronization to the mobile network. Therefore, problems arise with using the DOCSIS network as the x-hall for cellular communications, because DOCSIS-compliant devices have timing considerations that must be adhered to, especially when phase correction is required, such as after regaining connection to the master clock after a holdover, which may conflict with the timing and synchronization requirements of the mobile network. The DOCSIS network may be a conventional CMTS, remote PHY system, remote MAC-PHY system, or another method. In some cases, the CMTS may have a large time offset, such as 20 microseconds, and may require significant phase adjustment to relock to the time source when switching between two master clock sources.

[0008] Therefore, improved systems and methods for synchronizing network devices such as CMTS to the timing requirements of cellular networks are desirable. [Brief explanation of the drawing]

[0009] To better understand the present invention and to illustrate how it can be implemented, the following accompanying drawings are provided for reference as an example.

[0010] [Figure 1]Figure 1 shows an exemplary timing configuration for an R-PHY architecture, where the CCAP core is used as the timing grandmaster (GM) and the remote PHY device (RPD) is the timing slave.

[0011] [Figure 2] Figure 2 shows an exemplary timing configuration for the R-PHY architecture, in which both the CCAP core and its RPD are timing slaves to an external grandmaster.

[0012] [Figure 3] Figure 3 shows an exemplary CMTS architecture used as backhaul for a cellular network.

[0013] [Figure 4] Figure 4 shows an exemplary R-PHY architecture used as backhaul for cellular networks.

[0014] [Figure 5] Figure 5 shows the DTP delay parameters.

[0015] [Figure 6] Figure 6 shows one technique for synchronizing timing.

[0016] [Figure 7] Figure 7 shows one technique for synchronizing the timing in Figure 6.

[0017] [Figure 8] Figure 8 shows another technique for synchronizing timing.

[0018] [Figure 9] Figure 9 shows the technique for synchronizing the timing in Figure 8. [Modes for carrying out the invention]

[0019] For the purposes of this disclosure and the claims, the following terms are defined to facilitate understanding of the described and claimed subject matter. Master clock: A clock that transmits timing information for its clock to a slave clock and synchronizes that time with the time of the master clock. Slave clock: A clock that receives timing information from a master clock and synchronizes that time with the time of the master clock. Grandmaster clock: A clock that operates only as a master clock and is a source of time for a packet network. Ordinary clock: A slave clock having a single port that receives timing information from a master clock. Boundary clock: A clock that operates as both a slave and a master by having one port in a slave state that receives time from a master clock and one or more ports in a master state that distribute timing information to downstream slaves. Evolved Node B (eNB): A base station used in a cellular network that includes not only an antenna, a receiver, and a transmitter, but also historically includes a resource management function and a logical control function that were included in a separate base station controller, thereby enabling eNB stations to communicate directly with each other. Additionally, a next-generation base station (gNB), or other methods may also be used. Synchronous Ethernet (SyncE): A computer networking standard that easily transfers timing signals on an Ethernet physical layer for use by devices that require them. In this embodiment, SyncE timing should be derived from the same source as the grandmaster. Ethernet Equipment Clock (EEC): A slave clock of the SyncE protocol that receives synchronized data from an interface connected to an upstream master clock.

[0020] The DOCSIS Timing Protocol (DTP) specifies a way for a cable modem to provide timing references, particularly specific times with respect to another clock reference. To provide such proper synchronization, several common topologies are used. The DOCSIS Timing Protocol (DTP) is part of the CableLabs DOCSIS 3.1 specification, "Data Over Cable Service Interface Specification DOCSIS® 3.1, MAC and Upper Layer Protocol Interface Specification, CM-SP-MULPIv3.1-I21-201020", which is hereby incorporated by reference in its entirety.

