GENERATION OF DOCSIS CHANNEL PROFILES

Abstract

Dynamic generation of a DOCSIS (Data Over Cable Service Interface Specification) modulation profile. Conditions of a channel are examined to obtain a plurality of examination samples, which are data describing observed conditions of the channel. The plurality of examination samples are arranged in clusters to identify one or more clusters, each represented by one or more cluster centers. A dispersion value is determined for each cluster. The dispersion value is data describing a shape and size of the cluster with respect to the one or more cluster centers associated with it.For each cluster, a threshold modulation error (MER) is identified using the one or more cluster centers associated with it and the dispersion value determined for that cluster. The threshold MER identified for a respective cluster is converted into a bit load value. A modulation profile is generated for a respective cluster using the bit load value for that cluster.

Description

CLAIMING PRIORITIES

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 619,684, filed January 10, 2024, entitled "DOCSIS Channel Profile Generation," and U.S. Patent Application No. 19 / 015,172, filed January 9, 2025, entitled "DOCSIS Channel Profile Generation." The disclosures of both applications are hereby incorporated by reference in their entirety. FIELD OF THE INVENTION

[0002] Embodiments of the invention generally relate to the dynamic generation of a DOCSIS (Data Over Cable Service Interface Specification) modulation profile for an OFDM and / or an OFDMA (Orthogonal Frequency-Division Multiplexing; Orthogonal Frequency-Division Multiple Access) channel. BACKGROUND

[0003] DOCSIS (Data Over Cable Service Interface Specification) is a widely used industry standard for transmitting digital data over existing cable television (CATV) infrastructure. Version 3.1 of DOCSIS introduced the concept of an OFDM channel profile (Orthogonal Frequency-Division Multiplexing).

[0004] OFDM is a multiplexing technique used to simultaneously transmit multiple signals, or channels, over a single transmission medium, such that the signals forming the separate channels overlap. The overlapping signals do not interfere with each other because the signals are orthogonal, meaning that when a single signal is at its peak, the neighboring signals are at their zero point. OFDM provides greater data throughput by allowing signal overlap, unlike previous multiplexing approaches, such as frequency division multiplexing.

[0005] The profile for an OFDM channel can be used by a CATV operator to describe the information a cable modem must have to communicate over that OFDM channel. A CMTS (Cable Modem Termination System) can define one or more profiles for a specific OFDM channel. Each profile for a specific OFDM channel describes a different set of parameters, such as modulation order (commonly referred to as 'constellation' or bit load), FEC (Forward Error Correction), preamble, and guard interval, which define how data should be exchanged between a cable modem and the CMTS for that OFDM channel. As mentioned, a typical DOCSIS modulation profile identifies a bit load for each channel, which identifies how many bits are to be transmitted over the channel per unit of time.

[0006] The DOCSIS 3.1 specification states that up to 16 profiles can be defined for a specific OFDM channel. Version 3.1 of the specification further provides a way to assign a group of profiles to a cable modem and recommends ways to select the best profile for use when exchanging data with a particular cable modem over a specific OFDM channel.

[0007] Typically, each profile is assigned a letter, such as Profile A, Profile B, etc. Profile A is a common profile assigned to each cable modem for a specific OFDM channel, but the other profiles assigned for that specific OFDM channel may vary from cable modem to cable modem.

[0008] Each OFDM channel has its own specific set of profiles. For example, profile A on OFDM channel 1 is different from profile A on OFDM channel 2.

[0009] The parameters that describe an OFDM channel are defined in an OCD (OFDM Channel Descriptor) message, and each profile for an OFDM channel is defined in a DPD (Downstream Profile Descriptor) message. The OCD and DPD messages are sent on a PLC (PHY Link Channel) to all cable modems in the CATV system. When a cable modem initializes, it uses Profile A for a specific OFDM channel until instructed by the CMTS to use a different profile.

