“Last mile” networks typically refer to the final leg of a communications network that connects service providers to end users. It refers to both the physical infrastructure and technologies responsible for delivering telecommunications services from a provider’s network to individual homes, businesses and institutions (Jayant, 2005). This segment of the network is crucial, as it acts as the final bridge between the broader network ‘backbone’ (wider infrastructure managed by the provider) and end users.
The importance of the last mile lies in its direct impact on the quality, speed, and reliability of telecommunications services. Regardless of the sophistication of the ‘network backbone’, if the last mile of a network is poorly implemented, users will experience disruptions, slow connections, and in some cases – outright service failures. As such, understanding and optimising the last mile, the ‘weakest link’ of a network, is essential for ensuring reliable and seamless communication. Without doing so, telecommunications companies cannot adequately provide end users with reliable network connectivity – ultimately stunting economic, social, medical and technological development (Islam, 2008).
Having such a direct impact on connectivity, paired with the unique requirements of each end user, it’s no surprise last mile infrastructure is extremely complex. Not only must last mile networks find a compromise between expenditure, bandwidth limitations, security concerns and accessibility (Islam, 2008), however, must do so in a way that is scalable to everyone’s individual circumstances. Overcoming these complexities often requires a combination of technological innovation, strategic investments and planning, regulatory bodies and collaborative efforts among stakeholders – a task that often proves difficult in practice.
What are the main, typical, options for delivering Last Mile networks in Australia in 2024?
Modern last mile networking in Australia encompasses a wide range of solutions. Between the re-utilisation of existing copper telephone lines, newly installed fibre optic cabling, and various wireless options (ACMA, 2021) – these many methods are all a result of compromise, seeking to simultaneously reach as many individuals as possible and provide as much bandwidth as possible whilst still being cost-effective. Some common delivery methods include :
Copper Networks
Existing copper telephone lines have been used within various last mile networking solutions. While on the decline in the modern day (NBN Co, 2024a), copper cabling still makes up a large portion of last mile internet delivery – both in the form of traditional DSL/ADSL networks and in tandem with newer fibre optic solutions (FTTC & FTTN). While neither is as fast as direct fibre optic connections (FTTP), they’re viable alternatives – and have played a historical role in providing internet access to Australians, particularly those in regional areas.
Fiber Optic Networks (FTTx)
One of the newer advancements in last mile network solutions, FTTx networks make use of fibre optic cables to enable high speed data delivery. These come in a variety of methods (some of which use the existing copper connections mentioned earlier) and make up the majority of Australia’s last mile networking solutions (NBN Co, 2024a).
- FTTP (Fibre to the Premises) : FTTP involves directly running fibre optic cables from the exchange to the household. It’s most often seen in new developments and densely populated urban areas, providing the fastest consumer internet solution of the options.
- FTTB (Fibre to the Building) : FTTB is a variety of FTTP that is implemented in apartment complexes and office buildings. This involves fibre optic cables being run to a central point within the complex. From here, it is the building management’s responsibility to provide the tenants’ connection to the service.
- FTTC (Fibre to the Curb) : FTTC is often chosen when the cost of deploying FTTP is too expensive. Fibre to the curb runs fibre optics directly to a distribution point near the household (often the ‘curb’), which is then accessed by end users. Then, existing copper telephone cables complete the connection from the distribution point.
- FTTN (Fibre to the Node): Similarly to FTTC, FTTN is typically chosen when FTTP is deemed too costly or prohibitive. FTTN involves running fibre optic cables to street cabinets (nodes) within neighbourhoods, with existing copper lines being used to connect individual premises. This differs from FTTC, as one node acts as a central point for an entire neighbourhood, rather than a single street or property.
HFC (Hybrid Fibre-Coaxial) Networks
HFC networks leverage a combination of fibre optic and coaxial cables to deliver telecommunications services. HFC networks, much akin to FTTC, are seen when fibre cables are run to a central node. From here, the signal is translated onto coaxial cables and runs to the premises. These coaxial cables are typically pre-existing (from PayTV or Cable installations) and offer an improvement in bandwidth and connection reliability compared to FTTN/C solutions.
