Fiber Access Network

A Cable Operators Perspective

by
John A. Brouse, Jr.
Director of Network Implementation
Charter Communications, Inc.

Introduction
During the late 1980s, cable television operators in the United States reached a critical
point where bandwidth demands and physical extensions to their coaxial-based, tree and
branch service delivery networks stressed the technical limits for signal quality and
platform reliability. In response to the cable operators demand for a solution that
addressed both signal quality and reliability, the industry looked toward adapting existing
fiber optics technologies to accommodate frequency division multiplexed analog video
distribution. The initial application of broadband fiber based platforms was targeted to
simply reduce the cascading effects of active electronic devices. Within a short time, the
cable industry migrated fiber beyond this initial application and considered other services
beyond traditional broadcast video. As other economic forces emerged, such as the
launch of direct broadcast satellite service and the threat of telcos entering the video
market, cable operators recognized the need to evolve their networks to support two-way
interactive services and growing customer expectations. The result was development of a
hybrid network which integrated fiber optic cables and devices with coaxial cables and
RF based devices as a function of service based network requirements. This architecture
was named the Hybrid Fiber-Coax (HFC) network and moved fiber closer to the
customer, usually supporting service groups of 500 homes or less. The current list of
services which the cable television multi-system operator (MSO) is offering include
traditional video entertainment, video on demand, high speed connections to the Internet,
local and long distance telephony, with many more in development. While deployment
of the HFC network has certainly been the principal enabler for these high end two-way
services, it has not adequately addressed nor resolved other operational issues now
coming to the forefront. Todays two major concerns are

escalating outside plant operation and maintenance costs negatively impacts cash
flow growth
evaluating and readying outside plant for new service launches negatively impacts
product time-to-market

The Hybrid Fiber-Coax Network

Those unfamiliar with the MSOs services delivery platform often assume that the
introduction of fiber into the outside plant meant a reduction in the amount of coaxial
cable used. This is not the case. The initial deployment of fiber in the distribution
network was the addition of fiber cables overlaid on the existing coaxial cables. The
result was more cable to contend with not less. In fact, there are some HFC variations
that increased the amount of coaxial cables used rather than reducing it. The focus then
was on being able to reduce the number of customers sharing a given amount of
bandwidth for both the downstream and upstream portions of the RF spectrum. As these
serving areas became smaller in size, the signal quality and service reliability
immediately improved as a result of there being less things between the signal source
and the customer. However, as a cumulative effect, more elements were added to the
network. As a direct result, both preventative and demand maintenance requirements
increase; as well, more active devices are being added to the network resulting in everrising electric power consumption rates. The initial capital investment requirements for
the coaxial portions of the HFC network have stabilized with very little room for
lowering cost; the fiber based portions of the network will see some level of cost
reductions but will have a negligible effect, at least in the near-term. The focus then, is
on deployment costs and operations costs.
From the perspective of outside plant assessment and readiness to launch new services,
the incremental costs and time lines are a function of the type of service being launched.
For those services that have low or no impact on interactive sessions, little to no effort is
required. However, for new launches demanding high usage rates and superior reliability
the cost and time to ready the HFC platform can be significant. This may include service
node resizing (either internally or externally), headend RF combining and splitting
(mux/demux) network reconfiguration, service drop conditioning, and/or network
powering modifications. The time and effort required to complete the pre-launch work is
a function of the current state of the outside plant combined with any additional network
requirements demanded by the new service. Since each operating systems outside plant
may be at a different state of readiness, a universal impact model has not yet been
established. However, with varying degrees of additional effort, the HFC plant can be
readied for the successful launch of new services. These pre-launch network readiness
efforts include work focused on the headend and the outside plant. Depending on the
characteristics of the service, the customer drops may also require readiness
consideration.

