U.S. patent application number 11/982671 was filed with the patent office on 2008-03-13 for establishment of multiple upstream docsis logical channels based upon performance.
This patent application is currently assigned to Terayon Communication Systems, Inc.. Invention is credited to Yehuda Azenko, Robert James Fanfelle, Selim Shlomo Rakib.
Application Number | 20080062889 11/982671 |
Document ID | / |
Family ID | 34633879 |
Filed Date | 2008-03-13 |
United States Patent
Application |
20080062889 |
Kind Code |
A1 |
Azenko; Yehuda ; et
al. |
March 13, 2008 |
Establishment of multiple upstream DOCSIS logical channels based
upon performance
Abstract
A process for grouping cable modems by type/modulation profile
and/or throughput ability into different logical groups. Each
logical group is commanded to transmit on an upstream which has a
burst profile set to effectively use the throughput ability of the
cable modem. Some species monitor initial ranging bursts and
separate CMs with inadequate power onto a lower throughput
upstream. Some species monitor post registration data transmissions
for various factors and subdivide groups into subgroups of
overperformers and underperformers.
Inventors: |
Azenko; Yehuda; (San Jose,
CA) ; Fanfelle; Robert James; (Redwood City, CA)
; Rakib; Selim Shlomo; (Cupertino, CA) |
Correspondence
Address: |
THELEN REID BROWN RAYSMAN & STEINER LLP
P. O. BOX 640640
SAN JOSE
CA
95164-0640
US
|
Assignee: |
Terayon Communication Systems,
Inc.
Horsham
PA
19044
|
Family ID: |
34633879 |
Appl. No.: |
11/982671 |
Filed: |
November 2, 2007 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10729179 |
Dec 6, 2003 |
|
|
|
11982671 |
Nov 2, 2007 |
|
|
|
Current U.S.
Class: |
370/252 ;
370/255 |
Current CPC
Class: |
H04L 41/0816 20130101;
H04L 12/2801 20130101; H04L 41/083 20130101; H04L 41/0893 20130101;
H04L 41/0896 20130101 |
Class at
Publication: |
370/252 ;
370/255 |
International
Class: |
G06F 11/30 20060101
G06F011/30; H04L 12/28 20060101 H04L012/28 |
Claims
1. A process for optimizing transmission speeds on a distributed
transmission system which can support multiple upstream channels or
logical channels simultaneously, comprising: 1) gathering data
about each cable modem (CM) in a group of CM coupled to a cable
modem termination system (CMTS) through a distributed transmission
system; 2) dividing said group of CMs up into logical groups based
upon CM type and/or throughput ability; 3) creating an upstream
channel or logical channel on said distributed transmission system
for each logical group of CMs, each upstream channel having
transmission characteristics optimized for a particular logical
group of modems; and 4) assigning the modems in each logical group
to the upstream channel created for that logical group.
2. The process of claim 1 further comprising the steps of
monitoring the byte error rate of transmissions from each CM, and
if the byte error rate of any CM becomes too high or too low
relative to underperformance and overperformance standards,
respectively, sending a message to said CM whose byte error rate
has become too high or too low causing each said CM which is
overperforming or underperforming to switch to an upstream channel
with a burst profile which is compatible with the CM modem type and
suitable for more efficient communications of digital data between
said CMTS and said CM.
3. The process of claim 1 further comprising the steps of
monitoring the packet error rate of transmissions from each CM, and
if the packet error rate of any CM becomes too high or too low
relative to underperformance and overperformance standards,
respectively, sending a message to said CM whose packet error rate
has become too high or too low causing each said CM which is
overperforming or underperforming to switch to an upstream channel
with a burst profile which is compatible with the CM modem type and
suitable for more efficient communications of digital data between
said CMTS and said CM.
4. The process of claim 1 further comprising the steps of
monitoring the signal-to-noise ratio (SNR) of transmissions from
each CM, and if the SNR of any CM becomes too high or too low
relative to underperformance and overperformance standards,
respectively, sending a message to said CM whose SNR has become too
high or too low causing each said CM which is overperforming or
underperforming to switch to an upstream channel with a burst
profile which is compatible with the CM modem type and suitable for
more efficient communications of digital data between said CMTS and
said CM.
5. The process of claim 1 further comprising the steps of
monitoring the received power of transmissions from each CM, and if
the received power of any CM at said CMTS becomes too high or too
low relative to underperformance and overperformance standards,
respectively, sending a message to said CM whose received power has
become too high or too low causing each said CM which is
overperforming or underperforming to switch to an upstream channel
with a burst profile which is compatible with the CM modem type and
suitable for more efficient communications of digital data between
said CMTS and said CM.
6. The process of claim 1 wherein step 1 comprises gathering data
about each modem's throughput ability by monitoring post
registration upstream CM data transmissions and determining the
value for one or more of a plurality of factors that indicate
whether each said CM is overperforming or underperforming the burst
profile and throughput ability of the upstream channel upon which
said CM is transmitting, said factors including RS codeword error
rate, SNR, received power, bit error rate, byte error rate and/or
packet loss rate, and creating and directing said overperforming
CMs to transmit upstream on one or more new upstream channels with
burst profiles which are suitable for more efficient communication
upstream by said overperforming CMs, and creating and directing
said underperforming CMs to transmit upstream on one or more new
upstream channels with burst profiles which are suitable for more
efficient communication upstream by said underperforming CMs.
7. The process of claim 1 wherein step 1 comprises gathering data
about each modem through a registration process and wherein step 2
comprises dividing modems into logical groups by modem type as
learned from said registration process, and wherein step 1 further
comprises gathering data about each modem's throughput ability by
monitoring post registration data transmissions and determining the
value for one or more of a plurality of factors that indicate
whether said modem is overperforming or underperforming the burst
profile and throughput ability of the upstream upon which said
modem is transmitting, and wherein step 2 further comprising
subdividing any logical group with one or more modems which are
overperforming or underperforming into overperforming and
underperforming logical subgroups, and wherein step 3 further
comprises creating one or more upstream channels with burst
profiles tailored to the throughput ability of said overperforming
modems and wherein step 4 further comprises assigning said
overperforming modems to an upstream channel with a burst profile
tailored to the throughput ability of said overperforming modem(s),
and wherein step 3 further comprises creating one or more upstream
channels with burst profiles tailored to the throughput ability of
said underperforming modems, and wherein step 4 further comprises
assigning said underperforming modems to an upstream channel with a
burst profile tailored to the throughput ability of said
underperforming modem(s).
8. The process of claim 1 wherein step 1 comprises gathering data
about each modem through a registration process and wherein step 2
comprises dividing modems into logical groups by modem type as
learned from said registration process with DOCSIS 1.0 modems in
one logical group and DOCSIS 1.1 modems in another logical group
and DOCSIS 2.0 modems in a third logical group operating in SCDMA
or ATDMA mode only, each logical group having created for it an
upstream having a burst profile suited to the throughput ability
and modulation profile of the modems in said logical group in step
3, and all modems in each logical group being assigned in step 4 to
an upstream having a burst profile tailored to the modems in said
logical group, and wherein step 1 further comprises gathering data
about each modem's throughput ability by monitoring post
registration data transmissions and determining the value for one
or more of a plurality of factors that indicate whether said modem
is overperforming or underperforming the burst profile and
throughput ability of the upstream upon which said modem is
transmitting, and wherein step 2 further comprising subdividing any
logical group with one or more modems which are overperforming or
underperforming into overperforming and underperforming logical
subgroups, and wherein step 3 further comprises creating one or
more upstream channels with burst profiles tailored to the
throughput ability of said overperforming modems and wherein step 4
further comprises assigning said overperforming modems to an
upstream channel with a burst profile tailored to the throughput
ability of said overperforming modem(s), and wherein step 3 further
comprises creating one or more upstream channels with burst
profiles tailored to the throughput ability of said underperforming
modems, and wherein step 4 further comprises assigning said
underperforming modems to an upstream channel with a burst profile
tailored to the throughput ability of said underperforming
modem(s).
9. The process of claim 1 wherein step 1 comprises gathering data
about each modem through a registration process and wherein step 2
comprises dividing modems into logical groups by modem type as
learned from said registration process with DOCSIS 1.0 modems in
one logical group and DOCSIS 1.1 modems in another logical group
and DOCSIS 2.0 modems grouped into a logical group operating in
SCDMA mode only and/or a logical group operating in ATDMA mode
only, each logical group having created for it an upstream channel
having a burst profile suited to the throughput ability and
modulation profile of the modems in said logical group in step 3,
and all modems in each logical group being assigned in step 4 to an
upstream having a burst profile tailored to the modems in said
logical group, and wherein step 1 further comprises gathering data
about each modem's throughput ability by monitoring post
registration data transmissions and determining the value for one
or more of a plurality of factors that indicate whether said modem
is overperforming or underperforming the burst profile and
throughput ability of the upstream channel upon which said modem is
transmitting, and wherein step 2 further comprising subdividing any
logical group with one or more modems which are overperforming or
underperforming into overperforming and underperforming logical
subgroups, and wherein step 3 further comprises creating one or
more upstream channels with burst profiles tailored to the
throughput ability of said overperforming modems and wherein step 4
further comprises assigning said overperforming modems to an
upstream channel with a burst profile tailored to the throughput
ability of said overperforming modem(s), and wherein step 3 further
comprises creating one or more upstream channels with burst
profiles tailored to the throughput ability of said underperforming
modems, and wherein step 4 further comprises assigning said
underperforming modems to an upstream channel with a burst profile
tailored to the throughput ability of said underperforming
modem(s).
10. The process of claim 1 wherein step 1 comprises gathering data
about each modem through an initial ranging process and a
registration process, and wherein step 2 comprises dividing modems
into logical groups by modem type as learned from said registration
process with DOCSIS 1.0 modems in one logical group and DOCSIS 1.1
modems in another logical group and DOCSIS 2.0 modems in a third
logical group operating in SCDMA mode only or ATDMA mode only, and
wherein each logical group has created for it an upstream having a
burst profile suited to the throughput ability and modulation
profile of the modems in said logical group in step 3, and wherein
all modems in each logical group being assigned in step 4 to an
upstream channel having a burst profile tailored to the modems in
said logical group, and wherein step 1 further comprises gathering
data about the received signal power and/or signal-to-noise ratio
(SNR) of initial ranging transmissions, and if any modem has
inadequate received signal power and/or signal to noise ratio after
a plurality of attempts to correct the problem, dividing said
modems into one or more low power and/or high power subgroups
and/or one or more low SNR and/or high SNR subgroups in step 2 and
creating one or more lower throughput, more robust upstream
channels for each low power and/or low SNR subgroup in step 3 and
sending messages to said modems that have low power and/or low SNR
directing said modems to switch to said one or more lower
throughput, more robust upstream channels, each lower throughput,
more robust upstream channel having a burst profile tailored to
achieve reliable communications with said modems in said low power
and/or low SNR subgroup assigned to said lower throughput, more
robust upstream channel such that registration can be completed,
and creating one or more higher throughput, less robust upstream
channels for each high power and/or high SNR subgroup in step 3 and
sending messages to said modems that have high power and/or high
SNR directing said modems to switch to said one or more higher
throughput, less robust upstream channels, each higher throughput,
less robust upstream channel having a burst profile tailored to
achieve reliable communications with said modems in said high power
and/or high SNR subgroup assigned to said higher throughput, less
robust upstream channel such that registration can be completed;
and wherein step 1 further comprises gathering data about each
modem's throughput ability by monitoring post registration data
transmissions and determining the value for one or more of a
plurality of factors that indicate whether said modem is
overperforming or underperforming the burst profile and throughput
ability of the upstream upon which said modem is transmitting, and
wherein step 2 further comprising subdividing any logical group
with one or more modems which are overperforming or underperforming
into overperforming and underperforming logical subgroups, and
wherein step 3 further comprises creating one or more upstream
channels with burst profiles tailored to the throughput ability of
said overperforming modems, and wherein step 4 further comprises
assigning said overperforming modems to an upstream channel with a
burst profile tailored to the throughput ability of said
overperforming modem(s), and wherein step 3 further comprises
creating one or more upstream channels with burst profiles tailored
to the throughput ability of said underperforming modems, and
wherein step 4 further comprises assigning said underperforming
modems to an upstream channel with a burst profile tailored to the
throughput ability of said underperforming modem(s).