[0021] For example, referring to FIG. 1, a first topology may include a CCAP core 2 synchronized with an RPD3, both connected together via a plurality of network switches 4. The RPD3 is then connected to one or more cable modems 5. Synchronization is achieved by a grandmaster clock 6 in the core 2 that sends timing information to a slave clock 7 in the RPD3. Although FIG. 1 shows only one RPD3 connected to the core 2, those skilled in the art will understand that a number of such RPDs may be connected to the core 2 simultaneously, and each RPD has a slave clock 7 that receives timing information from the grandmaster clock 6 in the core. The switches 4 can be any combination of participating and non-participating switches. Although FIG. 1 shows only the RPD3 connected to the core 2, the RPD3 may be omitted, and those skilled in the art will understand that each CM typically has a slave clock that receives timing information from the grandmaster clock 6 in the core, typically through the CCAP core 2.

[0022] Figure 2 shows a second topology 10 for providing synchronization between the CCAP core 14 and RPD 16 which are reconnected to one or more cable modems 18. However, unlike the system in Figure 1, a separate timing grandmaster device 12 provides timing information to both the CCAP core 14 and the RPD 16. Specifically, the timing grandmaster 12 has a first master port 20a connected to the boundary clock 22 of the CCAP core 14 and a second master port 20b connected to the slave clock 24 of the RPD 16. Thus, the boundary clock 22 is a slave to the grandmaster 12 but can be a master to the slave clock 24 of the RPD. Those skilled in the art will recognize that the respective clocks of the CCAP core 14 and the RPD 16 could both be connected to a single master port in the timing grandmaster device 12, and that the use of separate timing ports 20a and 20b in Figure 2 is simply used to more easily illustrate separate timing processes. The CCAP core 14 may be connected to the timing grandmaster 12 through one or more switches 26, while the RPD 16 may be connected to the timing grandmaster 12 through one or more switches 28. The switches 26 may be any combination of participating and non-participating switches. Again, although Figure 2 shows only one RPD 16 connected to the timing grandmaster 12, a number of such RPDs may be connected to the grandmaster 12 simultaneously, and each RPD has a slave clock 24 that receives timing information from port 20b in the grandmaster clock 12. Again, although Figure 2 shows only one RPD 16 connected to the timing grandmaster 12, such RPDs may be omitted, and each cable modem has a slave clock that receives timing information from port 20b in the grandmaster clock 12.

[0023] While the CMTS14 and / or RPD16 and / or cable modem 18 are locked to the timing groundmaster 12, no major problems occur. However, if any of the CMTS14 and / or RPD16 and / or cable modem 18 lose their connection to the timing groundmaster 12, problems arise. During such a holdover period, when one or more devices are not connected to the timing clock of the groundmaster 12, the unconnected devices will deviate in frequency and phase from the timing groundmaster 12 and from the other devices. The magnitude of this deviation depends on many factors, including the length of the holdover period, temperature fluctuations, and the performance of the internal oscillator. For example, in an RPD with a typical TCXO oscillator, the phase can deviate by as much as 1 ms even within an hour.

[0024] When connection to the timing groundmaster 12 is restored, the CMTS 14 and / or RPD 16 and / or cable modem 18 will need to measure their own phase offset from the groundmaster 12 and employ one of two methods to compensate for the offset and re-establish synchronization in both phase and frequency. Ideally, since any sufficiently large phase step adjustment would cause the connected cable modem to go offline, the RPD 16 and / or core 14 would only change its own frequency relative to the frequency of the groundmaster 12 to compensate for the phase offset until the timing, such as phase, becomes acceptable and then the frequency is set to the frequency of the groundmaster 12. However, to maintain DOCSIS compliance (e.g., CableLabs DOCSIS 3.1), the frequency adjustment must be small below a specified threshold, and therefore compensating for phase deviation by frequency adjustment can often take a long time to achieve the desired phase adjustment.

[0025] As mentioned above, the increasing consumption of mobile data is forcing mobile network operators to build small cell networks where all network traffic must be x-hole into the mobile core. Modern cell architectures employ LTE (Long Term Evolution) 4G or 5G standards or others, which provide wireless communication fast enough for mobile users to access the internet and experience services such as data, voice, and video from the mobile network. While such standards allow both time-division duplexing (TDD) and frequency-division duplexing (FDD), this disclosure assumes that the disclosed system employs FDD for illustrative purposes only.