Fixed Wireless Technologies
Fixed wireless broadband uses radio frequencies to deliver wireless network access to households. This method requires both parties (the end user and telecommunications provider) to install antennas, both of which have line-of-sight connectivity to one another. This is particularly useful in rural or remote areas where laying fibre optic cables is impractical or cost-prohibitive.
Satellite Internet (Starlink & Sky Muster)
Finally, satellite internet services are implemented in extremely remote or regional areas where terrestrial infrastructure (line-of-sight connectivity to a telecommunications tower) is unavailable. Satellite dishes, installed on properties, communicate with orbiting satellites which then relay signals to a telecommunications tower. Satellite internet comes in various forms, including the privately owned ‘Starlink’ and NBN’s ‘Sky Muster’. More often than not, this is the only viable networking option for users in these areas.
What are the inherent (physical) limitations on data rates across typical Last Mile networks?
Data rates of last mile networks vary based on the physical medium through which the data is sent. Some limitations are as follows:
Copper Cables (ADSL, VDSL, FTTC, FTTN)
Copper cabling (being the traditional approach to telephone connectivity) was not installed with the internet in mind – and contains several physical limitations preventing fast data speeds. Notably, a signal over copper attenuates significantly as the cable increases in length, losing energy to both heat and physical resistance. Electrical interference is also an issue, with crosstalk from nearby cables, electromagnetic interference and radio-frequency interference all introducing noise, consequently reducing data throughput. These limitations mean copper cabling can significantly throttle a broadband signal. This is most prevalent on FTTC/N networks, where copper is the ‘weakest link’, throttling the extreme amounts of data that can be delivered by fibre optic cables. (Department of Communications, 2013)
Coaxial Cables (HFC)
While coaxial cables are still ‘copper’ (and thus suffer from similar limitations), their improved shielding and cable design allow them to provide much faster data rates. Here, noise is much less of a problem, as is crosstalk, with HFC cables implementing additional twisting, shielding and foils to reduce noise pollution. Similarly, added insulation means attenuation through heat is lessened – though not negated entirely. Notably, coaxial cabling still suffers from distance limitations, and cannot be effectively run over 500m (CISCO, 2003). Regardless, in comparison to traditional copper cabling, HFC networks enable significantly higher data rates.
Optical Fibre Cables (FTTP, FTTB)
Optical fibre cables are much faster than both copper and coaxial cables. These cables use light, rather than electricity, allowing for speeds of up to 1 Gbps, or even higher. Optical fibre cables are used in FTTx solutions, providing gigabit speeds to both nodes and households alike (ACCC, 2023). Unlike copper and coaxial cables, neither noise nor distance are huge issues when it comes to fibre optics. Whilst the mode (number of light sources running through a cable) do influence data rates (with single-mode fibre offering a further possible cable length), this is trivial when considering their role in a consumer-grade internet solution. Rather, they are mostly limited by external environmental factors or manufacturing flaws.
Fixed Wireless
As fixed wireless technologies rely on radio signals, they have somewhat different shortcomings compared to cabled solutions. Physical limitations of fixed wireless include line-of-sight obstructions (buildings, trees or terrain features), as well as signal interference from nearby wireless devices and current weather conditions. Fixed wireless connections may also suffer from issues relating to the transceivers, be it the one installed at the premises or the link at the telecommunications tower. With such a large range of possible limitations, fixed wireless data rates can range anywhere from ~10Mbps to ~100Mpbs (ACCC, 2023) and often fluctuate with atmospheric or environmental changes.
Satellite Internet (Starlink & Sky Muster)
Satellite internet’s limitations tend to mirror that of fixed wireless. As satellite internet relies on geostationary or low-earth orbit satellites to transmit data – the distance between the satellite and the Earth, atmospheric conditions such as rain or cloud cover, and the alignment of the satellite dish can all directly affect data throughput. Again, while these technologies are constantly improving, they are still prone to huge fluctuations in both latency and estimated speeds.
Describe (briefly) the various approaches that could be deployed and what deployment would look like in each situation
The FTTP Approach
Deploying FTTP, though costly, would offer the highest possible performance of these solutions. According to NBN’s design specifications (ACCC, 2012), FTTP can produce gigabit speeds for over 14 km, allowing properties to see speeds of up to 1000Mbps/400Mbps with minimal latency.