[0010] The DOCSIS 3.1 specification describes how the CMTS can obtain information regarding the SNR (signal-to-noise ratio) for a specific OFDM channel from a specific cable modem. The specification also provides means for the CMTS to request a specific cable modem to access a specific profile and report information regarding its SNR and FEC to the CMTS. The DOCSIS 3.1 CM-SP-PHY specification includes an algorithm that can be used by a CMTS to select a profile to be used in conjunction with a specific OFDM channel by a specific cable modem, based on information and statistics regarding SNR, FEC, and related parameters obtained from that cable modem. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Embodiments of the invention are illustrated by way of example and not limitation in the figures of the accompanying drawings, wherein like reference numerals refer to similar elements. In the drawings: Fig. 1 is a block diagram of a CCAP (Converged Cable Access Platform) platform in which an embodiment of the invention may be employed; Fig. 2 is a flowchart illustrating the steps of dynamically generating a DOCSIS (Data Over Cable Service Interface Specification) modulation profile for an OFDM (Orthogonal Frequency-Division Multiplexing) channel according to an embodiment of the invention; Fig. 3 is a diagram illustrating an exemplary graph of test results from many cable modems around a specific modulation error (MER), according to an embodiment of the invention; Fig. 4 is a diagram showing test results of many cable modems around a specific modulation error (MER) according to an embodiment of the invention; Fig. 5 is an illustration of the result of clustering study samples into a set of clusters according to an embodiment of the invention; Fig. 6 is a diagram illustrating how large the variance is between the study samples represented by a respective cluster, according to an embodiment of the invention; and Fig. 7 is a representation of a heatmap of different sets of optimized modulation profiles according to an embodiment of the invention. DETAILED DESCRIPTION

[0012] Embodiments relate to dynamically generating a DOCSIS modulation profile for an OFDM (Orthogonal Frequency-Division Multiplexing) channel. The modulation profiles generated using the techniques discussed herein may also be used with OFDMA (Orthogonal Frequency-Division Multiple Access) channels. In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a basic understanding of the embodiments of the invention described herein. It should be understood, however, that the embodiments of the invention described herein may be practiced without these specific details.In other instances, well-known structures and devices are shown in block diagram form or are discussed at a higher level in order to avoid unnecessarily obscuring the teachings of embodiments of the invention.

[0013] While DOCSIS modulation profiles have existed in the prior art, embodiments of the invention are directed to an innovative new and novel approach to generating them, offering a variety of improvements and advantages over the current state of the art. The assignment of the dynamically generated profiles of one embodiment to cable modems is not based on the physical location of the cable modems or any cable modem-related characteristics of the service provider's customers, but rather based on the periodic channel conditions, including interference and other interference patterns, that may be encountered by the cable modems to which the profile has been assigned.

[0014] Embodiments of the invention enable the dynamic generation of optimized DOCSIS modulation profiles that can be used in both the upstream (US) and downstream (DS) directions. While specific examples are discussed herein with respect to the US or DS directions, those skilled in the art will understand that these specific examples do not limit the generation or use of modulation profiles of an embodiment to either the US or DS directions. When discussing certain portions of the spectrum, different terms may be used to identify certain levels of granularity, but those skilled in the art will understand that these terms may have implications in the other direction. For example, a DOCSIS DS-OFDM bit load is provided across the channel frequencies at subcarrier resolution.The subcarrier resolution is either 25 kHz or 50 kHz, depending on the type of DOCSIS OFDM subcarrier spacing used. On the other hand, a DOCSIS upstream OFDMA bitload is provided across the channel frequencies at the minislot resolution of 400 kHz.

[0015] As another example, a DOCSIS DS-OFDM channel can consist of up to 7600 subcarriers. On the other hand, a DOCSIS US-OFDMA channel can consist of up to 237 minislots. Each OFDMA minislot consists of 8 or 16 subcarriers of 50 kHz or 25 kHz, respectively.

[0016] Before explaining certain embodiments of the invention, it is useful to consider the operational context in which certain embodiments may be used. Therefore, a brief discussion of upstream and downstream communications in a CCAP (Converged Cable Access Platform) platform will now be presented. UPSTREAM AND DOWNSTREAM COMMUNICATIONS ON A CCAP PLATFORM

[0017] Embodiments of the invention are directed to the dynamic generation of a DOCSIS (Data Over Cable Service Interface Specification) modulation profile for an OFDM and / or OFDMA upstream or downstream channel. As is well known to those skilled in the art of cable networks, DOCSIS is a widely used industry standard for transmitting digital data over existing cable television (CATV) infrastructure and is widely used in CCAP (Converged Cable Access Platform) deployments. CCAP is an industry-standard platform for transmitting video data and voice content.