11. The process of claim 10 further comprising the step of
continuing to monitor post registration data communications and
determining the values of one or more factors that indicate whether
a modem is overperforming or underperforming, and if any modem is
overperforming or underperforming its upstream channel's throughput
ability, creating a new logical subgroup and new upstream channel
for said modem and assigning said modem to transmit on said new
upstream channel, said new upstream channel having a burst profile
tailored to make efficient use of the throughput ability of said
modem.
12. A process for optimizing transmission speeds on a distributed
transmission system which can support multiple upstream channels
simultaneously and which has a plurality of cable modems coupled to
said distributed system, each having different upstream
transmission modes, comprising: transmitting one or more DOCSIS
downstreams from a cable modem termination system (CMTS); for each
DOCSIS downstream, transmitting: an upstream channel descriptor
message which establishes a DOCSIS 1.0 upstream; an upstream
channel descriptor message which establishes a DOCSIS 2.0 SCDMA or
DOCSIS 2.0 ATDMA upstream; receiving initial ranging bursts from
each of a plurality of cable modems (CM) and processing said bursts
to conduct initial training of each CM which transmitted an initial
ranging burst, and sending downstream messages to each CM to cause
any needed adjustments in power, frequency, timing and/or
equalization coeffients; receiving registration transmissions from
each CM which has successfully completed initial ranging, and
determining the type of each CM from registration data; creating a
separate logical group for all DOCSIS 1.1 CMs and one or more
separate 1.1 upstream channels for said DOCSIS 1.1 cable modems,
each said 1.1 upstream channel having a burst profile tailored for
the throughput ability of DOCSIS 1.1 CMs and linked to a downstream
to which a DOCSIS 1.1 CM is tuned, and sending downstream messages
to each DOCSIS 1.1 CM causing each DOCSIS 1.1 CM to switch to an
1.1 upstream channel linked to the downstream to which said CM is
tuned.
13. The process of claim 12 further comprising the steps of
monitoring the received power of each CM during initial training
thereof, and, for any CM which has inadequate received power after
a plurality of attempts to adjust transmit power of said CM have
failed to cause said CM's signal to arrive at said CMTS with
adequate received power, causing said CM to switch to an upstream
channel with a burst profile which is compatible with the CM modem
type and suitable for adequate communications of digital data
between said CMTS and CM despite said power shortfall problem.
14. The process of claim 12 further comprising the steps of
monitoring the received power of each cable modem in each logical
group during initial training, and, if the received power from a CM
is not adequate after a predetermined number of tries to adjust the
transmit power of said CM, then concluding said CM has a power
shortfall problem and either creating a low power, more robust
upstream channel with a burst profile suitable to allow adequately
reliable reception from said CM with said power shortfall problem
and sending a message to said CM with said power shortall problem
so as to cause said CM with said power shortfall problem to switch
to said low power, more robust upstream channel, or sending a
message to said CM with said power shortfall problem so as to cause
it to switch to a low power, more robust upstream channel which
already exists and which is compatible with the type of DOCSIS
modem said CM with said power shortfall problem is and which is
linked to a downstream to which said CM with said power shortfall
problem is tuned.
15. The process of claim 12 further comprising the steps of
monitoring the signal to noise ratio of transmissions from each CM
during initial training of said CM, and if the signal to noise
ratio of any CM is still unacceptable after multiple attempts to
complete initial training, sending a message to said CM whose
signal to noise ratio has become unacceptable causing said CM to
switch to an upstream channel with a burst profile which is
compatible with the CM modem type and suitable for adequate
communications of digital data between said CMTS and CM despite
said inadequate signal to noise ratio problem.
16. A process for optimizing transmission speeds on a distributed
transmission system which can support multiple upstream channels
simultaneously and which has a plurality of cable modems coupled to
said distributed system, each having different upstream
transmission modes, comprising: transmitting one or more DOCSIS
downstreams from a cable modem termination system (CMTS); for each
DOCSIS downstream, transmitting: an upstream channel descriptor
message which establishes a DOCSIS 1.0 upstream; an upstream
channel descriptor message which establishes a DOCSIS 1.1 upstream;
an upstream channel descriptor message which establishes a DOCSIS
2.0 upstream operating in SCDMA or TDMA mode; receiving initial
ranging bursts from each of a plurality of cable modems (CM) and
processing said bursts and sending downstream messages to each CM
to cause any needed adjustments in power, frequency, timing and/or
equalization coeffients; receiving registration transmissions from
each CM which has successfully completed initial ranging, and
determining the type of each CM from registration data and sending
any necessary downstream messages to any CM that is transmitting on
an upstream not having a burst profile optimized for the modulation
profile of said CM causing said CM to move to an upstream having a
burst profile optimized for the CM's modulation profile; monitoring
the received power of each CM during initial training thereof, and
determining any CM which has inadequate received power or
inadequate signal-to-noise ratio after a plurality of attempts to
adjust transmit power of said CM have failed to cause said CM's
signal to arrive at said CMTS with adequate received power or
adequate signal to noise ratio; sending a downstream message to
each CM which has inadequate received signal power or
signal-to-noise ratio to cause said CM to switch to a lower
throughput upstream channel with a burst profile which is
compatible with the CM type and suitable for adequate
communications of digital data between said CMTS and CM despite
inadequate received signal power or inadequate signal-to-noise
ratio.
17. The process of claim 16 further comprises the steps of
gathering data about each modem's throughput ability by monitoring
post registration data transmissions and determining the value for
one or more of a plurality of factors that indicate whether said
modem is overperforming or underperforming the burst profile and
throughput ability of the upstream upon which said modem is
transmitting, and further comprising the step of subdividing any
logical group with one or more modems which are overperforming or
underperforming into overperforming and underperforming logical
subgroups, and further comprising the step of creating one or more
upstream channels with burst profiles tailored to the throughput
ability of said overperforming modems, and further comprising the
step of assigning said overperforming modems to an upstream channel
with a burst profile tailored to the throughput ability of said
overperforming modem(s), and further comprising the step of
creating one or more upstream channels with burst profiles tailored
to the throughput ability of said underperforming modems, and
further comprising the step of assigning said underperforming
modems to an upstream channel with a burst profile tailored to the
throughput ability of said underperforming modem(s).
18. A process for optimizing transmission speeds on a distributed
transmission system which can support multiple upstream channels
simultaneously and which has a plurality of cable modems coupled to
said distributed system, each having different upstream
transmission modes, comprising: transmitting one or more DOCSIS
downstreams from a cable modem termination system (CMTS); for each
DOCSIS downstream, transmitting: an upstream channel descriptor
message which establishes a DOCSIS 1.0 upstream; an upstream
channel descriptor message which establishes a DOCSIS 2.0 upstream;
receiving initial training bursts from each of a plurality of cable
modems (CM) and deducing the cable modem type from the upstream
upon which each said initial training burst was received thereby
creating defacto logical groups of cable modems grouped by modem
type. receiving registration communications from each CM; after
registration, receiving upstream data transmissions from each CM;
monitoring one or more of the following parameters of communication
of data from each CM: the received power; the signal to noise
ratio; the bit error rate; the byte error rate; the Reed-Solomon
codeword error rate; and the packet error rate; if performance of
any CM becomes either too good or too bad, as measured by comparing
the monitored parameter for said CM to limits that establish what
performance level is too good or too bad, sending a message to said
CM to cause it to change to an upstream channel which has a burst
profile which is suitable for the CM's performance.
19. An apparatus comprising: any DOCSIS compatible cable modem
termination system having a control computer programmed to carry
out a process comprising the steps: receiving in a cable modem
termination system (CMTS) registration messages from each cable
modem coupled to said CMTS and determining the modem type from each
registration message; in said cable modem termination system
assigning each cable modem to a group based upon the modem type
with DOCSIS 1.0 compliant modems in a first group, DOCSIS 1.1
compliant modems in a second group, DOCSIS 2.0 compliant modems in
a third group; in said cable modem termination system, generating
and transmitting downstream to all said cable modems a plurality of
Upstream Channel Descriptor (UCD) messages, each UCD message
establishing a logical upstream channel to which one of the groups
of modems will be assigned and defining a burst profile for said
logical upstream channel which is appropriate for the group of
modems that will be assigned to transmit on that upstream logical
channel; generating in said cable modem termination system and
transmitting to each said cable modem which has registered a
message which tells each cable modem the upstream logical channel
to which it has been assigned.
20. The apparatus of claim 19 wherein said control computer is
further programmed to carry out the steps of: 1) monitoring the
received power from each cable modem in a group; 2) if the received
power at said CMTS is less than a required value for a cable modem,
commanding said cable modem to increase its transmit power in a
downstream message and repeating steps 1 and 2 until said cable
modem's transmitted signal arrives at said CMTS at the required
power; 3) if after reaching its maximum power available, a cable
modem's upstream transmissions still do not arrive at said CMTS at
the required power level, subdividing said cable modem into a
subgroup comprised of all modems of the same type and whose signals
do not arrive at said CMTS at the required power level despite each
said modem in said subgroup transmitting at the maximum available
power; 4) generating a UCD message for a new logical upstream to
which said modems in said subgroup will be assigned, said UCD
message establishing a burst profile for said new logical upstream
which is sufficiently robust in its forward error correction,
modulation type, symbol rate and/or other burst parameters such
that modems in said subgroup can transmit upstream with an
acceptable error rate.
21. The apparatus of claim 19 wherein said control computer is
further programmed to carry out the steps of: monitoring the
received power of each CM during initial training thereof, and, for
any CM which has inadequate received power after a plurality of
attempts to adjust transmit power of said CM have failed to cause
said CM's signal to arrive at said CMTS with adequate received
power, causing said CM to switch to an upstream channel with a
burst profile which is compatible with the CM modem type and
suitable for adequate communications of digital data between said
CMTS and CM despite said power shortfall problem.