While there are a handful of ways to deploy such a network, Fig 1.1 demonstrates a simple solution – where fibre optic cables are run horizontally down each of the 5 streets. These cables would be run into a DPU (distribution point unit) installed outside of each property, with additional fibre optic cabling installed down every driveway. A rough estimate suggests this solution would require 113.5km of fibre optic cabling, alongside the installation of 99 utility boxes. An expensive solution, but one that would provide unparalleled speeds and reliability.
Fig 1.1 – The implementation of an FTTP solution.
FTTC/N
A more efficient approach would be to deploy a FTTC/N network – running fibre to a common distribution point and making use of the properties’ existing copper cabling to complete the connection.
FTTC Approach
While FTTC is typically favoured over FTTN, when considering the size of the farms, this may not be the most efficient approach. Per NBNs design specifications (ACCC, 2012) (Pearce, 2018), when implementing FTTC, a copper cable should not run more than 150m from the distribution point. This means that each of the farms would require its own DPU (99), with 111.52km of fibre optics needing to be laid. With FTTC requiring this much cabling – the town would be better off purchasing an additional 1.98km and implementing an FTTP strategy.
Regardless, this solution would receive better speeds than FTTN (~100Mbps/40Mbps), reducing both potential network congestion at peak hours, and signal degradation over lengthy copper.
Fig 1.2 – The implementation of an FTTC solution.
FTTN Approach
A more cost-effective approach would be to deploy an FTTN network similar to that in Fig 1.3. Per ACMA (2019), copper connections running at 1050m can deliver download speeds of ~20Mbps. Understanding this, we can map a solution that ensures no property is more than 700m from the node, allowing for speeds of ~30Mbps/10Mbps (Aussie Broadband, 2018) even in peak hours.
This approach would involve running optical fibre cabling to 6 distribution points (nodes). Each node would service 15-20 premises’, with additional copper wiring extending the connection from each house’s driveway to the node. This solution would require an estimated 39.7km of additional copper cabling and 67.9km of fibre optics.
Fig 1.3 – The implementation of an FTTN solution.
HFC Approach
Another option would be to deploy a HFC network. These are unfavourable, however, in comparison to the FTTC approach, as the existing telephone lines would not be usable. Rather, coaxial cabling would need to be installed up the driveway of each of the properties. Per Tunmann (1995), broadband coaxial cabling cannot be run more than 100m without seeing performance degradation – meaning each premise would need to be equipped with a powered fibre-to-coaxial converter. This results in 99 cabinets, 111.52km of fibre optics and 1.98km of coaxial cabling. Were the premises’ to already have coaxial cabling installed, this would quickly become the most feasible solution, being able to provide nearly gigabit speeds (ACMA, 2019)
Fig 1.4 – The implementation of a HFC solution.
ADSL Approach
Finally, an older approach would be to deploy a fully copper network, running over ADSL. This would look similar to the previous FTTP approach outlined in Fig 1.1, but use the existing slower copper cabling rather than fibre optics. Assuming the network uses the newer ADSL2+ standard, the provided upstream data rates would still not come close to those suggested in the requirements (~12Mbps/1Mbps) (ITU, 2009).
4G/5G Wireless Approach
Depending on the existing 4G cell infrastructure, a surprisingly cost-effective method would be to simply allow the farms to make use of the 4G network. Were the existing infrastructure using a mid-band frequency spectrum, the network would effectively cover the entire town, performing at relatively high speeds (Simmons, 2024). This being said, this approach does fall short when considering peak hour traffic – with 4G networking speeds halving (or more!) on a congested network (Fenwick, 2023). Still, this is a temporary solution that would require minimal effort and expenditure, and perhaps one that could be utilised while another approach was implemented.
Fixed Wireless Approach
The more typical wireless solution would be to implement fixed wireless. Deploying a fixed wireless network would involve installing a directional transceiver at each of the properties, all pointing to the exchange. The exchange would then need to be equipped with a larger, omnidirectional antenna, whose radius covered the whole town. Per NBN (2024b), fixed wireless is a viable solution for up to 14km, providing estimated speeds of ~75Mbps/10Mbps. Fixed wireless in this scenario may outclass some cabled options, offering a significantly cheaper and less involved approach to internet access.
