802.3bn Bright House
Networks per-node SNR
Downstream Characterization
Edwin Malle+e 04/20/2012 Introduction
!   We’ve had a lot of discussion about this topic
of “multiple simultaneous Modulation and
Coding Schemes.”
!   We’ve even seen some average receive
Signal to Noise Ratio values in some
contributions.
!   This contribution goes a little deeper to
depict the average SNR values at a per-HFCNode grouping.
!   This presentation intends to provide the
distribution of SNR for stations on a many
coax cable distribution network (CCDN) –
HFC Node Serving Groups.
2 Hybrid Fiber Coax: A Review
!   Fiber to HFC Node, AM modulation, then active coax
network past the node.
!   CCDN extends from the node to the connected station.
!   The data in this presentation only includes the average
SNR (reported by the Cable Modem) on the DOCSIS
downstream carriers.
!   The per-node grouping allows a discussion about the
distribution of stations on the CCDN.
AMP
AMP
Coax
CATV Services
RF
Combining
Coax
Optical
TX & RX
TAP
Fiber
AMP
AMP
CMTS
TAP
TAP
TAP
HFC
Node
TAP
TAP
MTA
AMP
AMP
TAP
Coax
Splitter
Cable
Modem
Coax
Drop
TAP
Digital TV
Set Top
Downstream SNR for all CMs within an HFC Node 3 Analog TV
CMTS vs EPoC SNR
CMTS data provides Total SNR end to end on a Node basis EPOC may operate only on the Coax segment SNR on a Coax Segment basis Hub / CMTS Hub / CMTS Node OCU Node 4 Procedure
!   We have a tool that periodically queries cable-modems
for many performance values including downstream
SNR.
!   We took several polled intervals over a one-week
period.
!   We calculated the number of cable-modems within an
HFC node grouping that reported an SNR values…
§  E.G. 30+SNR, 35+SNR, 36+SNR, etc.
!   We found that the number of modems within a node
grouping reporting a given SNR value varied very little.
!   As a result the SNR values reported come from a
single polling cycle (as opposed to per-modem
average SNR over some period.)
!   The reported CMs include all in the population – DSGs,
MTAs, DOCSIS1.1, 2.0 and 3.0 CMs.
5 Average SNR with 2 Standard
Deviations
CABLE MODEM POPULATION -­‐ DOWNSTREAM SNR 100.00% 90.00% 80.00% PERCENTAGE 60.00% 50.00% 40.00% 2-­‐SD 1-­‐SD 70.00% 30.00% 20.00% 10.00% 0.00% 30+SNR 35+SNR 36+SNR 37+SNR 38+SNR 39+SNR 40+SNR AVG 99.76% 90.71% 82.86% 68.78% 50.32% 28.83% 13.69% 1SD(+) 100.00% 97.09% 93.72% 84.11% 68.05% 44.57% 23.57% 1SD(-­‐) 99.32% 84.32% 72.01% 53.45% 32.59% 13.10% 3.81% 2SD(+) 100.00% 100.00% 100.00% 99.44% 85.77% 60.31% 33.45% 2SD(-­‐) 98.88% 77.94% 61.15% 38.12% 14.86% 0.00% 0.00% 6 SNR versus Node Size
Perentage of total PopluaKon DistribuKon of Node Grouping Sizes 16.00% 14.00% 12.00% 10.00% 8.00% 6.00% 4.00% 2.00% 0.00% !   Distribution of Node
Grouping Sizes forms a
nice bell curve.
!   The outlier is the >500
CMs number which
includes node grouping
sizes between 500CMs
and 848 CMs.
CMs per Node DistribuKon of Avg SNR to Node Grouping Size Signal to Noise !   On the surface it is
questionable that there is
a direct correlation
between small node
groupings and average
SNR.
!   It is somewhat apparent
that the very small (less
than 50 CMs per node)
does have a consistently
better downstream SNR.