22. The apparatus of claim 19 wherein said control computer is
further programmed to carry out the steps of: monitoring one or
more of the following parameters of communication of data from each
CM: the received power; the signal to noise ratio; the bit error
rate; the byte error rate; the Reed-Solomon codeword error rate;
and the packet error rate; if performance of any CM becomes either
too good or too bad, as measured by comparing the monitored
parameter for said CM to limits that establish what performance
level is too good or too bad, sending a message to said CM to cause
it to change to an upstream channel which has a burst profile which
is suitable for the CM's performance, and creating an new upstream
channel with suitable burst profile if necessary to which said CM
whose performance is too good or too bad may be changed.
23. The apparatus of claim 19 wherein said CMTS has line cards
which receive upstream signals from upstreams on a plurality of
data paths from different cable nodes, and wherein said line cards
have switches therein controlled by said CMTS to gate bursts from
said plurality of data paths to a controller, and wherein control
computer is further programmed to control said switch of any line
card coupled to a cable upon which a TDMA burst is expected to turn
on during a gap before said TDMA burst and turn off during a gap
after said TDMA burst and to keep all other switches of other line
cards turned off such that only said expected burst is gated
through to said combiner, and wherein said control computer is
further programmed to control the switch of any line card coupled
to a cable upon which an SCDMA burst is expected to turn on during
the ramp up of said SCDMA burst and to turn off during a ramp down
of said SCDMA burst and to control other switches of other line
cards to behave in the same way for expected SCDMA bursts on cables
coupled to said other line cards.
Description
CROSS REFERENCE TO RELATED APPLICATIONS
[0001] This application is a divisional of U.S. patent application
Ser. No. 10/729,179, filed on Dec. 6, 2003, entitled "Establishment
of Multiple Upstream DOCSIS Logical Channels Based Upon
Performance".
BACKGROUND OF THE INVENTION
[0002] Prior art cable modem termination systems (CMTS) in DOCSIS
systems establish upstream burst profiles, one for each DOCSIS
upstream. The burst profile for each upstream controls the data
throughput rate and other burst parameters that all cable modems
transmitting on that upstream must use. The burst profile must be
set to the lowest common denominator such that the least capable
cable modem (CM) on the system can comply with it. Typical systems
have hundreds or thousands of cable modems, some of which may be
legacy CMs or CMs purchased by customers that are not capable of
high throughput. This penalizes all CMs on the system to transmit
at a lower throughput than they are capable of if even one CM on
the same upstream is not capable of transmitting at high
throughput. This results in a lower average throughput on the
channel. In other words, in order to ensure that the worst CM has a
robust burst profile with adequate error correction protection, the
whole channel is forced to operate at a lower throughput.
Therefore, the upstream throughput of all CMs on the upstream will
be penalized because of the inadequacy of a small number of
CMs.
[0003] Therefore, what is needed is a way to get higher average
throughput by using different upstream logical channels, each with
its own burst profile. A way is needed of segregating the CMs in
the system into groups based upon throughput capability, each group
suitable for transmission on one of the upstream logical channels
having a matching burst profile. A way is needed to segregate CMs
with high throughput capability onto a logical channel with a less
robust burst profile but higher throughput. A way is needed to
segregate lesser CMs onto a logical channel with a more robust
burst profile yielding higher protection at the expense of lower
throughput.
SUMMARY OF THE INVENTION
[0004] The genus of the invention is defined by any process which:
[0005] 1) divides cable modems (CM) up into logical groups based
upon either the cable modem's DOCSIS type/modulation profile
(DOCSIS 1.0, DOCSIS 1.1, DOCSIS 2.0 ATDMA and DOCSIS 2.0 SCDMA)
and/or based upon throughput ability as measured by performance
characteristics; [0006] 2) creates a separate upstream logical
channel for each logical group, each upstream logical channel
having a burst profile tailored to the type and/or throughput
ability of the cable modems in the associated logical group; and
[0007] 3) sends downstream messages to the cable modems to cause
them to switch to an upstream logical channel having a burst
profile tailored to the logical group to which the cable modem
belongs.
[0008] Numerous species are within this genus. Several of the most
important species are defined below. DOCSIS only recognizes four
different types of upstreams: DOCSIS 1.x that supports no DOCSIS
2.0 TDMA features; mixed mode upstreams which support DOCSIS 1.x
and DOCSIS 2.0 TDMA bursts during different time intervals; DOCSIS
2.0 advanced time division multiplexing (ATDMA) only; and DOCSIS
2.0 synchronous code division multiplexing (SCDMA) only. It is the
CMTS that decides which modulation profile a CM operates in, but
DOCSIS 1.0 and 1.1 compatible CMs cannot be ordered to operate in
any 2.0 mode. The invention creates multiple logical groups based
upon type/modulation profile or throughput ability or both, and
creates an upstreams for each logical group, each upstream having a
burst profile tailored to the modem types/burst profile in the
logical group and/or the throughput ability of the modems in the
group. In some species, type/burst profile alone is used for the
grouping. In other species, throughput ability alone (within 1.x
and 2.0 groups with no distinction between 1.0 and 1.x and ATDMA
and SCDMA) is used to do the grouping, and in other species,
grouping is first done on modem type/modulation profile (hereafter
simply referred to as type) and then subgrouping within each group
is done based upon throughput ability. In other species, grouping
is first done on type with monitoring of SNR and/or received power
during initial training and subgrouping of CMs with power shortfall
or bad SNR to lower power, more robust channels so that
registration can be accomplished successfully. After registration
is accomplished in this species, the modem types are learned and
the modems are divided by type into separate logical groups, each
with its own logical channel. In still other species, after the
initial training and grouping into low power channels for some CMs
and after registration and grouping by type/modulation profile,
monitoring of data transmissions for communications quality is done
on each group, and any CMs who are performing too well or too
poorly for their upstream channels are moved to different channels
having burst profiles tailored to the CM's throughput ability.
[0009] The details of several of the most important species
follow.
[0010] 1) One species divides CMs up into logical groups strictly
by DOCSIS type. This species separates DOCSIS 1.0 and 1.1 CMs into
separate logical groups and divides DOCSIS 2.0 modems up into two
logical groups, one operating in advanced time division
multiplexing (ATDMA) and one operating synchronous code division
multiplexing (SCDMA). Each logical group of 2.0 modems operates on
a different physical upstream channel or on a different logical
channel on the same physical upstream channel. A physical upstream
channel is an upstream carrier modulated with digital data. A
logical upstream channel can be a time interval on a physical
upstream channel. Thus, all ATDMA 2.0 CMs can transmit during
different timeslots during a time interval devoted to logical group
1 on the physical upstream channel, and all SCDMA CMs can transmit
during a second interval devoted to 2.0 SCDMA bursts on the same
physical upstream channel. In this species, no monitoring of other
criteria like received power or signal to noise ratio is done.
Modem types are learned by DOCSIS registration messages. Generally,
the 2.0 CMs will use either SCDMA or ATDMA. In most cases, the CMTS
will not open two separate logical channels for ATDMA and
SCDMA.
[0011] 2) Another species divides CMs up into logical groups
strictly by DOCSIS type and, optionally, separates DOCSIS 1.0 and
1.1 CMs into separate logical groups and optionally divides DOCSIS
2.0 modems up into two logical groups, one operating in advanced
time division multiplexing (ATDMA) and synchronous code division
multiplexing (SCDMA). Each different type CM transmits upstream on
a different logical and/or physical channel. Then the modems in
each group are monitored for communications link quality as
measured by any one or more of the following factors:
signal-to-noise ratio (SNR), received signal power, bit error rate,
byte error rate, Reed-Solomon codeword error rate, and packet error
rate. CMs with one or more of the link quality factors indicating
the CM is operating too well or too poorly for its upstream channel
are then moved to an upstream channel or logical channel having a
burst profile (throughput and various forward error correction
factors) tailored to allow communication with an acceptable error
rate. Hereafter physical upstream channels and logical upstream
channels will simply be referred to as upstream channels since
either will suffice to segregate different performance level
modems. Generally, the 2.0 CMs will use either SCDMA or ATDMA. In
most cases, the CMTS will not open two separate logical channels
for ATDMA and SCDMA.
[0012] 3) Another species divides CMs up into logical groups
strictly by DOCSIS type with DOCSIS 1.0 and 1.1 CMs grouped in the
same logical groups and DOCSIS 2.0 modems grouped in the same
logical group. Then the modems in each group are monitored for
communications link quality as measured by any one or more of the
following factors: signal-to-noise ratio (SNR), received signal
power, bit error rate, byte error rate, Reed-Solomon codeword error
rate, and packet error rate. CMs with one or more of the link
quality factors indicating the CM is operating too well or too
poorly for its upstream channel are then moved to an upstream
channel having a burst profile (symbol rate, modulation type and
various forward error correction factors) tailored to allow
communication with an acceptable error rate.
[0013] 4) Another species divides CMs up into logical groups
strictly by DOCSIS type with DOCSIS 1.0 and 1.1 CMs grouped in
different logical groups on different upstream channels carried on
different physical channels, or on different logical channels, each
logical channel comprising a different time region of a DOCSIS 1.x
mixed mode channel, each separate logical channel having a burst
profile tailored for DOCSIS 1.0 or 1.1 (or 1.0 and 1.1 CMs can be
grouped together into one logical channel of DOCSIS 1.x). Likewise,
DOCSIS 2.0 modems in ATDMA would be grouped together and assigned
to a DOCSIS 2.0 ATDMA upstream channel, and DOCSIS 2.0 SCDMA modems
would be grouped in the same logical group and assigned to an
upstream logical channel having a burst profile tailored to the
modem type. During initial training or ranging of each modem, the
CMTS calculates SNR and/or monitors the received power. If after
the maximum number of attempts to correct inadequate received power
or inadequate SNR, acceptable received power or SNR is not
achieved, the CMTS moves the CMs with problems to a lower
throughput, more robust upstream channel so that effective
communications for registration purposes can be achieved with each
CM. Likewise, CMs that are overperforming their logical channels
are moved to a higher throughput upstream channel having a higher
symbol rate, more complex modulation or less forward error
correction.