[0018] Fig. 1 is a block diagram of a CCAP platform in which an embodiment of the invention may be used. Fig. The CCAP platform shown in Figure 1 includes a CCAP core 110 (CCAP Core), a remote PHY node 120, a plurality of cable modems 130, 132, 134, and network devices 140, 142 located at cable subscribers. Fig. The CCAP platform illustrated in Figure 1 can be implemented using a virtual CCAP platform running on hardware components including a COTS switch / router and one or more COTS computing servers (COTS = Commercial Off-the-Shelf). A commercial example of a virtual CCAP is CableOS™, available from Harmonic, Inc., San Jose, California.

[0019] Fig. 1 shows a downstream direction (DS direction) (i.e., from the CCAP core 110 to the cable subscribers' network equipment) and an upstream direction (US direction) (i.e., from the cable subscribers' network equipment to the CCAP core 110). According to embodiments of the invention, the generation of DOCSIS modulation profiles may be performed by a CCAP core 110 or at a location accessible thereto. The CCAP core 110 may send OCD (OFDM Channel Descriptor) and DPD (Downstream Profile Descriptor) messages to the cable modems 130, 132, 134 to inform these cable modems of the generated DOCSIS modulation profiles. The cable modems 130, 132, 134 may then use the generated DOCSIS modulation profiles when sending communications to the CCAP core 110.

[0020] The term CCAP core 110, as used broadly herein, refers to a CCAP core as described in the Remote PHY family of specifications known as the MHAv2 specifications and maintained by CableLabs®, Louisville, Colorado. The CCAP core 110 may be accessible via the Internet 102 (as described in Fig. 1) or via one or more private networks (not shown in Fig. 1) communicate.

[0021] The CCAP core 110 may correspond to an integrated CCAP that performs the functions of a CMTS and an EQAM (Cable Modem Termination System; Edge QAM). Alternatively, the CCAP core 110 may communicate with one or more devices that perform functions associated with a CMTS and / or an EQAM.

[0022] The CCAP core 110 is typically located at a headend and is used to provide high-speed data services to network devices. For example, Fig. 1 two network devices 140, 142, each of which exchanges data with the cable modem 132, which in turn exchanges data with the CCAP core 110 via the remote PHY node 120. As can be easily understood, a practical implementation of a CCAP platform includes many different remote PHY nodes 120, many different cable modems, and many different network devices; however, for reasons of simplicity and ease of explanation, a large number of these entities are not shown in Fig. 1 shown.

[0023] The remote PHY node 120, assisted by a remote PHY device 220, performs digital-to-analog conversion of downstream DOCSIS data, MPEG video, and out-of-band (OOB) signals, as well as analog-to-digital conversion of upstream data, video, and OOB signals. A non-limiting illustrative example of a remote PHY node 120 is the CableOS™ Ripple-1 remote PHY node, available from Harmonic, Inc., San Jose, California. Although only a single remote PHY node 120 is Fig. 1, however, in practical implementations, a large number of remote PHY nodes are in communication with the CCAP core 110.

[0024] The remote PHY node 120 is designed to be deployed outdoors near the physical locations of cable modems 130, 132, 134. The remote PHY node 120 consists of an outer housing designed to provide a hermetically sealed environment for the interior of the remote PHY node 120 to protect internal components from outdoor environmental factors such as moisture, water, dirt, and pressure fluctuations. Although only three DOCSIS cable modems are in Fig. 1, however, in practical implementations, there are varying numbers of cable modems served by a single remote PHY node 120. For example, it is not uncommon for more than 100 cable modems to be served by a single remote PHY node 120.

