Analysys Mason considers that NBN Co’s design of its FTTP network reflects an efficient and
prudent network design.
In particular:
► Technology decisions
NBN Co’s decision to implement Gigabit passive optical network (GPON) for the massmarket segment prudently implements the requirements of the Australian Government under
its Statement of Expectations.
NBN Co’s decision to use point-to-point (P2P) technology for the delivery of services to the
enterprise and government segment is a prudent design choice for the supply of higherbandwidth and symmetrical services to large government and enterprise customers, and
represents international best practice.
NBN Co’s choice of Ethernet as a Layer 2 protocol is both efficient and prudent, as the choice
of Ethernet aligns with global standards and is a proven technology, and will facilitate
competitive vendor pricing and minimises technology risk/risk of stranded assets.
2
Available at http://www.nbnco.com.au/our-network/industry-consultation.html.
Review of the efficiency and prudency of NBN Co’s fibre, wireless and satellite network design | 3
Ref: 35497-382
► Architecture-related decisions
NBN Co’s adoption of a centralised GPON architecture is both efficient and prudent, as it
represents the best choice of architecture from a long-term cost-management perspective and
from a network scalability and flexibility perspective.
NBN Co’s network design is prudent from a resiliency perspective for the following reasons:
– the design of NBN Co’s distribution fibre network is based on a ring topology, which
provides path diversity from the fibre access node (FAN) to every fibre distribution hub
(FDH) and will prevent any single fibre cut within the distribution fibre network from
being service affecting
– NBN Co could implement all standardised GPON protection option types, if required,
using its proposed architecture
– NBN Co plans to have at least two independent entry/exit locations in each FAN, which
represents best practice and will ensure that each segment of the rings is diversely routed
– a centralised architecture provides greater flexibility in the implementation of protection in
the FDH as it is easier to design a ring topology around fewer sites hosting splitters.
NBN Co’s choice of ribbon technology for fibre cables is both efficient and prudent for the
following reasons:
– ribbon technology is modular and can provide adequate fibre counts for all parts of the
network, standardising cable size and associated deployment processes
– ribbon technology minimises operational expenditure (opex) as it allows the operational
team to deal with bundled fibres simultaneously rather than as single individual fibres
– each fibre in a ribbon is colour-coded, which mitigates against human connection errors,
thereby minimising opex
– fibre ribbon suits the pre-connectorised system being used by NBN Co as part of its fibre
network roll-out
– fibre ribbon cable is also better suited for aerial deployment (where required) because it
weighs 60% less than a traditional stranded fibre cable, maximising the number of existing
poles that can be potentially used for FTTP deployment
– fibre ribbon cable is extensively deployed by leading FTTP operators internationally,
including Verizon, NTT and Korea Telecom.
NBN Co’s proposed end-to-end service availability target of 99.9% is prudent from a network
design perspective, having regard to the geography of Australia and specifically due to the
significantly longer fibre runs in Australia compared to most overseas jurisdictions.
► Infrastructure-related design decisions generally
As NBN Co has a mandate to provide services to 100% of the Australian population through a
combination of FTTP, fixed wireless and next generation satellite technology, it is important
that a prudency and efficiency analysis has regard to this fact. Analysys Mason considers that
Review of the efficiency and prudency of NBN Co’s fibre, wireless and satellite network design | 4
Ref: 35497-382
the methodology used by NBN Co to determine the boundary between the FTTP network and
the fixed wireless network is both prudent and efficient, as it will ensure that a maximum
number of end users are covered by the FTTP network, while at the same time not resulting in
NBN Co incurring disproportionate costs in the relevant circumstances. In particular, based on
NBN Co’s estimate of FTTP coverage of 92.3% for existing premises and 93% of existing and
future premises (taking account of population growth), NBN Co will meet the Australian
Government’s minimum fibre coverage obligation of 90% and its objective of connecting 93%
of premises with fibre. In other words, NBN Co’s decision to set the reach of the fibre network
at 92.3% for existing premises (and at 93% when taking account of both existing and future
premises) serves as an efficient breakpoint for determining the boundary of the fixed and
wireless network footprints.
NBN Co’s decision to re-use Telstra’s infrastructure is prudent from an operational
perspective. There are strong operational reasons to use underground infrastructure wherever it
exists and is fit for purpose. The re-use of Telstra’s existing infrastructure will provide more
certainty (and therefore reduce risks) in a number of areas, including significantly reducing the
need for NBN Co to construct its own duct infrastructure (which would increase NBN Co’s
construction costs and delay the roll-out of its fibre network). It will also overcome some of
the downside that may be associated with aerial deployments, such as lower levels of
reliability and higher associated opex. We also welcome the provide-or-pay (PoP) provisions
contemplated in the NBN Co–Telstra deal, which will further increase the certainty of the
available infrastructure.
► Infrastructure-related design decisions at the end-user premises and the local fibre network
NBN Co’s design of the local fibre network is prudent, as it uses a standard design for FTTP
with a centralised architecture.
NBN Co’s approach to the architecture and features provided on the network termination
device (NTD) in areas served by fibre infrastructure is prudent, as it will allow simultaneous
delivery of multiple applications and services by multiple service providers (SPs) and is
consistent with industry best practice.