HFC Operating Cost Assessment

The assessment model selected is a typical HFC network delivering multiple two-way
services. The system serves nearly 35,000 customers and operates 1,410 miles of outside
plant. The network contains 199 nodes plus an additional 4,380 RF amplifiers. 584
outside power supplies are required to operate the network with each having three 12 volt
batteries for stand-by use. The system experiences a monthly customer trouble call rate

of 2.5% with half related directly to the RF portions of the network including the drop.
The following table details the annual expenses incurred, normalized to cost per outside
plant mile, related to operating and maintaining the outside plant. Rather than applying
the actual costs incurred which are specific to this system and its location, statistical
industry averages are used. This allows the model to be applied more generically and is
believed to minimize anomalies that may be introduced by specific employee behavior,
the specific operating environment or local cost of living index.
Technical Supervision
Service Trouble Truck Rolls
Plant Maintenance Truck Rolls
Material Inventory
Electric Utility Bills
Power Supply Battery Replacement
Power Supply Repairs
RF Line Equipment Repairs
Vehicle Accident Loss
Worker Injury Loss
Emergency Cable Repair
Total Annual Operating Expense per Plant Mile

$
$
$
$
$
$
$
$
$
$
$
$

42.03
226.15
235.50
49.64
446.81
43.49
1.77
35.46
8.80
5.01
8.51
1,103.17

FTTH Operating Cost Assessment

To completely address the primary operational issue, escalating O&M costs, the
assessment model selected is the passive optical network (PON). For this effort, the
outside O&M costs are PON specific (APON, EPON, etc) agnostic. For comparative
purposes, the system used is the same one identified for the HFC cost model above.
Based on the passive attributes of the PON, the OSP network elements are limited to fiber
optic cables, optical couplers and WDMs. Accordingly, the major identifiable cost is
emergency cable restoration. Due to the shift in work requirements, cable restoration
would be contracted to a third party resulting in a significantly higher restoration cost per
mile. The following table details the annual expenses incurred, normalized to cost per
outside plant mile, related to operating and maintaining the outside plant.
Technical Supervision
Service Trouble Truck Rolls
Plant Maintenance Truck Rolls
Electric Utility Bills
Power Supply Battery Replacement
Power Supply Repairs
RF Line Equipment Repairs
Vehicle Accident Loss
Worker Injury Loss
Emergency Cable Repair
Total Annual Operating Expense per Plant Mile

$
$
$
$
$
$
$
$
$
$
$

0
0
0
0
0
0
0
0
0
85.11
85.11

In addition to dramatically reducing the outside plant O&M requirements, deployment of

the PON to the home provides the highest level of reliability and negates the need for
future physical plant upgrade as bandwidth requirements increase. As well, services can
be monitored and tested to each customer thereby providing immediate fault notification.
And any new service launch plan is no longer dependent on outside plant modifications
which then shortens the preparation time preceding the service launch date.
There does remains, however, the need to conduct a headend readiness assessment and
complete any required pre-launch preparations. While these do impact the new service
time-to-market equation, for the purpose of this HFC FTTH comparison, it is assumed
the amount of effort and time required to ready the headend for the new service launch
closely approximates parity. Accordingly, the focus for the remainder of this study is lifecycle costs.

The Capital Investment Assessment

Now that we have examined the annualized operating and maintenance costs associated
with the HFC and FTTH network for the MSO, the looming question becomes How
does the capital investment requirements for deploying the FTTH PON today differ from
the cost to deploy the HFC network? or Is it economically feasible to take advantage of
the FTTH PON today to reduce operating expenditures?

HFC Deployment Costs Assessment

The costs to deploy the HFC network are categorized according to where the various
assets will be located. Specifically, they are grouped as being housed in the headend,
comprising the outside plant or defining the customers drop. Further, each grouping
identifies those costs that are related to materials and those related to labor. And, as in
the O&M assessments, the following table presents the identified capital costs in a
normalized fashion based on investment cost per mile of outside plant.
Outside Plant Deployment
Materials
Labor
Total OSP deployment Cost per Mile

$12,273.93
$16,408.61
$28,682.54

Headend Equipment Deployment

Materials
Labor
Total Headend Equipment Deployment Cost per Mile of OSP

$
$
$

736.51
83.50
820.02

Drop (Material & Labor) per Customer Cost

125.00

FTTH PON Deployment Costs Assessment

For consistency, the costs to deploy the PON are also categorized according to where the
various assets will be located. Specifically, they are grouped as being housed in the
headend, comprising the outside plant or defining the customers drop. Further, each
grouping identifies the costs to materials and those related to labor. And, as in the O&M
assessments, the following table presents the identified capital costs in a normalized
fashion based on investment cost per mile of outside plant.
Outside Plant Deployment
Materials
Labor
Total OSP deployment Cost per Mile