[0025] One such internal component embedded within a remote PHY node 120 is a remote PHY device 220. The remote PHY node 120 may include one or more remote PHY devices 220. The remote PHY device 220 is a computerized device that performs many of the functions encountered in converting downstream DOCSIS data, MPEG video, and out-of-band (OOB) signals from digital to analog, as well as upstream data, video, and OOB signals from analog to digital. A non-limiting illustrative example of a remote PHY device 220 is the CableOS™ Pebble-1 remote PHY device from Harmonic, Inc. IDENTIFYING DISTURBANCE INPUT PATTERNS

[0026] Cable modems, such as the cable modems 130, 132, 134 from Fig. 1 occasionally experience interference or other temporary impairments to certain portions of a CCAP platform's upstream spectrum. Such interference and / or impairments degrade the quality and capacity of the upstream channel, which may result in a noticeably negative user experience due to retransmissions and increased network latency. The occasional interference and / or impairments may be of such magnitude that they cause a specific cable modem to enter partial mode, which refers to a mode of the cable modem in which the cable modem no longer has the ability to transmit on one or more channels but can otherwise continue to operate.For example, if a specific cable modem experiences a prolonged impairment on a specific upstream channel, that cable modem may enter partial mode and be unable to use the upstream channel experiencing that impairment.

[0027] Approaches for automating and optimizing the generation of DOCSIS modulation profiles are presented here. DOCSIS modulation profiles are used during transmission over a DOCSIS channel. Advantageously, embodiments of the invention enable increased quality and capacity of the upstream spectrum of a CCAP by enabling dynamic generation and use of profiles based on current and recently observed spectrum conditions. As a result, embodiments of the invention require a smaller number of cable modems to operate in partial mode because the profiles used by the cable modems were generated with understanding and insight into the current operating conditions of the upstream spectrum.Embodiments of the invention usefully enable generation and use of an optimized DOCSIS modulation profile for a channel such that the current SNR ratio of that channel meets the minimum requirements of that DOCSIS profile (SNR = signal-to-noise ratio).

[0028] Fig. Figure 2 is a flowchart illustrating the steps of dynamically generating a DOCSIS (Data Over Cable Service Interface Specification) modulation profile according to one embodiment of the invention. The modulation profile may be, for example, an OFDM (Orthogonal Frequency-Division Multiple Access) modulation profile or an OFDMA (Orthogonal Frequency-Division Multiple Access) modulation profile. The steps of Fig. 2 are explained with reference to a simplified example of an embodiment including a two-minislot OFDMA DOCSIS channel. As is well known to those skilled in the art, a minislot of an OFDMA channel is a short, repetitive, scheduled time-frequency segment allocated to a specific cable modem during which that specific cable modem is permitted to perform upstream transmission over that specific channel.

[0029] Although the following explanation of the steps of Fig. 2 is given largely with reference to a concrete example involving a modulation profile used in the upstream direction, it should be clear to those skilled in the art that the steps of Fig. 2 can also be used to generate a modulation profile for downstream use.

[0030] Initially, at step 250, to generate a set of profiles for a single channel (hereinafter referred to as "the channel being profiled"), the CCAP core 110 examines conditions of the channel being profiled to obtain study samples. As used herein, the term "study sample results" refers to data collected by the CCAP core 110 that describe conditions of the channel being profiled. Study samples may correspond to receiver MER vectors or SNR vectors. The CCAP core 110 may obtain study samples of the channel being profiled upstream by examining upstream transmissions (US transmissions) using a modulation error rate (MER), hereinafter referred to as "the reference MER."The CCAP core 110 may obtain downstream probe samples of the channel currently being profiled in a variety of ways, for example, by examining downstream transmissions (DS transmissions) sent by the CCAP core 110 to cable modems over the channel currently being profiled using Simple Network Management Protocol (SNMP) queries. Another approach that the CCAP core 110 may take to obtain downstream probe samples for the channel currently being profiled is to use OPT-REQ messages to instruct cable modems to report information regarding their RxMER (MER per subcarrier) to the CCAP core 110 in an OPT-RSP message.Other approaches to measuring the quality of a channel may be used without departing from the spirit and scope of the embodiments of the invention described herein.

[0031] The DOCSIS 3.1 specification requires that a certain minimum average modulation error ratio (MER), expressed in dB, be supported at the receiver for a given channel (abbreviated as 'RxMER'). The MER value is the ratio, in decibels, of the average symbol power to the average error power. The MER provides a measure of the "blur" or scattering of clouds of graphically represented symbol points of a constellation. The DOCSIS 3.1 specification also requires that a certain modulation bit load (bits per symbol) / modulation magnitude be supported by the receiver for a given channel.