NBN Co’s decision to provision a single fibre in the local fibre network for the initial service
connection to the premises, along with a second fibre to meet future capacity requirements in
respect of the relevant premises (e.g. to take account of subdivision of the relevant property),
is both efficient and prudent, as we would recommend a strict minimum of two fibres per
premises in the local fibre network for operational, growth and potential protection reasons.
While the number of fibres that are needed in the local fibre network to cover non-addressable
premises is challenging to evaluate at this point, we consider that NBN Co’s overall
provisioning of fibre in the local fibre network is prudent.
NBN Co’s decision to pre-build the final drop is efficient and prudent, having regard to current
levels of broadband penetration in Australia and the deal between NBN Co and Telstra, which
provides for the migration of end users from the public switched telephone network (PSTN)
Review of the efficiency and prudency of NBN Co’s fibre, wireless and satellite network design | 5
Ref: 35497-382
and hybrid fibre coaxial (HFC) network to the national broadband network (NBN). It is
reasonable for NBN Co to assume a take-up profile of 70%3 in light of these factors and a
decision to pre-build the final drop is the most cost-effective approach.
NBN Co’s decision to implement FDHs using street cabinets is a prudent choice, as it provides
greater levels of flexibility over time than underground splitter enclosures and is also
consistent with the approach that is implemented by the majority of operators using centralised
GPON architectures worldwide.
► Infrastructure-related design decisions in relation to the fibre distribution network
NBN Co’s design of the distribution network is prudent, as it allows for different levels of
protection to be implemented in the FTTP network, which will ensure high levels of resiliency.
NBN Co’s decision to re-use existing Telstra local exchanges for the FAN is both efficient and
prudent, particularly in the context of NBN Co’s deal with Telstra. As all ducts in Telstra’s
distribution network come back to local exchanges, the use of local exchanges as FAN sites will
minimise additional civil works that would otherwise be required as part of the roll-out, resulting in
cost savings relative to a situation where NBN Co was constructing its own facilities.
NBN Co’s approach to defining the size of fibre serving areas (FSAs) (which sets a maximum
size of 38 500 geocoded national address files or GNAFs) is prudent, having regard to
NBN Co’s deal with Telstra and NBN Co’s decision to use Telstra’s exchanges as FAN sites
and the geographical reach of GPON and P2P technology.
NBN Co’s decision to use an optical fibre distribution frame at FAN sites to connect to the FDH is
prudent, as it will provide a higher level of flexibility than an optical consolidation rack.
► Infrastructure-related design decisions in the fibre transit network
NBN Co’s design and architecture for the transit network is prudent for the following reasons:
– use of wavelength division multiplexing (WDM) technology is prudent as it is a mature
technology, which has been adopted by most operators in their core networks throughout
the world to minimise the number of fibres to be deployed
– a single dark fibre pair leased from Telstra will be sufficient to carry traffic in different
sections of the transit network for the medium and long term, thereby minimising opex
– NBN Co has adopted a ring topology, which adequately addresses the requirements for a
resilient transit network and provides an optimal solution for linking the points of
interconnection (PoIs) and the FANs
– NBN Co’s implementation of an overlapping physical ring topology is also prudent as this
makes the most efficient use of available infrastructure without compromising the
resiliency of the network.
3
The expected penetration of 70% does not take into account the agreement between NBN Co and Optus which
should increase the penetration further.
Review of the efficiency and prudency of NBN Co’s fibre, wireless and satellite network design | 6
Ref: 35497-382
NBN Co’s intention to dimension each TC_4 AVC to a minimum of 150kbps is prudent,
particularly in light of the forecast average fixed download volume for NBN Co services up to
2013. Our own analysis of busy hour estimates provide that the bandwidth per TC_4 AVC
should be around 166kbps, which suggests NBN Co’s own initial dimensioning is broadly
consistent with our own calculations.
NBN Co’s implementation of a semi-distributed PoI architecture is consistent with the
requirements of the Australian Government in its Statement of Expectations and has been prudently
implemented by having regard to the availability of competitive backhaul in accordance with the
ACCC’s ‘competition criteria’ and by having regard to duct space, power and cooling.
► Future-proofing of NBN Co’s fibre network
NBN Co’s network design is likely to have a sufficient upgrade path to meet the reasonably
anticipated requirements of access seekers and end users for bandwidth over the next 30 years.
In terms of bandwidth evolution, the GPON standard has a clear evolution path as the
downlink bandwidth can be upgraded from 2.5Gbps to 10Gbps.
While it is difficult to predict how the technology will evolve in the next 30 years, we have not
found any bottlenecks in the choice of the technology or design of the physical network that
would mean the network cannot be upgraded in terms of bandwidth or functionality for the
fibre network. In 2010, on behalf of Ofcom, Analysys Mason undertook a large-scale study
regarding the capacity limitations in fibre access networks.4 In that study, we concluded that
we did not believe that capacity will be the main limiting factor in GPON fibre access
networks, and we do not foresee a situation where supply is unable to meet the growing
demand of users. Instead, the study suggests that the bottlenecks in the access network may be
in the operational upgrade of one generation of FTTP technology to the next, but we are
confident these issues will be resolved in time.
The proposed GPON architecture is future-proof, especially regarding the dimensioning of the
local fibre.