[0014] 5) The same as any of the species defined above but
including post-registration monitoring. In this species, the DOCSIS
2.0 CMs which do not have a power shortfall problem or bad SNR such
as to be segregated during initial training into lower throughput,
more robust channels are all grouped into a single logical group
operating into SCDMA or ATDMA after registration. Subsequent
monitoring may cause the CMTS to conclude that certain CMs need to
have their burst profiles changed to handle changing conditions
such as falling SNR, burst noise, impulse noise etc. The CMTS will
then create a new physical upstream channel or a new upstream
logical channel (hereafter new upstream channel) with an
appropriate burst profile and send downstream message(s) to a CM to
be changed to the new upstream channel telling it which new burst
profile to assume and ordering it to change to the new upstream
channel. Specifically, after registration, the modems in each
logical group are monitored for communications link quality as
measured by any one or more of the following factors:
signal-to-noise ratio (SNR), received signal power, bit error rate,
byte error rate, Reed-Solomon codeword error rate, and packet error
rate. CMs with one or more of the link quality factors indicating
the CM is operating too well or too poorly for its upstream channel
are then moved to an upstream channel having a burst profile
(throughput and various forward error correction factors) tailored
to allow communication with an acceptable error rate. If an
upstream channel does not already exist having the needed burst
profile and linked to the appropriate downstream to which a CM is
listening, then an upstream with the appropriate burst profile will
be created. As many different burst profile upstreams as are
necessary are created. There are many factors which can be changed
in a burst profile. The logical channel parameters include: a) the
symbol rate which can be any one of 6 different rates from 160
ksym/sec to 5.12 Msym/sec in octave steps; b) the center frequency;
and c) the 1536-bit preamble superstring that is prepended to at
least some bursts; and d) the SCDMA channel parameters. The burst
profile transmission characteristics of a logical channel, in the
preferred embodiment, include: modulation (QPSK, 64 QAM, 128 QAM
etc.), differential encoding on or off; Trellis or TCM encoding on
or off; preamble length, preamble value offset; preamble type (QPSK
0 or QPSK1), RS (Reed-Solomon) error correction T value from 0 to
16 where 0 is no FEC bits to 16 for the maximum where the number of
codeword parity bytes is 2.times.T, RS codeword information bytes
length (16 to 253), scrambler seed, max burst length in minislots,
guardtime from 5 to 255 symbols for TDMA channels and 1 symbol for
SCDMA channels, last codeword (fixed or shortened), scrambler on or
off, byte interleaver depth, byte interleaver block size, SCDMA
spreading on or off, codes per subframe, and SCDMA interleaver step
size. User unique parameters may vary from user to user even when
on the same channel and same burst type and include such things as:
power level; offset frequency (defines center frequency of channel
to transmit on); ranging offset to achieve minislot boundary
alignment at CMTS (which also achieves upstream chip clock
alignment between the upstream chip clock generated at the CMTS and
the chip clock embedded in the received signal at the CMTS
receiver--a state which is referred to herein as "phase
coherence"), burst length in minislots if variable on the specified
channel (changes from burst to burst); and the transmit equalizer
coefficients (up to 64 coefficients specified by 4 bytes per
coefficient--2 real and 2 complex). The CMTS can create many
different upstreams by varying the values for these logical channel
parameters and burst profile transmission characteristics. The CMTS
will vary these burst profile characteristics based upon the
results of monitoring data transmissions of each CM after
registration and the initial grouping of a CM into a logical group
in a preferred species.
[0015] In species 1, the CMTS receives upstream data such as DOCSIS
registration messages from cable modems in a system and divides
them up into logical groups based upon modem type or throughput
ability. In this embodiment, all DOCSIS 1.0 and 1.1 CMs are
classified as different types and all DOCSIS 1.0 modems ordered by
the CMTS to operate in a 1.0 burst profile, and all DOCSIS 1.1
modems ordered by the CMTS to operate in a 1.1 burst profile, and
all DOCSIS 2.0 CMs are ordered to operate in SCDMA or ATDMA mode
and grouped into a single group. The process then creates a
separate upstream channel for each logical group of modems, each
upstream having transmission characteristics optimized for the
modem type in the group assigned to it. The process then assigns
each modem to the upstream having characteristics optimized for
that type of modem.
[0016] DOCSIS compatible CMTS can create upstream channels for each
logical group of CMs and control their initial training by sending
Upstream Channel Descriptor and MAP messages to all cable modems
listening to a downstream. The CMTS controls which CMs are assigned
to each upstream channel by sending DOCSIS channel change messages
or ranging response messages to CMs that need to change their
upstream channel. These channel change messages are used to assign
each DOCSIS compatible cable modem to an upstream channel having a
burst profile optimized for the modem type and mode of transmission
for the modems in the group. The same receiver can process all
logical channels where each logical channel is during a different
time interval of the same upstream physical channel.
[0017] In the preferred embodiment, subgroupings within each
logical group are made both initially at registration time based
upon SNR and/or received power criteria measured by the CMTS during
the initial ranging or registration process, and further
subgrouping are made later based upon criteria measured during
subsequent data communications. Subgroupings can be based upon: the
received signal-to-noise ratio; the packet loss rate; the bit error
rate; the byte error rate; the received signal power; the RS
codeword error rate or the cable node to which a cable modem is
coupled.
[0018] The apparatus needed to carry out the invention is a
conventional DOCSIS CMTS with its control process programmed to
carry out any of the processes within the above defined genus of
the invention.
BRIEF DESCRIPTION OF THE DRAWINGS
[0019] FIG. 1 is a flowchart of a process to group modems into
different groups based upon modem type only (no subgrouping) and to
establish a different upstream logical channel for each and to
command each CM to transmit upstream on a logical channel having a
burst profile suited to the modem types assigned to the logical
channel. The main point is that 1.0 and 1.1 CMs are separated into
separate logical channels having burst profiles optimized for the
type of CM in each channel.
[0020] FIG. 2, comprised of FIGS. 2A through 2D, represents species
4 in the summary of the invention. It is the preferred method to
initially set up the groups based upon modem type with SNR and
received power monitoring during initial training to establish one
or more low SNR or low power channels if necessary to achieve
successful registration even for CMs with power shortfall
problems.
[0021] FIG. 3, comprised of FIGS. 3A through 3E, is a flowchart of
the process of species 5 for post registration monitoring and
subgrouping based upon over or under performance.
[0022] FIG. 4 is a flowchart of the process of species 2 of the
invention wherein DOCSIS 1.1 and 1.0 modems are separated and all
2.0 modems are left as one group. Post registration monitoring for
overperformance or underperformance is used to create new subgroups
within each group.
[0023] FIG. 5 is a flowchart of the process of species 3 of the
invention wherein DOCSIS 1.1 and 1.0 modems are grouped in one
group with a 1.x burst profile logical channel, and all 2.0 modems
are left as one group. Post registration monitoring for
overperformance or underperformance is used to create new subgroups
within each group.
[0024] FIG. 6 represents the modified circuitry of a CMTS which
uses switches controlled by the CMTS to eliminate the aggregated
noise of the combiner.
[0025] FIGS. 7A and 7B show timing diagrams for switching in TDMA
and SCDMA logical channels, respectively, to avoid noise
aggregation.
DETAILED DESCRIPTION
[0026] A logical upstream channel is defined in the DOCSIS
specifications as a MAC entity identified by a unique channel ID
and for which bandwidth is assigned by an associated MAP message. A
single physical upstream channel is an upstream carrier signal
modulated with digital data at any one or a plurality of different
symbol rates so as to have any of four different bandwidths. A
single physical RF carrier may include multiple logical upstream
channels, each having a different burst profile. Each logical
upstream channel is described by an Upstream Channel Descriptor
Message (UCD). The UCD and MAP messages associate with an upstream
logical channel completely describe the logical channel.
[0027] A UCD message is a MAC message which is sent downstream by
the Cable Modem Termination System (CMTS) to a plurality of cable
modems (CMs) to define the characteristics of the logical upstream
channel to which the UCD message pertains. UCD messages include:
modulation rate; frequency; preamble pattern; extended preamble
pattern; first burst descriptor; second burst descriptor; third
burst descriptor; etc. The UCD message for each upstream logical
channel must be transmitted by the CMTS at a periodic interval to
define the characteristics of the upstream logical channel to which
it pertains. A separate UCD message must be sent for each logical
upstream that is currently available for use. The details of many
of the concepts used herein are given in the DOCSIS 1.0, 1.1 and
2.0 specifications, all of which are hereby incorporated by
reference.
[0028] Cable modems have different upstream throughput because of
the following reasons. [0029] 1) DOCSIS 1.0 CMs have low throughput
because they do not have a transmit equalizer and they usually use
QPSK modulation. [0030] 2) DOCSIS 1.1 CMs have a transmit equalizer
which is only 8 taps. This gives a lower throughput than DOCSIS 2.0
CMs. DOCSIS 1.1 CMs generally use 16-QAM modulation so they have a
higher througput than DOCSIS 1.0 CMs. [0031] 3) DOCSIS 2.0 CMs have
a 24-tap transmit equalizer, so they can use higher throughput
modulation up to 64-QAM. [0032] 4) CMs which cannot overcome high
path attenuation with their power amplifiers are received at the
CMTS with lower received power thereby yielding lower
signal-to-noise ratio. [0033] 5) Different CMs have different paths
through the cable plant, and these different paths contribute
different distortions and echoes which can create different levels
of receive errors for each CM. [0034] 6) Different CMs can have
defects in their design or manufacture which can cause signal
distortion. This causes errors in reception and requires
downgrading the burst profile to reduce the throughput and/or add
more error detection and correction capability. [0035] 7) DOCSIS
2.0 SCDMA and ATDMA modes have different throughput because they
have different schemes that handle noise and they handle impulse
noise differently.
[0036] The DOCSIS specifications in the prior art suggest dividing
upstream logical channels into groups based upon modem types:
DOCSIS 1.0 and 1.1 modems should be grouped on one logical channel,
and DOCSIS 2.0 ATDMA and DOCSIS 2.0 SCDMA modems should each have
their own groups and operate on separate upstreams.
[0037] This approach is better than no grouping but still is
inadequate for the reasons stated above since there are other
factors which affect throughput which the DOCSIS specification do
not take into account.
Grouping Factors
[0038] The approach according to the invention is to divide the CMs
into groups or logical channels based upon one or more of the
following factors: [0039] 1) modem type with DOCSIS 1.0 and DOCSIS
1.1 CMs are each segregated into different groups each with their
own upstream, and DOCSIS 2.0 CMs are initially grouped together.
Subsequent monitoring of 2.0 modem performances can lead to the
conclusion that they need to be subdivided into separate logical
groups each with its own upstream with a burst profile tailored to
the throughput ability of the group.
[0040] The following factors can also be used to segregate CMs into
logical groups over and above the logical groups 1.x, ATDMA and
SCDMA recommended by the DOCSIS specifications. In one species
within the genus, logical grouping by modem type with 1.x and 2.0
modems separated into two logical groups with different upstream
burst profiles is performed first. This is followed by subgrouping
within each one of these type groups, based upon one or more of the
following throughput ability factors.
[0041] 2) received signal SNR;
[0042] 3) the packet loss rate;
[0043] 4) the bit error rate;
[0044] 5) the byte error rate;
[0045] 6) the received signal power;
[0046] 7) the cable node to which a CM is connected since some
cable nodes may be less capable than others or be suffering from
problems.
[0047] The packet loss rate can be calculated using the HCS and CRC
error. The MAC header of each burst has a two byte Header Check
Sequence (HCS) to ensure the integrity of the MAC header. The HCS
is calculated using CRC-CCITT polynomial defined in ITU-T-S.25. The
variable length PDA includes a pair of 48 bit addresses, data and a
CRC field. The DOCSIS MAC protocol protects against errors through
the use of checksum fields across both the MAC header and the data
portions of the packet. All MAC messages for management are
protected by a CRC covering the entire message. Any message with a
bad CRC must be discarded by the receiver.