Satellite Approach (Starlink & Sky Muster)
Starlink and satellite internet services are also another great option – however, would depend entirely on the location of the town. Assuming the town is based somewhere in Australia, the Starlink service would already be accessible by the townsfolk, offering a cheap and high-performance networking solution. Similarly to fixed wireless, this would involve each property purchasing a directional transceiver, allowing for speeds of up to ~220Mbps/20Mbps (Starlink, 2024). This being said, were the satellites not already existing, the cost of doing so goes up exponentially and becomes a less viable alternative to simply using fixed wireless technologies.
Fig 1.5 – The implementation of a Fixed Wireless solution.
b/c) Pick one cabled and one wireless approach, and explain in more depth for those which you would recommend, and why. Estimate the deployment costs for your two designs.
FTTP Implementation
Based on my earlier research, I had planned to recommend the FTTN design outlined in Fig 1.3. However, after doing some initial cost estimations, it became clear that the FTTP design (Fig 1.1) would be similar in long-term cost, yet provide a vastly superior networking solution.
Implementing an FTTP solution boasts many advantages. Obviously, speed and reliability are big factors. As mentioned, FTTP can support up to 1000Mbps/400Mbps – speeds that will presumably future proof the network for many years to come. This is unlike FTTN, whose 30Mbps speeds may become insufficient in 5 years time. Similarly, the use of high quality fibre optic cabling means connectivity will be extremely reliable, with the ACCC (2023) suggesting FTTP households experience peak performance nearly 100% of the time, compared to ~75% seen by FTTN users. Finally, as outlined below, this option may potentially reduce long-term expenditure, requiring significantly less maintenance than an FTTN implementation.
Nevertheless, it is still critical to acknowledge some of FTTP’s downsides. Notably, implementing an FTTP solution means connecting new properties to the internet will be extremely costly. These properties cannot simply lay copper to the node but rather must run lengthy fibre optics to the central exchange. Similarly, though more sturdy, if something were to happen to the fibre optic cabling, there would be greater costs involved in replacing and fixing them.
Pricing Estimate of FTTP Deployment & Operations
Based on the FTTP Design outlined in Fig 1.1
| Item | Price | No. Required | Total |
| DPU | $99 per unit | 99 | $9,801 |
| Fibre Cabling | $10/m | 113,500 | $1,135,000 |
| Fibre Termination | $500 per cable | 198 | $99,000 |
| One Off Deployment Total | $1,243,801 | ||
| Fibre Maintenance
per Annum |
$31 per household | 99 | $3,069 |
| Operational Total Per Annum | $3,069 | ||
| Operational Total over 30 Years | $92,070 | ||
| Total estimated cost, including deployment and operational costs over 30 years. | $1,335,871 | ||
Assumptions Made :
- Per the CTC Technology and Energies Report (2015), newly placed fibre will last 30 years if provided with respective maintenance.
- Per the CTC Technology and Energies Report (2015), fibre optic maintenance costs are averaged at 31 AUD per ‘passing’ (household), annually.
- No major environmental disruptions or un/intentional vandalism during this time.
Comparative Pricing Estimate of FTTN Deployment & Operations
Based on the FTTN Design outlined in Fig 1.3
| Item | Price | No. Required | Total |
| Fibre Cabling | $10/m | 67,900 | $679,000 |
| Fibre Termination | $500 per cable | 99 | $49,500 |
| Copper Cabling | $3/m | 39,700 | $119,100 |
| Copper Termination | $150 per cable | 99 | $14,850 |
| Nodes | $2000 + $6250 per node | 6 | $49,500 |
| One Off Deployment Total | $911,950 | ||
| Node Servicing
per Annum |
$85 p/a | 6 | $510 |
| Node Power
per Annum |
$1404 p/a | 6 | $8,424 |
| Fibre Maintenance
per Annum |
$31 per cable | 99 | $3,069 |
| Copper Maintenance
per Annum |
$6.6 per cable | 99 | $653.40 |
| Operational Total Per Annum | $12,656.40 | ||
| Operational Total over 30 Years | $379,692 | ||
| Total estimated cost, including deployment and operational costs over 30 years. | $1,291,642 | ||
Assumptions Made :
- Per Pearce (2014), nodes will have an initial cost of $6250 to connect to the electrical grid. They will also cost $85 p/a to service, and $1404 p/a to power.