[0026] While such backhaul has traditionally been implemented via fiber and microwave, hybrid fiber-coaxial (HFC) networks have been adopted as x-haul mechanisms due to their increased capacity, cost-effectiveness, and rapid deployment. Distributed cable access architectures (DAAs), including remote PHY (R-PHY) architectures or conventional non-distributed architectures, can also be adopted as x-haul mechanisms.

[0027] Referring, for example to Figure 3, the architecture 200 may include a CMTS 210 connected to a grandmaster clock 214 via a wide area network 216, such as the Internet, which then receives its timing information from a GNSS satellite 218. The CMTS 210 provides a transmission path for CATV services to a cable modem 220, which may include television channels, internet services, video on demand, and any other services provided to customers over the transmission network. In addition to providing conventionally transmitted CATV services over the transmission network, the architecture 200 also functions as an x-hole for a cell network 222, which includes an eNB 224 that relays communications between user devices, such as mobile phones, and the x-hole network to the Internet. In some embodiments, the eNB 224 may be part of a picocell, small cell, femtocell, macrocell, or something else.

[0028] The architecture 200 shown in Figure 3 implements a combination of IEEE 1588 and SyncE. The GM clock 214 may include a master port 230 that provides the IEEE 1588 timing protocol to a server port 232 on a device in the network 216. The same or a different device in the network may provide a master port 234 that provides the IEEE 1588 timing protocol to a slave port 236 on the CMTS 210. The CMTS 210 may use other timing protocols as desired. The GM clock 214 may also include an EEC clock 240 that provides SyncE to an EEC clock 242 on a device in the network 216. The same or a different device in the network may include an EEC clock 244 that provides SyncE to an EEC clock 246 on the CMTS 210. The network 216 may include one or more boundary clocks 250 synchronized with the grandmaster clock 214, and the CMTS 210 may include a boundary clock 252 synchronized with the grandmaster clock 214.

[0029] The cable modem 220 may include a master port 260 that provides the IEEE 1588 timing protocol to a slave port 262 on the eNB 224. The cable modem may also include an EEC clock 264 that provides SyncE to an EEC clock 266 on the eNB 224. The cable modem 220 may use other timing protocols for the eNB 224. The eNB 224 may include a standard clock 268.

[0030] The CMTS210 includes a master entity 270 that provides DOCSIS timing protocol (DTP) messages to a slave entity 274 on the cable modem 220. The CMTS210 also includes a master port 276 that provides the IEEE 1588 timing protocol to a slave port 278 on the cable modem 220. The cable modem 220 may use other timing protocols as desired.

[0031] Mobile base stations must achieve frequency synchronization of within 16 ppb (parts per billion) relative to an absolute time standard, and phase synchronization of up to 1.5 microseconds. This error budget represents the total error accumulated in the path between the time source, such as the grandmaster clock 214, and the mobile station's end application. IEEE-1588 also assumes symmetric network delays in the forwarding and inversion paths. Any delay asymmetry in the network effectively causes a phase error of 50% of the asymmetry value. Furthermore, phase synchronization is sensitive to packet delay variations within the network. Providing accurate frequency and phase on the network when using hybrid fiber coaxial networks in combination with DOCSIS as x-holes for mobile communications presents many challenges. Because DOCSIS is a packet-based network, it suffers from network asymmetry and packet delay variations. Additionally, DOCSIS presents additional challenges related to asymmetry due to the nature of DOCSIS upstream scheduling, packet delay variations due to upstream scheduling, low-bandwidth channels that increase packet transition times, and unknown delays and asymmetries in CMTS and cable modems.

[0032] Referring to Figure 4, a system similar to that shown in Figure 3 is applicable to a system including a remote PHY device (RPD). Furthermore, a system similar to the one illustrated in Figure 3 is applicable to a system including a remote MACPHY device (RMD).