38.00 37.80 37.60 37.40 37.20 37.00 7 % of CMs SNR Capabilities vs Node
Grouping Size
% of CMs SNR vs Node Grouping Size 100.00% 90.00% 5-­‐49 CMs 80.00% 50-­‐99 CMs 70.00% 100-­‐149 CMs 150-­‐199 CMs Axis Title 60.00% 200-­‐249 CMs 50.00% 250-­‐299 CMs 300-­‐349 CMs 40.00% 350-­‐399 CMs 30.00% 400-­‐449 CMs 450-­‐499 CMs 20.00% 500+ CMs 10.00% 0.00% 30+db 35+db 36+db 37+db 38+db 39+db 40+db 8 % of CMs SNR Capabilities vs Node
Grouping Size
% of CMs SNR vs Node Grouping Size 30+db 100.00% 35+db 90.00% 36+db 80.00% 37+db Axis Title 70.00% 60.00% 38+db 50.00% 40.00% 39+db 30.00% 20.00% 40+db 10.00% 0.00% 5-­‐49 CMs 50-­‐99 CMs 100-­‐149 CMs 150-­‐199 CMs 200-­‐249 CMs 250-­‐299 CMs 300-­‐349 CMs 350-­‐399 CMs 400-­‐449 CMs 450-­‐499 CMs 500+ CMs 9 Further Analysis (SNR)
100% CMs 30db Percentage of Nodes 63.08% 35db 2.53% Average Node Size 241CMs/Node 18CMs/Node Average SNR >97.5 <100% CMs 37.54db 30db 38.65db 35db Percentage of Nodes 36.33% 3.95% Average Node Size 323CMs/Node 219CMs/Node Average SNR > 95 < 97.5 CMs 37.12db 30db 38.42db 35db Percentage of Nodes .03% 43.9% Average Node Size 123CMs/Node 291CMs/Node Average SNR 35.99db 37.54db !   The first table is 100%
of Cable Modems on
that node that support
30db and 35db SNR.
!   The second table is of
> 97.5% < 100% of
Cable Modems that
support 30db and 35db
SNR.
!   The next table is of
>95% but < 97.5% .
!   The last is of the worst
10% of nodes
Worst 10% of Nodes Percentage of CMs on a <84.06% node Capable of >35db Average Node Size 254.36 Average SNR 10 35.93 More Work to Come
!   With the small nodes, it’s not clear (from our polling data) how large
the CCDN is or how many passings we have on a node.
§  The nodes could have a small number of CMs but be far apart, for
instance.
§  Alternatively we could have a large number of passings but an
abnormally low penetration number in some cases.
!   Something we’re looking at is the how far the cable-modems are
from the CMTS.
§  Perhaps timing offset would help.
§  Or we could simply map the individual cable-modem locations to the
plant (this is harder.)
!   Its also not clear what the impairments that reduce SNR for some
CMs and not others.
§  It could be noise isolated to the home.
§  It could also be compound noise near the end of line.
!   We’re also investigating what the downstream SNR would like if we
provided digital lasers to a device past the node (SLIDE 4).
§  Does it improve ?
§  Significantly ?
11 Released Data
!   We’re can release some samples of our node
grouping data in excel if that would be helpful
!   This data would include:
§  20 of our best performing nodes for SNR.
§  20 of our nodes that are centered around the average
SNR numbers.
§  20 of our worst performing nodes for SNR.
!   Examples are below:
NodeNAME TotalCMs AvgSNR 15+5SNR 20+5SNR 25+5SNR 30+5SNR 33+5SNR 34+5SNR 35+5SNR 36+5SNR 37+5SNR 38+5SNR 39+5SNR 40+5SNR
BEST_1
12
40.4
12
12
12
12
12
12
12
12
12
12
11
8
BEST_2
13
40.2
13
13
13
13
13
13
12
12
12
12
11
8
BEST_3
20
40.2
20
20
20
20
20
20
20
20
19
19
17
9
BEST_4
56
39.8
56
56
56
56
55
55
54
54
50
45
44
38
BEST_5
59
39.8
59
59
59
59
58
58
57
57
53
49
44
31
BEST_6
168
39.8
168
168
168
168
167
166
166
164
152
146
120
96
BEST_7
10
39.7
10
10
10
10
10
10
10
10
9
8
8
7
BEST_8
13
39.7
13
13
13
13
13
13
13
13
13
13
11
5
BEST_9
31
39.7
31
31
31
31
31
31
31
31
30
29
25
19
BEST_10
54
39.7
54
54
54
54
54
54
54
51
49
45
36
27
12 Thank You!