[0048] Moving of CMs to different upstreams based upon packet loss
rate is preferably done using hysteresis as follows: [0049] If
packet loss percentage>PACKET_LOSS_REQUIRED*(1+TH1)
[0050] Move CM to a lower throughput logical channel [0051] If
packet loss percentage<PACKET_LOSS_REQUIRED*(1-TH2)
[0052] Move CM to a higher throughput logical channel.
End
[0053] TH1 and TH2 are packet loss constants, preferably
programmable, used to provide hysteresis so that switching does not
happen too frequently (chatter) caused by momentary dips or
increases in packet loss percentage. The packet loss percentage is
calculated for data bursts that do not have contention grants or
unsolicited grant service because these grants may be empty or have
collisions. The packet loss percentage is averaged over a
sufficiently long time (for example 1000 packets) so that channel
switching is not done when channel conditions change for a short
time. After switching, the packet loss counter should be reset and
restart the count.
[0054] The same holds true for subgrouping by SNR and moving CMs to
lower or higher throughput upstream channels. Hysteresis constants
should be used for the comparison and SNR should be calculated for
data burst not transmitted during contention grants. SNR should be
averaged over a sufficiently long time to avoid changing channels
on momentary dips or increases in performance. The SNR_REQUIRED
constant is a constant set on a per logical channel basis. Typical
SNR values are set to get packet loss rates of about 0.5%.
[0055] Byte error rate is calculated according to the following
formula: ByteErrorRate = ( T + 1 ) * n_error ( k + 2 .times. T ) *
n_total ( 1 ) ##EQU1## where
[0056] T is the number of the maximum correctable bytes in an RS
codeword;
[0057] T+1: when there is an uncorrectable RS codeword, the is a
high probability of having T+1 errored bytes;
[0058] n_error: the total number of uncorrectable RS codewords of a
specific IUC (burst type) in a certain time interval;
[0059] n_total: the total number of received RS codewords of a
specific IUC in a certain time interval;
[0060] k: the number of information bytes in an RS codeword of a
specific IUC; and
[0061] k+2T: the codeword length in bytes.
[0062] Since on the average only half the bits are erroneous, the
bit error rate is given by: B .times. .times. E .times. .times. R =
0.5 * ( T + 1 ) * n_error ( k + 2 .times. T ) * n_total ( 2 )
##EQU2## Generally, the long data grant is used to calculate a the
BER or byte error rate.
[0063] The Interval Usage Code (IUC) is defined by DOCSIS as a
field in the MAP and UCD messages to link burst profiles to grants.
In other words, the IUC tells what kind of burst may be transmitted
during each grant. The IUC codes map to sets of burst parameters
such as modulation, RS error correction capability T, etc.
[0064] One way to increase the number of supported logical channels
by the CMTS physical layer is that the logical channels will use
only the long data grant and not the short data grant. Then the
CMTS PHY chip has 2 IUCs (short and long data grants) that can be
used for 2 logical channel data grants. The other IUCs of REQ
ranging) should be common to all logical channels.
[0065] Burst Profiles Examples TABLE-US-00001 Use for Impulse Burst
Noise Net Data Profile Channel? Rate @ % bit rate # Yes/No 2.56
Msps Modulation RS from max 1 Yes 2.3 Mbps QPSK k = 16 22% t = 10 2
Yes 3.0 Mbps QPSK k = 28 29% t = 10 3 Yes 4.0 Mbps QPSK k = 78 39%
t = 10 4 No 4.7 Mbps QPSK k = 235 46% t = 10 5 Yes 5.1 Mbps 16-QAM
k = 20 50% t = 20 6 Yes 6.4 Mbps 16-QAM k = 39 62% t = 10 7 Yes 8.1
Mbps 16-QAM k = 78 79% t = 10 8 Yes 9.4 Mbps 16-QAM k = 235 92% t =
10 9 No 10.24 Mbps 16-QAM k = 16 100% t = 0
Definitions
[0066] A logical channel in the DOCSIS upstream is defined in the
DOCSIS specifications as, "a MAC entity identified by a unique
channel ID and for which bandwidth is allocated by an associated
MAP message. The associated UCD (Upstream Channel Descriptor--the
burst profile) and MAP messages (which CMs can transmit and when)
completely describe the logical channel." A physical upstream
channel may support multiple logical upstream channels in different
time intervals. Both physical upstream-channels and logical
channels will be referred to as upstream channels or upstreams in
the specification and the claims since it does not make any
difference for purposes of the invention exactly what the form of
the upstream implementation is.
[0067] A downstream group is a group of one or more downstreams to
which cable modems (CMs) are coupled that are to share one or more
DOCSIS upstreams which may be of different DOCSIS types (1.0, 1.1,
2.0 ATDMA, 2.0 SCDMA etc. and may have different throughput levels
even within a type).
[0068] The UCD message is defined by the DOCSIS specs as "the MAC
Management message used to communicate the characteristics of the
upstream physical layer to the cable modems." Basically, the
transmission characteristics of each logical channel, as defined by
the corresponding UCD message, are separated into three portions:
1) channel parameters; 2) burst profile attributes, and 3) user
unique parameters.
[0069] In the appended claims, the term "burst parameters" is
intended to include all those channel parameters, burst profile
attributes and user unique parameters needed by the shared back end
demodulator to properly process each burst. The logical channel
parameters include: a) the symbol rate which can be any one of 6
different rates from 160 ksym/sec to 5.12 Msym/sec in octave steps;
b) the center frequency; and c) the 1536-bit preamble superstring
that is prepended to at least some bursts; and d) the SCDMA channel
parameters. These characteristics are shared by all CMs on a given
channel or sub-channel (hereafter logical channel may be used to
refer to either channel or sub-channel).
[0070] The burst profile transmission characteristics of a logical
channel, in the preferred embodiment, include: modulation (QPSK, 64
QAM, 128 QAM etc.), differential encoding on or off; Trellis or TCM
encoding on or off; preamble length, preamble value offset;
preamble type (QPSK 0 or QPSK1), RS (Reed-Solomon) error correction
T value from 0 to 16 where 0 is no FEC bits to 16 for the maximum
where the number of codeword parity bytes is 2.times.T, RS codeword
length, scrambler seed, max burst length in minislots, guardtime
from 5 to 255 symbols for TDMA channels and 1 symbol for SCDMA
channels, last codeword (fixed or shortened), scrambler on or off,
byte interleaver depth, byte interleaver block size, SCDMA
spreading on or off, codes per subframe, and SCDMA interleaver step
size. User unique parameters may vary from user to user even when
on the same channel and same burst type and include such things as:
power level; offset frequency (defines center frequency of channel
to transmit on); ranging offset to achieve minislot boundary
alignment at CMTS (which also achieves upstream chip clock
alignment between the upstream chip clock generated at the CMTS and
the chip clock embedded in the received signal at the CMTS
receiver--a state which is referred to herein as "phase
coherence"), burst length in minislots if variable on the specified
channel (changes from burst to burst); and the transmit equalizer
coefficients (up to 64 coefficients specified by 4 bytes per
coefficient--2 real and 2 complex).
[0071] Profile robustness refers to the UCD factors defined above
and how they are set for any particular logical channel. More
robust profiles have burst parameters that have lower symbol rate,
less complex modulation constellations and/or allow for greater
error detection and correction capability (forward error correction
or FEC). Greater FEC results in lower throughput because there is
more overhead consumed in each burst with forward error correction
factors such as code word length, byte interleaver depth and block
size, RS error correction T value, RS codeword length, guardtime,
preamble length, Trellis encoding on or off, etc.
Grouping Based Upon DOCSIS 1.0 and 1.1
[0072] Although the DOCSIS specifications suggest grouping DOCSIS
1.0 and 1.1 modems in the same group, one species of the invention
segregates them into separate groups and, in some species
subdivides each 1.0 and 1.1 group into separate logical channels
based upon other factors such as received signal power or SNR or
packet loss rate, bit error rate, etc. Other species may lump 1.0
and 1.1 CMs into one logical group and then subdivide the group
into one or more other upstream channels based upon monitoring for
overperformance or underperformance. If some CMs are
underperforming, an upstream channel which is more robust will be
created and these underperforming modems moved to it. If some CMs
are overperforming, an upstream channel with greater throughput
will be created and these overperformers moved to it. Monitoring
for overperformance or underperformance can be by SNR, received
power, packet loss rate, bit error rate, byte error rate, etc.
[0073] DOCSIS 1.0 CMs generally require a more robust burst profile
because they do not have a transmit equalizer. This yields lower
throughput than DOCSIS 1.1 CMs which have a transmit equalizer and
can therefore transmit at higher throughput rates and require less
robust burst profiles. DOCSIS 1.1 CMs generally can use 16-QAM
modulation (4 bits per symbol transmitted--16 constellation points)
and symbol rate of 2.56 megasymbols per second (Msps), while DOCSIS
1.0 CMs generally use QPSK (2 bits per symbol--4 possible symbols).
Because of these very different modulation types and symbol rates
and the lack of a transmit equalizer (more errors resulting in the
need for burst profiles with more overhead for forward error
correction to detect and correct errors) in a DOCSIS 1.0 CM, less
data can be transmitted per second upstream than can be transmitted
by a DOCSIS 1.1 CM. This is why DOCSIS 1.0 and DOCSIS 1.1 CMs are
separated into separate logical groups in the invention contrary to
the teachings of the DOCSIS specification.
[0074] CMs may be moved to a different upstream either by the CMTS
via downstream UCC (upstream channel change) or DCC (dynamic
channel change) messages or manually by the cable operator.
Grouping Based Upon DOCSIS 2.0 ATDMA and DOCSIS 2.0 SCDMA
[0075] DOCSIS 2.0 ATDMA CMs have different schemes for handling
noise and impulse noise than DOCSIS 2.0 SCDMA modems. This results
in different throughput, so these modems should be grouped in
different groups, as recommended by the DOCSIS specifications, but
subgrouping by throughput ability is used in the invention to
further subdivide the ATDMA and/or SCDMA logical groups into
subgroups based upon throughput ability as indicated by link
quality parameters: the received signal-to-noise ratio; the packet
loss rate; the bit error rate; the byte error rate; the received
signal power; the RS codeword error rate or the cable node to which
a cable modem is coupled.
[0076] FIG. 1 is a flowchart of a process to group modems into
different groups based upon upstream throughput capability based
upon modem type alone (no subgrouping) with 1.0 and 1.1 CMs in
separate groups, and to establish a different upstream logical
channel for each and to command each CM to transmit upstream on a
logical channel having a burst profile suited to the modem types
assigned to the logical channel. The process starts at 10 and
proceeds to step 12 where the CMTS receives initial registration
communications from each CM. Registration occurs for each cable
modem after it powers up and does its ranging and training, but
cable modems may re-register whenever they change upstream logical
channels in some embodiments. The step of receiving registration
communications in the CMTS in the appended claims is intended to
cover both cases.
[0077] During this registration process, the CMTS learns the type
of the CM which is registering. CMs scan for a valid DOCSIS
downstream when they first power up. They then determine which
upstreams are linked to that downstream and join the upstream most
compatible with the modem type and send initial training bursts.