[0033] As generally mentioned above, the DOCSIS Timing Protocol (DTP) is a set of extensions to the DOCSIS protocol, and DOCSIS cable modem and CMTS implementations are intended to support precise timing transmission over DOCSIS by utilizing OFDM extended timestamp 272 and the DOCSIS system's range setting capabilities. Frequency is addressed by coupling cable modem Ethernet timing to DOCSIS downstream symbol frequencies, while time is addressed by coupling cable modem PTP timestamp messages to DOCSIS extended timestamps. Time offsets and asymmetric corrections on DOCSIS segments are addressed by measurement, signal transmission, and range setting by providing time adjustment values ​​through DTP messages. Generally, DTP involves phase (e.g., time) synchronization. CMTS can be pinned to a northbound GM clock (directly or through a series of one or more boundary clocks) using conventional PTP messages. Cable modems can provide PTP master functionality to applications present on their network.

[0034] When using DOCSIS as an x-hole for mobile networks, one challenge is its ability to support two different timing applications with different timing requirements on a single clock residing within the CMTS. While DOCSIS has strict limitations on phase and frequency changes, mobile does not have the same limitations. In some cases, the CMTS may need to perform large phase and / or frequency adjustments to relock to its time source. These may occur, for example, when switching between two master clock sources, where there may be some time offset between them, such as 20 microseconds. Or, it may occur after a long holdover when there is no connection to the grand master clock and the phase of one of the devices is significantly deviated. When the connection to the master clock is restored or switched, the CMTS can only compensate for such phase offsets by slow frequency changes, as DOCSIS does not allow the execution of phase steps and is constrained to a limited frequency change rate of 10 ppb / s. Correcting large phase offsets is conventionally done using various techniques. One technique involves using a sufficiently large step to the DOCSIS extended timestamp. Another technique could allow the DOCSIS symbol rate to exceed the permitted frequency change rate, rapidly aligning the DOCSIS symbol rate and extended timestamps (and the cable modem clocks that depend on them) to mobile requirements. This could result in significant downtime across the entire DOCSIS system, as the cable modem may require a reset and / or range change. Another technique involves phase correction, which only adjusts the DOCSIS symbol rate within the DOCSIS limits. This could substantially prolong downtime on the mobile backhaul path until the phase is corrected to mobile requirements. A 50-microsecond phase offset could take up to 8 minutes to correct.

[0035] Referring to Figure 5, the illustrative diagram shows the parameters modeled and measured for DTP asymmetric calculations. Delay asymmetry in the data paths and PHYs of the CMTS and cable modems may need to be determined, modeled, and measured for each deployment. Delay asymmetry in the HFC (amplifier, etc.) is also measured. After these parameters are characterized, the main factor that may influence the time adjustment parameter (e.g., "t-cm-adj") is the change in range setting that can be reflected in t-tro.

[0036] It is preferable to use a DTP correction mechanism, which is a static parameter traditionally used to compensate for DOCSIS and HFC asymmetry, as a dynamic parameter that is also corrected to compensate for ingress timing offsets, such as those from PTP, that cannot be quickly compensated within the limits imposed by DOCSIS, rather than causing a malfunction. When large phase correction is required, a DOCSIS DTP master such as a CMTS, RPD, RMD, or other method determines whether a phase step (for large phase offsets) and / or a series of phase steps and / or frequency rotations up to the DOCSIS limit of 10 ppb / s can be used for synchronizing DOCSIS and DTP parameters. This allows the cable modem's PTP master clock or other timing output to jump to a time detected by the boundary clock slave of the CMTS (or other method) without resetting and / or changing the range of the cable modem.