- Existing household copper lines (to the driveway) are connected to newly laid copper cables at no extra cost.
- Properties marked x can connect to the exchange directly, rather than connecting to the node, as it is closer.
- Per Reichert (2015), half the placed copper will require replacing within a 30 year time frame.
- The same fibre optic assumptions as the FTTP solution.
- No major environmental disruptions or un/intentional vandalism.
Evidently, despite FTTN’s initial guise of cost-effectiveness – when considering the long term, a more robust FTTP implementation would be a better investment. With significantly better data rates, reliability and scalability (only requiring an additional ~$50,000 in capital), FTTP is the cabled network I would suggest.
Starlink Implementation
All things considered, however, my personal recommendation would be to implement a satellite network. Assuming the farm is based in Australia, Starlink would not only offer a combination of scalability, reliability and performance but would provide faster deployment times and significantly lower overhead costs – especially in comparison to FTTx / fixed wireless alternatives.
According to Pekhterev (2021), Starlink can service up to 2080 users within a 20km radius before quality is impacted. This means individuals would not suffer from peak hour network congestion, nor would they struggle with bandwidth or data limitations seen in FTTN & FW. This also ensures the service is scalable. If other properties were to be built new residents would simply need to purchase a dish for instant internet connectivity.
Moreover, Starlink (and satellite internet technologies) are also being continually refined and improved. According to Ookla (Fomon, 2022), Starlink saw a 58% uptick in speed over 2021, with average download speeds moving from ~100Mbps to ~150Mbps – a number still increasing in 2024. Not only will these improvements help future-proof the network, but reduce potential additional expenditure that physical solutions might incur.
This being said, Starlink does have some notable limitations. As outlined earlier, atmospheric disruptions and environmental interference may affect internet connectivity and reliability. Having said this, the town’s existing 4G network lessens the issue, allowing the townsfolk access to emergency services and critical infrastructure if the Starlink network were to experience downtime. Starlink is also currently unable to compete with the gigabit speeds offered by FTTP (though, this may change given time). Finally, Starlink being a private corporation means its services are subject to the company’s terms. Were Starlink to remove, reposition, or decommission its satellites, the town could lose internet access with little to no backup plan.
Fig 1.6 – The implementation of a Starlink solution.
Pricing Estimate of Starlink Deployment & Operations
Based on the design outlined in Fig 1.6
| Item | Price | No. Required | Total |
| Satellite Transceiver | $599USD per unit | 99 | $59,301 |
| Shipping | $30USD per unit | 99 | $2,970 |
| One Off Deployment Total (Australian Dollars) | $95,387 | ||
| Internet Connectivity (Cost of the Household) | $139USD/mo | N/A | N/A |
| Powering Per Annum
(Cost of the Household) |
$120AUD | N/A | N/A |
| Transceiver Replacement Per Annum | $126USD | 99 | $12,474 |
| Operational Total Per Annum (AUD) | $19,108 | ||
| Operational Total over 30 Years (AUD) | $573,240 | ||
| Total estimated cost, including deployment and operational costs over 30 years. (AUD) | $668,627 | ||
Assumptions Made :
- Per the bandwidth requirements, each household will only require a ‘standard dish’ rather than a ‘high performance’ dish (Starlink, 2024a).
- Current pricing, as indicated on Starlink’s website, is accurate and reliable and will remain similarly priced in the future (Starlink, 2024b).
- Per Starlink’s Design Specifications (Starlink, 2024), each household will need their dish replaced every 5 years.
- These new dishes will provide upgraded, if not better, connectivity and speeds.
- Per Antenna Direct (2022), Starlink dishes will cost roughly $10/mo to power.
Overall, while Starlink does have its issues – its scalability, potential performance and cost-effectiveness make it the best of the wireless and cabled options. Saving upwards of $700,000 compared to FTTN/P approaches and providing excellent speeds and coverage – of those covered, this is the networking option I would recommend.