[0032] For clarification, Table 1 lists the average RxMER required by the DOCSIS 3.1 specification as MER thresholds per modulation magnitude.

[0033] As shown in Table 1, the DOCSIS 3.1 specification requires that a transmission sent with a modulation magnitude of 512 has a MER of 30.5 dB.

[0034] In one embodiment of the invention, the probe samples collected at step 250 may be filtered for invalid or bad samples or data. It should be appreciated that the specific times or intervals at or over which a collection of probe samples is performed is of less importance than representing the variety of channel conditions as accurately as possible. IDENTIFYING CLUSTERS WITHIN THE STUDY SAMPLES

[0035] After examining channel conditions to obtain study samples for the channel currently being examined for profiling, at step 252 the CCAP core 110 performs clustering of the study samples obtained at step 250 around the reference MER for the purpose of identifying one or more clusters of study samples. Step 252 will be described with reference to Fig. 3, which is a diagram illustrating an exemplary plot of investigation samples for many cable modems around a reference MER according to an embodiment of the invention. In the example of Fig. 3, the test samples for many cable modems are represented by “X”, and the unit of both the X-axis and the Y-axis is dB. Fig. Figure 3 shows a threshold for two different minislots, labeled minislot 1 and minislot 2. Real OFDMA samples have hundreds of minislots, resulting in points of high dimensionality. To provide a graphical representation of the clustering process of step 252, Fig. 3 shows a simplified example that only includes two minislots and thus two dimensions, so that the clustering can be represented in a two-dimensional format that facilitates visualization of clusters.

[0036] In one embodiment, at step 252, a K-means clustering algorithm, which is widely known to those skilled in the art, may be used to identify one or more cluster centroids that may be used as a proxy or representation of a respective cluster. For example, in performing the simplified example as illustrated in Fig. 3, a center point 202 of a cluster 204 may be determined in step 252 using a K-means clustering algorithm. A K-means clustering algorithm may receive as input N points in a certain dimension (such as the study samples of cluster 204) and return as output K points corresponding to the centers of these N points such that the accumulated (Euclidean) distance of the points to their assigned centers is approximately minimal. The purpose of determining the K center points using the K-means clustering algorithm at step 252 is to obtain an accurate representation of the N study samples.In fact, the goal is to cluster the N study samples, which may reflect noise and / or interference, using a smaller number of K center points that represent most (or at least one representative sample) of the N study samples. Note that the use of a similar clustering algorithm or a variant thereof, such as, but not limited to, DBSCAN, GMM with other distance metrics, such as, but not limited to, Manhattan, Minkovsky, or an SNR-quantized dependent distance, may be performed by embodiments of the invention without departing from the spirit or scope of the invention.

[0037] In this example, it is assumed that the channel over which the probe samples were obtained is stable, and as such, the probe samples are centered relatively closely around a center point 202 in a single cluster 204. Note that if the probe results were not centered relatively closely around the center point 202, then multiple clusters in Fig. 3. As explained below, embodiments of the invention perform dynamic generation of a profile for each cluster of profile samples identified in step 252.

[0038] Fig. 4 is a diagram illustrating test results for many cable modems according to an embodiment of the invention. Fig. Figure 4 shows over 35,000 probe samples collected over the course of a day. The probe samples can be obtained at regular intervals from all cable modems on a specific upstream port (or downstream port). Probe samples can also be obtained from all cable modems on a specific virtualized port, which corresponds to several physical channels sharing a common profile scheduler.

[0039] In the Fig. 4, each investigation sample describes an upstream channel that has 273 minislots; accordingly, in the example of Fig. 4 Each probe sample is a vector of 273 numbers corresponding to MER measurements of a particular minislot. A lower value for a probe sample vector means that a higher average error rate was measured for the region of the spectrum associated with that value (such as a minislot in the case of the upstream direction), and consequently, that the minislot only supports lower modulation or less data, i.e., a lower number of bits per minislot. On the other hand, a higher value for the probe sample vector means that the region of the spectrum associated with that value was measured as relatively 'clean', and consequently, that the region of the spectrum can support higher modulation (i.e., more data or a larger number of bits per minislot). In Fig. 4, the scale is given in dB, and the higher the vector values, the whiter the grayscale coloring. Embodiments of the invention can be used with study samples having any number of minislots; for example, study samples can be represented as a vector of M values of MER in dB representing K clusters of channel conditions.