After training is completed, the CM registers with the CMTS and the
registration messages tell the CMTS what type of service (DOCSIS
1.0, 1.1, ATDMA or SCDMA) each CM is. The registration messages
and/or training bursts also tell the CMTS to which downstream the
CM is tuned and on which upstream it is operating.
[0078] In step 14, the CMTS assigns each CM which registers to a
group comprised of other CMs of the same type. All DOCSIS 1.0 CMs
are put into one group and DOCSIS 1.1 CMs are put into another
group. All DOCSIS 2.0 ATDMA CMs are put into a third group, and all
DOCSIS 2.0 SCDMA modems are put into a fourth group. In alternative
embodiments, the 2.0 CM can all be grouped into one logical channel
which is either ATDMA or SCDMA. Also in step 14, the CMTS sends a
DCC (Dynamic Channel Change) or other DOCSIS message to each CM
that needs to be moved to another upstream channel telling that CM
to which upstream logical channel it has been assigned. The CM then
looks up the UCD message for that logical channel and uses the
burst profile information therein to configure itself to transmit
according to the modulation type, symbol rate, center frequency,
various forward error correction factors, etc. for the upstream
logical channel, as defined in the UCD message.
[0079] In step 16, the CMTS generates an sends an appropriate
Upstream Channel Descriptor (UCD) message for each group. Each UCD
message establishes a separate upstream logical channel having a
burst profile which is appropriate for the CM throughput
capabilities of the CMs in the group which will be assigned to
transmit on that logical channel. These UCD messages are
transmitted to all CMs and each CM receives each UCD message and
stores it. Each UCD message has an ID for the logical upstream
channel to which it pertains.
[0080] Step 20 ends the process.
Subgrouping Based Upon Receiving Power, SNR, etc.
[0081] Some cable plants, notably in Asia, have cable path
attenuation that is so high, it cannot be overcome by the CM's
power amplifier. In other words, even though the CMTS commands the
CM to transmit with more power, the power amplifier has reached the
top of its power range and still its signal does not have enough
power when it reaches the CMTS. To overcome this problem, cable
operators use the following techniques.
[0082] 1) They can reduce the required power for upstream
transmissions (the required power level at which an upstream burst
must be received at the CMTS) to a level which is low enough that
even CMs that have a high attenuation path have enough transmit
power to meet the requirement. The lower received power reduces the
SNR for the entire upstream channel, and penalizes all modems for
just a few who have high attenuation paths. The lower SNR causes
more errors thereby requiring more overhead for forward error
correction to keep the errors under control.
[0083] 2) They can modify the modem to add a more powerful transmit
power amplifier. This requires the cable modem manufacturers to
manufacture a special class of CMs just for these customers or
requires the customers to do it themselves. With thousands of
deployed modems, this is a very expensive solution.
[0084] 3) They can use SCDMA modems with a power management mode
which uses fewer spreading codes for CMs which have a high
attenuation path so there is more power transmitted per spreading
code. This requires the cable operator use SCDMA modems and have
head end equipment compatible therewith which many cable operators
do not yet have. This is a good solution because only the SCDMA
modems with the power problem are penalized by having to use fewer
spreading codes, but this solution does not work for TDMA cable
modems so systems which have not yet upgraded their equipment to be
DOCSIS 2.0 compatible cannot use this solution.
[0085] The method of the invention proposed here solves the problem
for modems which have insufficient power or inadequate SNR in
either a TDMA or SCDMA upstream environment so all systems can use
the method. This method divides up the CMs into at least two
groups. One group transmits on one logical channel with a high
throughput and which contains only CMs which do not have a high
attenuation path or at least which have adequate power and/or
adequate SNR to meet the received power and/or SNR specification at
the CMTS. The second group is comprised of CMs that have inadequate
power and/or inadequate SNR to meet the upstream received power
and/or SNR specification of the CMTS. These modems are grouped to
transmit on a different logical channel with a more robust burst
profile (more overhead for FEC) and lower throughput. The burst
profile for this logical channel will be such as to be able to
handle the lower received power and/or the lower SNR. Also, lower
complexity modulation constellations such as QPSK modulation may be
used. Further, more aggressive FEC overhead may be used such as RS
error correction with a sufficient number of error detection and
correction (ECC) bits to correct worst case scenarios for errors
caused by the low received power or inadequate SNR can be used on
the lower throughput logical channel.
[0086] On the higher throughput channel with CMs that have
sufficient power and/or a high SNR, fewer errors will result so,
for example, 16-QAM or better modulation, a higher symbol rate and
RS encoding with fewer numbers of ECC bits may be used. For DOCSIS
2.0 CMs, 64-QAM modulation can be used.
[0087] A flowchart shown in FIG. 2 comprised of FIGS. 2A through 2C
represents species 4 in the summary of the invention. It is the
preferred method to initially set up the groups based upon modem
type with SNR and/or received power monitoring during initial
training to establish one or more low power channels if necessary
to achieve successful registration even for CMs with power
shortfall and/or SNR problems. The process starts at step 22, and
at step 24, the CMTS assumes that 1.X and 2.0 CMs are present and
creates a 1.x upstream and 2.0 upstream for each downstream the
CMTS is transmitting.
[0088] In step 26, each CM powers up (this happens at random times
whenever the user wishes to use the CM or resets it) and listens
for, finds and locks onto a valid DOCSIS downstream. That
downstream carries UCD messages that are broadcast to every CM that
define the upstreams that are linked to that downstream and the
burst profiles of each. There will be at least one 1.0 and at least
one 2.0 upstream present in the preferred embodiment. However, in
other embodiments, there may also be a 1.x upstream with a burst
profile optimized for 1.1 CMs and several types of 2.0 upstreams
present such as 2.0 ATDMA and/or 2.0 SCDMA (2.0 upstreams with
different burst profiles, one optimized for ATDMA and/or one for
SCDMA). The cable system operator will configure the CMTS to create
a separate upstream for at least 1.x and 2.0 modems for every
downstream using UCD messages (if both 1.x and 2.0 CMs are
available in the channel). Each CM looks at the UCD messages
transmitted in its downstream and determines which upstream matches
its type. The CM then picks the best available upstream for its
type (if there are multiple upstreams which the CM can join) or
picks the single available upstream for its type, and configures
its upstream transmitter to transmit on that upstream.
[0089] In step 28, each CM examines the MAP messages which pertain
to the upstream it has selected which are transmitted in the
downstream to which it is locked. The CM finds in these MAP
messages the location in time of an initial training interval.
[0090] In step 30, each CM transmits an initial training burst
during the initial training interval and includes in the burst the
CM's MAC address and an initialization service Identifier (SID).
For 2.0 CMs, this initial training burst is called an INIT-RNG-REQ
in the DOCSIS spec. For 1.x CMs, this initial training burst is
called a RNG-REQ burst. The SID in the initial training burst will
later be changed by the CMTS to a primary SID which will be sent to
the CM in a ranging response message. The initial training burst
also includes the downstream channel ID of the downstream channel
the CM locked onto and upon which a UCD message was received
describing the upstream on which the initial training burst was
sent.
[0091] The steps that follow are steps the CMTS performs for
initial ranging from the DOCSIS 2.0 specification, FIG. 11-8, but
these steps have been modified to reflect the changes that are
required to practice the invention. The initial ranging process is
a two phase process which is designed to perform enough iterations
of training that the CM gets into precise synchronization with the
CMTS such that a registration process can be carried out
efficiently with few errors. The iterations are comprised of the CM
transmitting a training burst followed by the CMTS making
calculations regarding various factors and sending back correction
messages to the CM which is followed by the CM making corrections
and sending another training burst and repeating until the CM has
been properly trained or the allowable number of iterations has
been exhausted.
[0092] In step 32, the CMTS waits for a recognizable initial
ranging burst from a CM. Step 34 represents receipt of the initial
ranging burst.
[0093] Step 36 determines if a SID has been assigned to the CM
already. CMs start out with initialization SIDs in their initial
ranging bursts, but the CMTS will assign a temporary SID after
receiving their initial ranging burst. The MAC address in the
initial ranging burst allows the CMTS to determine from which CM
each initial ranging burst came.
[0094] If step 36 determines that a SID has not been assigned to
the CM whose initial ranging burst was just received, step 38 is
performed to assign a temporary SID to this CM. Step 40 is then
performed to add the CM to a polling list for future MAPs. This
means that this CM will be added to the list of CMs which will
receive invitations to send further training bursts if necessary
during the training process to get the CM into synchronization with
the CMTS. If step 36 determines that a SID has already been
assigned to this CM, a retry count in the poll list for this CM is
reset.
[0095] After either step 42 or step 40 is performed, step 44 is
performed. Step 44 represents the process of making timing,
frequency and power offset measurements and calculation of the SNR
and upstream equalization coefficients on the initial ranging burst
in the CMTS. These measurements (except for the SNR) are then sent
in a downstream message (RNG-RES) to the CM that sent the initial
ranging burst to cause it to readjust its transmit parameters. In
step 46, the CM receives the ranging response message and makes the
requested adjustments in its transmit power, frequency and timing
and uses the equalization coefficients to adjust the coefficients
of its transmit equalization filter. Phase two of the initial
ranging process now starts.
[0096] Step 48 represents the process of polling the CMs on the
list of CMs that have sent initial training bursts with invitations
in MAP messages for the particular upstream each CM is on. Each
invitation invites a particular CM to send an additional ranging
burst during a specified interval. Step 48 also represents waiting
for these polled RNG-REQ training bursts. Each training burst will
contain the temporary SID assigned the CM by the CMTS. Step 50
represents the receipt by the CMTS of a ranging burst from a CM.
This ranging burst will have been sent using the new parameters
previously sent the CM in response to its initial ranging
burst.
[0097] Step 52 represents the process of determining if the CM's
ranging burst is good enough for the required synchronization.
Measurements of power, frequency and timing offset will be made and
equalization coefficients will be calculated. Other parameters such
as signal-to-noise ratio (SNR); bit error rate; byte error rate, RS
codeword error rate and packet error rate will also be measured in
some embodiments, but usually these measurements are not made until
after registration. These parameters as well as power shortfall can
be used in post registration processes in some embodiments to group
modems with problems onto more robust upstream channels (channels
with a lower throughput and more overhead devoted to forward error
correction). Any one or a combination of these parameters can be
used as an indicator that a CM has to be moved to a lower
throughput, more robust upstream channel. In the preferred
embodiment, test 52 determines if the power, frequency, timing and
equalization coefficients and/or SNR (hereafter the ranging
parameters) are within acceptable limits to end the initial
training process and proceed to registration. If all these ranging
parameters are acceptable, step 54 is performed to send a ranging
response message to the CM who sent the training burst indicating
training is complete and has been successful. Step 56 then removes
the CM from the polling list, and the process finishes as step 58
where processing moves on to the registration process.
[0098] If test 52 determines that one or more of the ranging
parameters is not within limits, test 60 is performed to determine
if the number of permissible retries has been exhausted. If so,
step 62 is performed to send a ranging response message to the CM
telling it to abort the training process, and step 64 is performed
to remove the CM from the polling list. Then, the CMTS waits for a
recognizable training burst from a CM, as symbolized by step
66.