$16,505.22
$ 9,578.79
$26,084.01

Headend Equipment Deployment

Materials
Labor
Total Headend Equipment Deployment Cost per Mile of OSP

$15,910.00
$ 208.76
$16,118.77

Drop (Material & Labor including an RF only ONU and power

supply) per Customer Cost

748.00

Life-Cycle Cost Comparison

To determine the near-term financial viability of FTTH deployments compared to the
current HFC architecture, the capital investment costs are combined with the annual
operating and maintenance costs to develop a simple life cycle cost analysis. Factors not
included in the early analysis are asset depreciation schedules, time value of money, or
end of life residual value. Of note, data presented in the immediately above two sections
indicate the CPE costs for FTTH are significantly higher than those for the HFC
platform. For this reason, there must be a significant increase of revenue potential from
the average to offset this cost in a reasonable timeframe. Accordingly, neither the
customer drop nor the CPE is included in the current life cycle cost assessment. In
developing the life cycle cost comparison, consideration was only given to those costs
that would need to be incurred regardless of penetration rates and those costs only
associated with delivering the service signal from the headend to the service drop and
from the service drop to the headend.
Following is the life cycle costs comparison presented in graphical format. Considering
only those factors already discussed, the analysis indicates that the HFC network is more
cost efficient until year 13. After 13 years, the FTTH become the most cost effective
delivery platform. As with the other assessments presented, the life cycle costs are based
on a cost per mile of outside plant, which is consistent with the cable industrys methods
for assessing plant values.

Life-Cycle Cost Analysis

HFC vs FTTH PON

Cost per Mile of Outside Plant

$55,000

$50,000

$45,000

HFC

$40,000

PON

$35,000

$30,000

$25,000
1

9 10 11 12 13 14 15 16 17 18 19 20 21

Years in Operation

Conclusion
While the fiber-to-the-home solution addresses the operational concerns by an
overwhelming margin, the current costs to deploy the FTTH network place it at a
significant disadvantage to todays HFC approach. Other models of the HFC network
that also reduce operating and maintenance costs will be further investigated in the future
as will other approaches for a FTTH solution. Since the MSOs world is based in RF
technology, the FTTH solution assessed did not include replacing the MSOs use of
cable modems with an Ethernet approach to high speed data services. Of interest is that
the OSP deployment cost for the FTTH is lower than for the HFC approach. However,
there is an overwhelmingly significant difference in headend equipment deployment
costs; the HFC architecture is more than an order of magnitude more financially efficient
than the FTTH approach. Also not considered in the operating and maintenance cost
assessments were any cost adjustments associated with operating the headend. With the
dramatically high number of active components in the headend for the FTTH design, it is
reasonable to expect some increase in operating cost; however, on the surface they
initially appear to be orders of magnitude lower than the differences in outside plant
O&M costs.

It should be noted that for the HFC model, 44% of the costs are for material while the
FTTH model realizes 77% of its cost in materials. The life cycle costs currently tips the
scales in favor of the HFC network for cable operators. However, FTTH is still in its
infancy and therefore opportunities for significant FTTH hardware cost reductions will be
realized. Bear in mind these early generation products carry high developmental costs
that are typically factored into early pricing schedules. As identified by this cost
assessment and comparative analysis, two areas where the manufactures must focus in
reducing costs are headend equipment and CPE.
Finally, the FTTH - HFC comparison model may yield differing results if the strict RF
product delivery approach is shifted to an Ethernet to the home approach. When it
becomes possible to consider an all IP delivery platform, a significant percentage of the
FTTH deployment costs for the headend equipment will likely be removed. For the nearterm, MSOs will continue refining the HFC platform and will focus more effort to design
O&M costs out of network by driving fiber closer to the curb - but not to the home.