[0040] Fig. Figure 5 is a representation of the result where study samples were organized into a set of clusters, according to one embodiment of the invention. For example, the over 35,000 in Fig. 4 shown sample values by clustering into the six in Fig. 5 using a K-means clustering algorithm. As shown in Fig. 5, the six clusters are numbered from 0 to 5, as identified by the Y-axis, whereas the minislot numbers are identified by the X-axis. The grayscale coloring shown for each region of the spectrum for each cluster represents the corresponding MER measurement, as shown in the vertical scale on the right side of Fig. 5, expressed in dB.

[0041] As a result of performing step 252, for each identified cluster, a set of one or more center points is identified to represent a respective single cluster. DETERMINE HOW DISTURBANCES AFFECT A PARTICULAR CLUSTER OF STUDY SAMPLES

[0042] After one or more cluster centers have been calculated for each cluster, in step 254, the CCAP core 110 determines a dispersion value for each cluster determined in step 252. The dispersion value of a cluster is data describing the shape and size of the cluster, which in turn reflects how noise and / or interference in the spectrum has affected the study samples. The greater the variation in the noise and / or interference present in the spectrum when the study samples were collected, the greater the dispersion of the study samples in a cluster; and conversely, the smaller the variation in the noise and / or interference present in the spectrum when the study samples were collected, the smaller the dispersion of the study samples in a cluster. Note that Fig. 3 is a simplified example and that embodiments of the invention may operate with any number of examination samples.

[0043] One must consider the shape and / or the dispersion of a particular cluster. For this purpose, the Fig. 3, the dispersion of cluster 204 in both the X and Y dimensions is determined by identifying the outer boundaries of the perimeter of each cluster in these directions. In the example of Fig. 3 there is only one cluster, namely cluster 204; consequently, the CMTS only needs to perform step 254 once to obtain a single dispersion value for cluster 204 in the Fig. 3 shown example. However, if in Fig. 3 several clusters were represented, then when performing step 254 a separate dispersion value would be generated for each cluster.

[0044] It is not possible for the CCAP core 110 to generate a profile for the cluster 204 associated with the center point 202 based solely on the MER associated with the center point 202, since roughly half of the probe sample results grouped in the cluster represented by the center point 202 are associated with a MER lower than that of the center point 202 and therefore could not be sent with the calculated profile.The MER associated with center point 202 represents average channel conditions for which approximately half of the probe samples were plotted on either side of the center point at step 252; therefore, any modulation profile requiring the channel conditions associated with center point 202 of cluster 204 would be too stringent for too many of the probe samples of cluster 204 to achieve successful transmission using that profile. Accordingly, one must consider how noise and / or interference has affected the distribution of the probe samples in cluster 204. The greater the magnitude of noise and / or interference in the spectrum over which the probe samples were obtained, the more widely the probe samples will be distributed within the cluster, which in turn causes the cluster to become wider and / or longer.Thus, the wider and / or longer the graphical representation of a cluster is determined, the lower the dispersion value for the cluster, as determined in step 254, must be to account for most of the samples assigned to that cluster.

[0045] One approach that can be used to calculate the dispersion value in step 254 is to calculate the dispersion of a specific cluster based on its variance or standard deviation of each dimension for the study samples assigned to a specific cluster. Another approach that can be used to calculate the dispersion value in step 254 is to calculate the variance or standard deviation of the difference between the samples and their assigned clusters. The calculations to determine the dispersion value are performed per minislot; as a result, in one embodiment, the results are a vector of dispersion values, where the length of the vector corresponds to the number of minislots in the channel. DETERMINING AN ADJUSTED OR MARGINALIZED MER FOR EACH CLUSTER

[0046] After obtaining a dispersion value for each cluster of study samples, in step 256, the CCAP core 110 calculates a threshold MER for each cluster of study samples. The threshold MER is the result of adjusting the MER value of the center points determined for that cluster based on the dispersion value for the cluster. In fact, in performing step 256, the CCAP core 110 subtracts some amount from the MER associated with the center points for each cluster based on the dispersion value, typically an amount equal to two or three times the standard deviation.The calculated threshold MER for a cluster is designed to be sufficient to allow the cable modems represented in the cluster to use the profile using the threshold MER, since the SNR is large enough to support most of the cable modems using this profile.