[0099] Returning to step 48, it is possible that the CMTS will poll
a CM inviting another training burst and then not receive any, as
symbolized by step 68. If that happens, the CMTS polls the CM
again, and test 70 determines if the maximum number of retries at
polling the CM has been exhausted. If not, step 48 is performed
again to send another invitation to the CM to sending a training
burst. If the maximum number of polling attempts to this CM has
occurred, processing proceeds to step 64 to remove the CM from the
polling list, and then processing proceeds to step 66 to wait for a
recognizable RNG-REQ training burst in step 66 from another CM.
[0100] The modification to this DOCSIS initial training process
occurs starting at step 72. In step 72, a test is made to determine
if the particular criteria (received signal power or SNR) being
used to determine if the CM which sent the training burst needs to
be moved to another more robust channel so as to complete
registration is within acceptable limits. Step 72 is performed if
step 60 determines that the number of retries is not exhausted and
if test 52 determines that the measured ranging criteria (power,
frequency, timing and equalization coefficients and/or SNR) are not
all within limits or otherwise acceptable. In one embodiment, the
criteria compared to limits in step 72 is the received power. In
another species, it is SNR. In another species, it is both SNR and
received power. Step 72 will be reached if any one of the ranging
criteria measurements is not acceptable and the retries are not
exhausted. The criteria that may be used to determine if a CM needs
to be moved to a more robust channel within its logical group will
be referred to herein as the measurement criteria and are: received
power and signal-to-noise ratio (SNR) although in other embodiments
where other measurements are made by the CMTS on the initial
training-bursts such as bit error rate (BER), byte error rate and
packet error rate (PER), those other criteria may also be used. The
particular criteria being used is measured in step 52, or, if not
in step 52, in step 72. Step 52 however only compares the ranging
criteria of power, frequency and timing offset and equalization
coefficients and/or SNR to standards of quality in reaching its
conclusion as to whether the training burst is good enough. Step 72
compares the received power to the desired received power level
again because it is possible to reach step 60 if any one of the
ranging criteria is not acceptable or if all but the received power
is acceptable. In alternative species where SNR is used, step 72
compares the SNR measured during the ranging bursts to an
acceptable SNR level). Further, test 52 does not compare the other
criteria such as BER or PER or SNR to limits in reaching its
conclusion even though they might be calculated. Therefore, the
exact status of the received power (or other criteria) must be
determined in order to draw a conclusion regarding whether the CM
must be moved to a more robust channel, and that is what step 72 is
for.
[0101] If the received power or other criteria examined in step 72
is within acceptable limits, step 74 is performed to send a
downstream message RNG-RES to the CM that sent the training burst
telling it to continue ranging and sending it new adjustments to
make. Then step 76 is performed to wait for a new polled training
burst from the CM while processing vectors back to step 48 to
invite the CM to send another training burst.
[0102] If step 72 determines that the power or other criteria being
compared is not acceptable, step 78 is performed to increment the
retry counter which keeps a separate record of the number of
retries to get the received power or other criteria correct. Step
80 then determines if the maximum number of retries for adjusting
received power (or the other measurement criteria) have
occurred.
[0103] If step 80 determines that the maximum number of retries
have not been exceeded, steps 74 and 76 are performed again to send
a downstream RNG-RES message giving the CM new adjustments on the
measurement criteria trying to be improved and telling the CM to
that ranging status is "continue". Step 76 waits for a new training
burst, and step 48 invites a new training burst.
[0104] If however, step 80 determines that the maximum number of
retries to get the received power or other measurement criteria
right have occurred, the CMTS concludes in step 82 that the CM has
a power shortfall or some other problem that cannot be overcome by
further adjustments. This means that the CM must be moved to a
channel with a lower throughput (lower symbol rate, less complex
modulation) and more robust forward error correction properties to
allow successful communication with this CM. Such a more robust
channel with lower throughput and aggressive forward error
correction burst profile will hereafter be referred to as a low
power channel even though a bad SNR, PER, BER or byte error rate
may be the problem in some species that causes the CM to be moved
to the more robust channel.
[0105] The first step in moving the CM to a low power channel is
the determination in step 84 as to whether such a low power channel
for this modem type has already been created. The process of FIGS.
2A-2C then finds the CMs that have a power shortfall or other
problem that cannot be resolved by adjustments during the initial
training interval and moves them to a low power channel for their
modem type.
[0106] If step 84 determines that a low power channel already
exists for the modem type, step 86 is performed to send a channel
change message to the CM. The channel change message assigns the CM
to the already existing low power channel for CMs of its type. Step
88 represents the process of the CM receiving this message and
changing its configuration to transmit on the low power channel to
which it has been assigned. Step 88 also represents the process of
the CM performing one of the training regimens specified in the
DOCSIS specification for retraining after an upstream channel
change.
[0107] Test 90 determines if ranging has been successful for the CM
on the new upstream channel. If not, step 92 is performed to send
an abort ranging RNG-RES downstream message to the CM to tell it to
stop ranging. A service call is in order for this CM. If test 90
determines that ranging has been successful on the new upstream
channel, step 94 is performed to exit the training process and go
to the registration process.
[0108] If test 84 determines that no low power channel has already
been created for the type of CM which step 82 determined needs to
be sent to a more robust channel, then step 96 is performed. In
step 96, a low power channel with robust forward error correction
and less complex modulation and/or lower symbol rate burst profile
is created by creating and broadcasting an appropriate UCD message
on the downstream to which the CM is tuned. After the new low power
upstream is created, the CMTS sends a channel change message to the
CM assigning it to the new low power channel. Thereafter, steps 88,
90, 92 or 94 are performed as previously described.
[0109] Both step 94 and step 58 transfer processing to the
registration process with the CM that just completed initial
training registers with the CMTS. The registration communications
tell the CMTS what type of DOCSIS modem the CM is, and which
downstream and which upstream it is on. This gives the CMTS the
information it needs to make logical groupings on the appropriate
upstreams. That process starts with step 100 on FIG. 2D to register
the CM. Then, step 102 is performed to determine the modem type and
which upstream channel the modem is transmitting upon from the
registration data.
[0110] Test 104 is then performed to determine if the modem is on
the correct upstream channel for its type and throughput ability.
In one embodiment, the CMTS makes a determination here if any
DOCSIS 1.1 CMs are transmitting on DOCSIS 1.x upstreams which have
burst profiles which are not tailored for maximum throughput for
DOCSIS 1.1 modems. Recall that DOCSIS 1.1 modems have 8-tap
transmit equalizers which allows them to use 16-QAM modulation as
compared to QPSK modulation which DOCSIS 1.0 modems usually use
because of their lack of a transmit equalizer. Therefore, it is
advantageous to separate DOCSIS 1.1 modems into a logical group
assigned to an upstream with a burst profile which is optimized for
the higher throughput ability of DOCSIS 1.1 modems instead of
penalize all 1.1 modems by making them transmit on a 1.x upstream
which has a burst profile tailored to the lower throughput of 1.0
modems. In other embodiments, the determination as to whether a CM
is on a suitable upstream channel is based on SNR, bit error rate,
byte error rate, packet loss rate, etc. Step 106 is then performed
to create whatever upstreams are needed (which have not already
been created) for the logical groups into which the CMs have been
divided. Each upstream will be tied to the downstream to which a CM
to be moved onto that upstream is listening and each will have a
burst profile tailored to the modem type/modulation profile of the
CM. As many different upstreams as are necessary to serve all the
CMs and all the logical groups are created. Step 108 is then
performed to send channel change messages to any CM that needs to
be moved to a different logical group and a different upstream
channel.
[0111] The above explains how 1.x modems are handled. What the CMTS
does about 2.0 modems can be categorized into several different
subspecies within this species of FIGS. 2A-2D. In some embodiments,
step 24 represents the process of the CMTS creating 2.0 ATDMA and
SCDMA upstreams and the CMs picking whatever upstream they are most
compatible with after they latch onto the downstream. In this
subspecies, steps 102, 104, 106 and 108 represents the CMTS leaving
the 2.0 CMs in separate logical groups for ATDMA and SCDMA as
established in defacto fashion by the CMs themselves when they
picked an upstream. In another subspecies, steps 102, 104, 106 and
108 represent the process of the CMTS grouping all DOCSIS 2.0 CMs
into one logical group and ordering them all to operate in SCDMA
mode on one upstream channel having a burst profile optimized for
2.0 SCDMA. In other subspecies, some or all of the 2.0 CMs may be
ordered to operate in ATDMA. ATDMA and SCDMA 2.0 modems need to be
segregated into separate logical groups however if they are
coexisting, each operating on an upstream having a burst profile
tailored for the modulation profile in which the modem is
operating.
[0112] The CMTS also determines which upstreams have already been
defined and determines which upstreams it needs for the CMs that
have registered. In step 106, the CMTS creates any additional
upstreams with the needed burst profiles that it needs for the
number of logical groups it has.
Species 5: FIGS. 3A through 3D
[0113] Species 5 is the same as species 4 from the summary of the
invention but including post registration monitoring and
subgrouping. The subgrouping is done within the already established
groups established by the processing up through step 108. The
subgrouping is based upon over or under performance of a CM for the
upstream upon which it operates. Steps 22 through 108 of FIGS. 3A
through 3D operate as previously described for FIGS. 2A through
2D.
[0114] The difference of species 5 over species 4 lies in the post
registration and grouping monitoring and subgrouping. That process
starts at step 110 on FIG. 3D. Step 110 symbolizes the process of
monitoring each CM's post registration data communications and
determination for each CM of one or more throughput ability
factors. These throughput ability factors are:
[0115] 2) received signal SNR;
[0116] 3) the packet loss rate;
[0117] 4) the bit error rate;
[0118] 5) the byte error rate;
[0119] 6) the received signal power;
[0120] 7) the cable node to which a CM is connected since some
cable nodes may be less capable than others or be suffering from
problems;
[0121] 8) RS codeword error rate.
These factors will change as conditions such as additive white
guassian noise, burst noise, impulse noise, degradation of the
cable plant, and other impairments improve or get worse over
time.
[0122] Step 112 represents the process of comparing each CM's
throughput ability factor(s) to overperformance and
underperformance limits to determine if the CM is over performing
the upstream channel it is on or underperforming. Overperforming
would generally mean that the throughput rate is too low for a CM
and the forward error correction factors are being underutilized
such that the most efficient communication is not occurring because
the symbol rate is too low, a more complex modulation could be used
or less overhead in FEC bits could be used with a still acceptable
error rate. An overperforming CM can be moved to an upstream
channel with a higher throughput, more complex modulation
constellation, less overhead consumed for forward error correction
factors or both. An underperforming CM will be having a higher than
acceptable bit error rate, byte error rate, packet loss rate, or
received signal power or SNR which is too low. It needs to be
switched to a lower throughput upstream channel with a lower symbol
rate, less complex constellation, and/or more overhead consumed in
FEC bits for more efficient communications.