[0037] Referring to Figures 6 and 7, a preferred technique for phase step recovery involves correcting the DTP timing adjustment parameter 700 (e.g., "t-cm-adj") by the phase error detected by the CMTS on its ingress PTP interface 710. This causes the cable modem's PTP master clock timestamp output 720 to "jump" and align with the detected CMTS ingress PTP, thereby effectively removing the timing error from the CM. To avoid actually jumping the DOCSIS timestamp, a "virtual" phase step 730 is introduced by incrementing the CMTS PTP ingress timestamp by the necessary offset, "interfering" with the PTP technique from which the phase error has been effectively removed. The DOCSIS symbol frequency and extended timestamp 740 are always slowly pushed out within the limits of the DOCSIS specification to recover the large phase offset and align it again with the PTP ingress timestamp. This can be achieved by dividing the large phase offset into small phase recovery corrections (e.g., 100 ns every 1-2 minutes), each of which can be completed relatively quickly by frequency changes. Each small phase recovery correction modifies both the "virtual" ingress phase step and the DTP time adjustment parameter.

[0038] Referring to Figures 8 and 9, for the recovery of frequency rotation, the DTP time adjustment parameter 800 can be corrected by the phase error detected by the CMTS on its ingress PTP interface 810. This causes the cable modem's PTP master clock timestamp output 820 to "jump" and align with the detected CMTS ingress PTP, effectively eliminating the CM output timing error. The DOCSIS clock slowly rotates within the limits of DOCSIS, aligning the DOCSIS timestamp back to the PTP ingress 840. The DTP time adjustment parameter 800 may be adjusted, for example, every few seconds, to compensate for the phase shift of the DOCSIS timestamp 840, and can be calculated by the amount of phase recovered by the average DOCSIS clock frequency offset over that period (or in another way). For example, if the DOCSIS clock averages an offset of 100 ppb over a 1-second interval to reduce a large phase offset, the DOCSIS timestamp will appear to have shifted by 100 ns. The DTP adjustment is also changed by 100ns over the same period and communicated to the cable modem via frequency DTP signal transmission (or by another means).

[0039] The present invention is not limited to the specific embodiments described, and it will be understood that modifications may be made to it so as to be interpreted in accordance with the principle of superiority law, including equivalent doctrines or any other principle that expands the legally enforceable scope of claims beyond their literal scope, without departing from the scope of the invention as defined in the appended claims. Unless the context indicates otherwise, a reference in the claims to the number of instances of an element, whether it is a reference to one instance or to multiple instances, requires at least a certain number of instances of that element, but is not intended to exclude claims, structures or methods having more instances of that element than described. The term “comprise” or its derivatives used in the claims is used in a non-exclusive sense and is not intended to exclude the presence of other elements or steps in the claimed structure or method.

Claims

1. A method for adjusting the phase error of a DOCSIS network, (a) The first device detects the phase error with respect to the time source of the DOCSIS network at the timing interface of the first device. (b) The first device modifies the DOCSIS timing protocol parameters based on the detected phase error. (c) The first device causes a modification of the timestamp output of the second device in a manner based on the DOCSIS timing protocol parameters, within the limits of the DOCSIS timing protocol regarding phase and frequency changes. (d) A method comprising the following: the DOCSIS timing protocol parameters are reduced over time after the modification.

2. The method according to claim 1, wherein the first device is a CMTS.

3. The method according to claim 1, wherein the second device is a cable modem.

4. The method according to claim 1, wherein the first device includes one or more devices of the R-PHY architecture.

5. The method according to claim 1, wherein the DOCSIS timestamp is modified over time as the DOCSIS timing protocol parameters are modified in the opposite direction.

6. The method according to claim 5, wherein the DOCSIS timestamp is modified at a rate within the range specified by the corresponding DOCSIS protocol.

7. The method according to claim 1, wherein the DOCSIS timing protocol parameters are t-cm-adj.

8. The method according to claim 1, wherein the timestamp output of the second device is a high-precision time protocol clock timestamp output.

9. The method according to claim 1, wherein the first device includes an R-MACPHY device.

10. The method according to claim 1, wherein the first device includes an RPD device.

Images