[0047] In one embodiment, the assumption is made that the MER measurements are normally distributed around the true MER center for that cluster. By setting the threshold MER to twice the standard deviation below the MER, approximately 95% of the time (as in the case of a normal distribution, where 95% of the samples fall within twice the standard deviation of the mean), the channel's SNR will be greater than the threshold MER, and therefore, sending the information with the specific bit load value (which matches this threshold MER via the lookup table) should result in the data arriving correctly at the other end.

[0048] Fig. 6 is a diagram illustrating how large the variance is between study samples represented by each cluster, according to one embodiment of the invention. Fig. Figure 6 shows a calculation of the standard deviation of sample values around their assigned cluster center per minislot. The vertical scale on the right side of Fig. Figure 6 shows a magnitude of variance for a given grayscale coloring. Fig. Figure 6 shows sections of the spectrum (e.g., 237 minislots as indicated by the x-axis) in lighter color at locations where the variance is large. When large variance is present, a profile should use a lower bit load value to allow more cable modems using a lower SNR channel to be able to use that profile (SNR = signal-to-noise ratio). Generating a profile for a particular cluster using a bit load value or a modulation magnitude for that cluster

[0049] At step 258, the CCAP core 110 converts the threshold MER associated with a respective cluster into a bit load value. In one embodiment, this may be done using a bit load lookup table. The bit load value (i.e., the number of bits per symbol / minislot that can be transmitted over the channel in a particular unit of time) is also known as the modulation magnitude. The bit load lookup table may be maintained in a location accessible by the CCAP core 110. The bit load lookup table is used to identify a specific bit load value for a specific marginalized MER. By looking up the bit load lookup table, the higher the marginalized MER, the larger the resulting bit load value.The relationship between the marginalized MER and the bitload value in the bitload conversion table is not linear but quantized in steps.

[0050] After obtaining a bit load value, the CCAP core 110 generates a profile for the channel currently being profiled using the determined bit load value. If multiple clusters have been identified, the CMTS is able to select a respective profile generated from one of the multiple clusters in real time for each cable modem, depending on the sample value it transmits.

[0051] Fig.7 is a representation of a heatmap of different sets of optimized modulation profiles generated according to an embodiment of the invention. Heatmaps 710 and 720 show different sets of optimized profiles. Because profiles may be generated using embodiments having a large number of different bitload "bands" (i.e., a continuous list of minislots having the same bitload), and because such information must be transported downstream to the cable modems to inform them which minislot has which bitload, a large number of different bands results in wasted downstream bandwidth. Therefore, advantageously, certain embodiments limit the number of bands to a fixed number, e.g., 10, by merging similar bands together in such a way as to minimize the impact of bitload.The final result of this compaction or band limiting is shown in heatmap 720, where the number of bands in each of the profiles is limited to 10, which is not the case in heatmap 710.

[0052] After the CCAP core 110 generates a profile for the channel currently being examined for profiling, it can inform the cable modems of the profile so that the cable modems can use that channel. Note that profiles generated using embodiments of the invention can be used in conjunction with an upstream channel or with a downstream channel.