[0123] If the CM is overperforming, the CMTS checks in step 114 if
there is an upstream with a higher throughput (for example 16-QAM
with smaller amounts of Reed Solomon encoding or higher modulation
such as 64-QAM for DOCSIS 2.0)) for the CM's modulation
profile/DOCSIS type which is linked to the downstream to which the
CM is tuned. If a higher throughput upstream already is in
existence which is linked to the downstream to which the CM is
tuned and which has a burst profile tailored to the CM's throughput
abilities, step 116 is performed to send a downstream command to
the CM to cause it to move to the upstream so identified.
[0124] If step 114 determines that a higher throughput upstream
which is linked to the downstream this CM is tuned to and which has
a burst profiled tailored to the needs of the CM, then step 118 is
performed to create a new upstream. This upstream will have a
higher throughput and a burst profile established to better serve
the CMs throughput ability. After step 118 is performed, step 116
is performed to send a downstream command to the CM to cause it to
move to the higher throughput upstream. Establishment of the new
upstream or finding a suitable higher througput existing upstream
and moving the CM to the higher throughput upstream creates a new
logical subgrouping within the grouping by modem modulation
profile/DOCSIS type. Then step 120 is performed to read the
throughput ability factors for the next CM, and processing vectors
back to test 112.
[0125] If test 112 determines that the CM being processed is
underperforming, processing vectors to step 122. There, the CMTS
determines if a lower throughput upstream already exists with a
suitable burst profile (for example QPSK and R-S error correction
used on lower throughput channels) for the underperforming CM and
which is linked to the downstream to which the CM is tuned. If such
an upstream already exists, step 124 is performed to send a
downstream command to the CM to cause it to move to the lower
throughput upstream located in step 122. If such an upstream does
not already exist, step 126 is performed to create it. Following
both steps 126 and 124, step 120 is performed to read the througput
ability factors for the next CM, and then processing vectors back
to step 112 to determine if the next CM is over or under
performing. This process continues until every CM in every logical
group by type/modulation profile has been processed. In some
species, the process of monitoring all CMs in all logical groups is
repeated periodically, and in other species, it is only done
once.
Species 2: FIG. 4
[0126] FIG. 4 is a flowchart of a process to group CMs by
modulation profile/type and break 1.0 and 1.1 modems into separate
logical groups, each with their own upstream, and group all 2.0 CMs
together in one logical group. Then monitoring of transmission
quality is done, and subgrouping within each logical group based
upon overperformance or underperformance is done. The process
starts at step 130 and then step 132 is performed where the CMTS
establishes a 1.0 and 1.1 and 2.0 ATDMA or SCDMA upstream for every
downstream. In step 134, initial training bursts are received from
each CM and are processed to train each CM. Optionally, in some
species, monitoring for low SNR or inadequate received power during
initial ranging can be performed and low power upstreams can be
established for CMs that have a power shortfall problem or low
SNR.
[0127] In step 136, the CMTS conducts registration communications
with each CM and learns its DOCSIS type/modulation profile and, if
not already known, the upstream and downstream the CM is tuned to.
In step 138, the CMTS creates separate upstreams per modem type for
each downstream, the new upstreams having burst profiles tailored
for DOCSIS 1.1 CM and 1.0 CM throughput ability. Downstream
messages are then sent to move all DOCSIS 1.1 CMs on each
downstream to the new 1.1 upstream and to move all DOCSIS 1.0 CMs
to the new 1.0 upstream. All DOCSIS 2.0 CMs are left in one logical
group operating on the upstream(s) created in step 132 with burst
profiles tailored for 2.0 SCDMA or ATDMA operation.
[0128] An optional step 140 is performed in some subspecies of this
species to create a separate upstream for each downstream. The
burst profile of this new 2.0 upstream will be tailored for DOCSIS
2.0 ATDMA operation or the first 2.0 upstream is SCDMA and will be
SCDMA if the first 2.0 upstream is ATDMA. Then some or all of the
2.0 CMs are moved to this upstream(s) and ordered to operate in the
appropriate mode.
[0129] In step 142, each CM's data transmissions post registration
are monitored for one or more throughput ability factors previously
named to determine if any CM is overperforming or underperforming
its upstream. Step 144 represents the process of creating one or
more separate logical groups within each existing logical group for
overperformers and underperformers. That is, overperformers may
have one or more subgroups and upstream channels created for them
with appropriate burst profiles, and underperformers may have one
or more subgroups and upstream channels created for them with
appropriate burst profiles.
[0130] In step 146, a new upstream is created for each
underperforming subgroup within an existing logical group. The new
upstream has lower throughput, less complex modulation, and/or more
aggressive forward error correction burst profile parameters to
allow more reliable communications with the CMs in the
underperforming group. Once these new upstreams are formed,
messages are sent to each underperformer moving it to the
appropriate new upstream.
[0131] In step 148, a new upstream is created for each
overperforming subgroup within an existing logical group. The new
upstream has higher throughput and/or less aggressive forward error
correction burst profile parameters to allow more reliable
communications with the CMs (and consume less overhead) in the
underperforming group. Once these new upstreams are formed,
messages are sent to each overperforming CM moving it to the
appropriate new upstream.
[0132] Step 150 is optional and defines subspecies within this
species to continually monitor the performance of each CM in each
group and subgroup and do further subgrouping based upon
performance if appropriate. Step 152 marks the end of this
process.
Species 3: FIG. 5
[0133] Species 3 is similar to species 2 except that all DOCSIS 1.0
and 1.1 CMs are grouped in the same 1.X logical group and all
DOCSIS 2.0 CMs are grouped in the same logical group operating on
an upstream created with a burst profiled tailored for SCDMA
operation. The invention comes with monitoring post registration
data communications for throughput ability parameters such as BER,
PER, received signal power, SNR, etc. to determine if any CMs are
overperforming or underperforming.
[0134] The process is the same as species 2 for steps 130, 134 and
136 to get all CMs registered. However, in step 133, only a 1.x
upstream for all the 1.0 and 1.x CMs is created for each downstream
group. Also in step 133, a 2.0 ATDMA or SCDMA upstream is created
for each downstream group. Then step 137 is performed to make sure
that all DOCSIS 1.x compatible CMs have selected an upstream that
has a burst profile tailored for the DOCSIS 1.0 modulation profile.
Step 137 also makes sure all DOCSIS 2.0 compatible CMs have
selected an upstream that has a burst profile that is tailored to
DOCSIS 2.0 SCDMA operation. The chances of this not happening are
minimal since every CM locks onto a downstream and then picks the
upstream associated with that downstream which is most compatible
with the CM's capabilities. Therefore, step 137 is eliminated in
some species.
[0135] Next, step 142 is performed to monitor each CM's post
registration data transmissions for quality as indicated by the
value(s) of one or more throughput ability factors previously
identified. Thereafter, steps 144, 146, 148 are performed as
previously described for species 2. In step 144, one or more
subgroups is created for the overperformers and one or more
subgroups is created for the underperformers. In step 146, a new
upstream channel with an appropriate burst profile is created for
each new underperforming subgroup created in step 144 and a new
upstream channel with an appropriate burst profile is created for
each new overperforming subgroup created in step 144. Optional step
150 can also be performed in species where continual monitoring and
further subgrouping or cancellation of upstreams and regrouping of
CMs is performed based upon changing conditions.
Logical Channels Based Upon Cable Nodes
[0136] Usually an upstream receiver is coupled to more than one
upstream through multiple cable nodes. In the prior art, combiners
have been used to combine the upstreams from cables from different
cable nodes onto a single cable coupled to the upstream receiver.
This has the unfortunate effect of summing the noise on all the
input cables and outputting the summed noise on the single cable.
This reduces the throughput of all CMs coupled to all the input
cables.
[0137] The invention solves this problem by creating a separate
logical upstream channel for the CMs of each cable node. The
upstream for each cable node will have a burst profile based upon
the SNR that the cable node has without the additional noise from
other cable nodes. Each cable node is connected via a switch with
the switch being activated only when a burst is expected from the
cable node. The switch is switched during the gap between bursts.
The gap between bursts can be the guard time that available at the
end of a TDMA burst or a quiet time that is allocated by the CMTS
via a null SID.
[0138] FIG. 6 represents the modified circuitry of a CMTS which
uses switches controlled by the CMTS to eliminate the aggregated
noise of the combiner. A combiner 200 in the CMTS is coupled to
switches 202 and 204 on the CMTS upstream line cards that receive
signals in DOCSIS upstreams transmitted on cables 206 and 208 from
two cable nodes (not shown). FIG. 7A is a timing diagram of the
timing of control of switches 202 and 204 in FIG. 6 for TDMA mode
bursts so as to eliminate the effect of the combiner. The CMTS
control logic controls switches 202 and 204 such that switch 202 is
closed during the gap just before a TDMA burst on an upstream on
cable 208 is expected and opened again during the gap just after
the burst. Switch 204 is left open so that only the burst on cable
208 is gated through to the combiner 200. Likewise, when a TDMA
burst is expected on an upstream on cable 206, switch 204 is closed
during the gap just before the expected burst and opened again
during the gap just after the burst. Switch 202 is left open so
that only the burst on cable 206 is gated through to the combiner.
This eliminates the aggregation of noise from cables 206 and 208
onto output cable 210. The CMTS controls when each CM may transmit
by grants on the MAP messages of each upstream, so it knows when to
expect a burst on each upstream to which it is connected.
[0139] Cable nodes that have logical channels of SCDMA do not
require a gap between bursts for switching cable nodes because the
SCDMA spreader on bursts in different cable nodes are synchronized
in time to each other, and there is no interference between bursts.
The gap between SCDMA bursts has zero duration. Therefore, the
switches must be operated during the ramp up and ramp down portions
of SCDMA bursts since there is no gap between bursts. In order to
process the ramp up and ramp down of the bursts from the two cable
nodes, the switches 202 and 204 of the two cable nodes should be
enabled during the ramp up and ramp down of the expected SCDMA
bursts as shown in FIG. 7B. If SCDMA burst #1 is received on cable
208, switch 202 is closed at time T1 during the ramp up of burst 1,
and switch 204 is left open. Burst 2 is expected to ramp up at time
T2, so switch 204 is closed at that time, but switch 202 is still
closed until the ramp down of burst #1 at time T3. Since the ramp
up and ramp down overlap duration is small, the effect of the
increased noise caused by the summation of the two cable nodes is
small.
[0140] If the CMTS is a single channel receiver that can only
handle one burst at a time, it will coordinate the MAP messages so
that there is never an overlap in time of TDMA bursts even though
the different upstreams are independent of each other. If the CMTS
is a multiple channel receiver that can process more than one burst
simultaneously, and overlapping bursts can happen in the MAPs, each
line card will have a buffer (not shown) therein to store any
overlapping burst for a time long enough that another burst can be
gated through to the combiner, and then the switch on the line card
with the buffered burst will be closed and the buffered burst sent
through to the combiner with all other switches in an open
state.
[0141] Although the invention has been disclosed in terms of the
preferred and alternative embodiments disclosed herein, those
skilled in the art will appreciate possible alternative embodiments
and other modifications to the teachings disclosed herein which do
not depart from the spirit and scope of the invention. All such
alternative embodiments and other modifications are intended to be
included within the scope of the claims appended hereto.
* * * * *