References
ACCC. (2023). Measuring Broadband Australia Program. Twenty-third Report, December 2023. Australian Government. https://www.accc.gov.au/system/files/measuring-broadband-australia-report-23.pdf
ACCC. (2012). NBN Co Network Design Rules provided to the ACCC in support of NBN Co’s Special Access Undertaking. Australian Government. https://www.accc.gov.au/system/files/NBN%20Co%20Network%20Design%20Rules.pdf
ACMA. (2019). NBN Wi-Fi modem study: ACMA modem performance testing and outcomes report. Australian Government. https://www.acma.gov.au/sites/default/files/2019-08/ACMA%20modem%20performance%20testing_outcomes%20report.docx
ACMA. (2021). Trends and developments in telecommunications 2022–23 Australian Government. https://www.acma.gov.au/communications-and-media-australia
Antenna Direct. (2022). Starlink – What You Need to Know. https://www.antennadirect.com.au/starlink/
Aussie Broadband. (2018). FTTP vs FTTN vs FTTC: NBN connections explained. https://www.aussiebroadband.com.au/blog/fttp-vs-fttn-vs-fttc-nbn-connections-explained/
CISCO. (2003) CCNA Self-Study: CCNA Basics (CCNAB) : Network Media Types. Cisco Press. https://www.ciscopress.com/articles/article.asp?p=31276&seqNum=2
CTC Technology and Energy. (2015). Cost Estimates for FTTP Network Construction, Prepared for City of Santa Cruz, California. Columbia Telecommunications Corporation. https://www.tellusventure.com/downloads/santacruz/ctc_santa_cruz_ftth_estimate_may2015.pdf
Department of Communications. (2013). Broadband Availability and Quality Report: December 2013. Australian Government. https://www.infrastructure.gov.au/sites/default/files/Broadband%20Availability%20and%20Quality%20Report.pdf
Fenwick, S. (2023). Mobile Network Experience Report. Opensignal. https://www.opensignal.com/reports/2023/10/australia/mobile-network-experience
Fomon, J. (2022). Here’s How Fast Starlink Has Gotten Over the Past Year. Ookla. https://www.ookla.com/articles/starlink-hughesnet-viasat-performance-q1-2022
Islam, R. et al. (2008). Wireless broadband technologies for regional and rural Australia: A last-mile perspective. Telecommunications Journal of Australia, 58(2-3), Monash University ePress.
ITU. (2009). Recommendation ITU-T G.992.3 – Asymmetric Digital Subscriber Line Transceivers 2 (ADSL2). International Telecommunications Union.
Jayant, N. (2005). Broadband Last Mile: Access Technologies for Multimedia Communications. Milton: Taylor & Francis Group.
NBN Co. (2024a). NBN Wholesale Market Indicators Report: December quarter 2023 report. ACCC. https://www.accc.gov.au/by-industry/telecommunications-and-internet/national-broadband-network-nbn-access-regulation/nbn-wholesale-market-indicators-report/december-quarter-2023-report
NBN Co. (2024b). Nbn® Fixed Wireless Explained. NBN. https://www.nbnco.com.au/learn/network-technology/fixed-wireless-explained
Pearce, R. (2018). Inside NBN’s fibre to the curb rollout. Computerworld. https://www.computerworld.com/article/3474725/inside-nbn-s-fibre-to-the-curb-rollout.html
Pearce, R. (2014). NBN FTTN’s power bill: $89m a year?. Computerworld. https://www.computerworld.com.au/article/549704/nbn_fttn_power_bill_89m_year_/
Pekhterev, S. (2021). The Bandwidth Of The StarLink Constellation…and the assessment of its potential subscriber base in the USA. SatMagazine. http://www.satmagazine.com/story.php?number=1026762698
Simmons, A. (2024). Cell Tower Range: How Far Do They Reach? Dgtl Infra. https://dgtlinfra.com/cell-tower-range-how-far-reach/
Starlink. (2024a). Starlink Service Specifications. SpaceX. https://www.starlink.com/legal/documents/DOC-1400-28829-70
Starlink. (2024b). Residential Plans. SpaceX .
https://www.starlink.com/residential
Tunmann, E. (1995). Hybrid Fiber-Optic Coaxial Networks: How to Design, Build, and Implement an Enterprise-Wide Broadband HFC Network. (1st ed.)
Reichert, C. (2015). NBN: ‘supercharged’ copper needs no replacing for FttN. ZDNet. https://www.zdnet.com/article/nbn-supercharged-copper-needs-no-replacing-for-fttn/