[0053] Advantageously, embodiments of the invention enable profiles to be assigned to cable modems not based on the physical location of the cable modems or any characteristics of the service provider customers associated with the cable modems, but rather based on the periodic channel conditions, including interference noise or other interference patterns, that the cable modems to which the profile has been assigned may experience. Furthermore, embodiments of the invention are computationally less complex than previous approaches to generating modulation profiles, since the complexity of embodiments for N cable modem profiles is C(N), as opposed to a computational complexity of C(N 3), which was common for previous approaches to generating modulation profiles. Another advantage that embodiments have over the prior art is that transmit samples of channel conditions are obtained throughout the day. This makes it possible to identify which channel bitloads can be used and to design the modulation profiles accordingly based on this information. Thus, modulation profiles generated using embodiments of the invention can specify optimal bitloads based on observed characteristics of the channel. Furthermore, if channel conditions recur in a systematic manner or pattern, then the modulation profile previously generated for the same or similar channel conditions can be used whenever the presence of such channel conditions is either detected or suspected. EXTENSIONS

[0054] The term "non-transitory computer-readable storage medium," as used herein, refers to any tangible, physical medium that participates in the persistent storage of instructions or operating instructions that can be provided to a processor for execution. Further details regarding the operation of a non-transitory computer-readable storage medium can be found in U.S. Patent Application No. 11,212,590, published December 28, 2015, entitled "Multiple Core Software Forwarding," the disclosure of which is hereby incorporated by reference in its entirety.

[0055] In the foregoing description, embodiments of the invention have been described with reference to numerous specific details that may vary from implementation to implementation. Thus, the sole and exclusive indicator of the scope of the invention, and what is intended by the applicant as the scope of the invention, is the set of claims that emerge from this application, in the specific form in which such claims are drafted, including any subsequent amendment. Any definitions expressly set forth herein for terms included in these claims are intended to govern the meaning of such terms as used in the claims. Therefore, no limitation, element, property, feature, advantage, or attribute not expressly recited in a claim should in any way limit the scope of such claim.The description and drawings are therefore to be understood as explanatory and not restrictive. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] US 63 / 619,684

[0001] US 19 / 015,172

[0001] US 11,212,590

[0054]

Claims

[1] A method for dynamically generating a DOCSIS modulation profile (DOCSIS = Data Over Cable Service Interface Specification), comprising: examining conditions of a channel to obtain a plurality of examination samples, the plurality of examination samples being data describing observed conditions of the channel; Arranging the plurality of study samples into clusters to identify one or more clusters, each of the one or more clusters being represented by one or more cluster centers, determining, for each of the one or more clusters, a dispersion value, the dispersion value being data describing a shape and size of the cluster with respect to the one or more cluster centers associated therewith; Identifying, for each of the one or more clusters, a threshold modulation error (MER) using the one or more cluster centers associated therewith and the dispersion value determined for that cluster, wherein the threshold MER identifies a specific adaptation to the modulation error of the one or more cluster centers; and Converting, for each of the one or more clusters, the threshold MER identified for that cluster into a bit load value and generating a modulation profile using the bit load value for that cluster. [2] The method of claim 1, wherein the modulation profile is an upstream modulation profile. [3] The method of claim 1, wherein the modulation profile is a downstream modulation profile. [4] The method according to any one of claims 1 to 3, wherein the modulation profile is one or more of an OFDM modulation profile and an OFDMA modulation profile (OFDM = Orthogonal Frequency-Division Multiplexing; OFDMA = Orthogonal Frequency-Division Multiple Access). [5] Method according to one of claims 1 to 4, further comprising: Assigning the modulation profile to a specific cable modem based on observed periodic channel conditions of that channel experienced by the specific cable modem. [6] The method of any one of claims 1 to 5, wherein the dispersion value determined for each of the one or more clusters is a vector of dispersion values, the length of the vector corresponding to a number of minislots in the channel. [7] A method according to any one of claims 1 to 6, wherein generating the modulation profile includes limiting how many bit load bands are present in the modulation profile, wherein a bit load band is a continuous sequence of minislots in the channel that have been assigned the same bit load value. [8] A transitory or non-transitory computer-readable storage medium storing one or more sequences of instructions for dynamically generating a DOCSIS modulation profile (DOCSIS = Data Over Cable Service Interface Specification) which, when executed, cause operations to be performed which include the operations of any one of claims 1 to 7. [9] Apparatus for dynamically generating a DOCSIS modulation profile (DOCSIS = Data Over Cable Service Interface Specification), comprising: one or more processors; and one or more transitory or non-transitory computer-readable storage media storing one or more sequences of instructions which, when executed, cause operations to be performed which include the operations of any one of claims 1 to 7.

Images