Is The NBNco Fast Yet?

March 2017.







Dear NBNco

Bring it on? Cut it out.

This article – much like the NBN roll out – is long overdue. But, it’s time that those of us in the technology sector spoke out. Since the Abbott–Turnbull government first imposed its NBN policy directives, the voices of experts in our industry have been ignored. They continue to be ignored.

Our concerns are wide ranging. But what bothers us most is that NBNco – the company tasked with rolling out Australia’s largest infrastructure project – is failing to do so. But it’s not over yet. We still believe that it’s possible to fix what’s broken. Swerving from the embarrassing road we’re barreling down, and aiming for the kind of NBN Australia deserves.


Quick history of the NBN

The NBN was born of the failure of the many political and private endeavours before it. Telstra was chiefly privatised to capture investment capital from the business community with the intent purpose of replacing the ageing copper network. This never came to pass. And now the NBN is in danger of repeating the failures from which it sprang. History is repeating itself.

While the original model (Fibre to the Premises, FTTP) offered a high-performance connection, the model we have settled for (FTTN, Fibre to the Node) relies on decades old copper infrastructure. It can be as unreliable as the infrastructure it purports to upgrade.

NBNco is continuing this trend, racing to the bottom with the Multi Technology Mix. As they deploy these different kinds of technologies we will end up with an NBN that is neither suitable nor cost effective. All while actively working to remove incentives for future upgrades to the network to keep up with demands of the citizens.

The Liberal NBN is broken, and the resulting network will be equal or worse than what it replaced. How can we be an innovation nation when we rank 60th in global Internet speeds?


Hidden costs behind the scenes

Telstra charges more than 2000% for network access compared to European counterparts.

Network Provider Cloudflare has stated "We pay about as much every month for bandwidth to serve all of Europe as we do to for Australia. Approximately 33x the number of people live in Europe". An NBN-owned network MUST be able to route around Telstra as a means of lowering costs – and, therefore, internet-usage bills for all Australians.


Purpose-built…for 1872

Most of the technologies available to the NBNco Multi Technology Mix were not originally designed for the internet. Copper networks were built for telegraphs and phone calls. Today, that technology is held together with plastic bags and Band-Aids, while the fibre-optic network’s core design has a single purpose: transmitting internet data as fast as possible.

Fibre may be initially more expensive because it’s a newer technology, but keeping the copper network on life support is plain irresponsible.


Why Focus on Fibre?

In 2017, broadband internet access shouldn’t be a luxurious toy, but a privilege all Australians enjoy. As more and more of our institutions come to rely on the NBN, the lack of equitable access will mean that those without will be severely disadvantaged.

The original FTTP model is non-discriminatory and utilitarian. It creates a society where an NBN-enabled home is the gold standard for speed and equality, with most homes being equal – including those in rural areas. Education, entertainment and future career prospects all depend on reliable internet access.

In that sense, the NBN is more than a piece of infrastructure, it's an economic enterprise – and should – be an instrument for social good. With a little more foresight, it could be both.

Where we have headed with the FTTN looks very different…


Speeds: FTTP vs FTTN

FTTP guarantees the best available speeds to all Australians, everywhere (at the time of writing, that would be up to 1000 Mbps). Not just the lucky ones. Whereas copper-based FTTN can be a roll of the dice. Unguaranteed speeds range from a ‘best case scenario’ of 65 Mbps to as low as 10 Mbps, depending on a number of factors (such as rain), with no ability to improve. Why do we accept an infrastructure project with such clear winners and losers? And a low benchmark for even the best connections.

You just have to look to the real-estate market to see how this is playing out. Increasingly, homes are advertised as “NBN enabled” to showcase the fact that they have ‘Fibre to the Premises’ (FTTP). The technology once promised to all is fast becoming a luxury item, available (for a price) to the lucky few.


Multi Technology Mix Backend Cost Blowout

An inescapable technical quagmire

The MTM model requires complex systems to pull these technologies together. Each piece of hardware increases complexity as multiple vendors and engineers struggle to glue incongruous systems together. Incompatible proprietary systems require costly, custom-designed software to communicate. The Turnbull model has resulted in skyrocketing costs with diminishing returns.


Complications

Every piece of technology has primary hardware, supporting hardware and operating software, as well as support contracts and training. Hardware quickly goes out of warranty or is discontinued; multiplying those hardware issues by using a mix of technology just makes no sense. Different brands and technologies are commonly unable to directly speak to each other, requiring system administrators to glue the disparate pieces together – at considerable cost.

In short, NBNco’s model of adapting and maintaining Frankenstein networks is incredibly expensive and complex. Choosing a single technology would reduce upfront and ongoing maintenance costs.


Meanwhile, Turnbull continues to raise concerns that “people don’t use higher speeds” while failing to mention this is because high bandwidth services are not yet built. This is a chicken and egg issue – build it and they will use it. The business community will feast on this increased bandwidth especially during the #ideasboom with entrepreneurs brimming with possibilities for this untapped bandwidth. (Streaming services like Netflix, for example, are a prime example of the kind of opportunistic entrepreneurialism Australia could see with increasing bandwidths.)

And – perhaps, as someone has suggested, if you don’t have the kind of connectivity you think you should, maybe you should ask if it’s your own fault. After all, that’s what Malcolm Turnbull thinks.

Written by nbnopenletter@protonmail.ch

Extra Reading: Technical Issues

Line Length

Using FTTN, the best of the speeds available are only maintained for the first kilometre. Considering that speed continues to drop with every additional metre, every home on the street is likely to be connected at a different speed. FTTN requires nodes to be located at a visible distance from the property for a chance of higher-than-average speeds.

Fibre maintains its maximum speed for the length of the entire run, which means that node placement is much more flexible. For example, if NBNco decided to limit fibre runs to five kilometres, this would require massively fewer nodes. In turn, that would mean a lower cost for installation, power and ongoing maintenance.

Weather

Copper-based broadband networks are affected by the weather. Rain causes connection problems, thanks to insufficient protection and non-waterproofed access panels. TV and radio broadcast towers also degrade copper cable. To fix this you need additional hardware, such as RF filters (another extra cost).

Wireless broadband

While offering homes internet access without digging up the streets, wireless also suffers from congestion issues. The more people connecting to each tower, the smaller the share of bandwidth each person gets. Even in the best conditions, wirelessly connected homes will never reach acceptable levels for video conferencing, online gaming or other similar services.

Latency

It’s an issue often raised by network engineers, but latency is often lost in the wider political discussion. Internet connections are primarily advertised by connection speed (25Mbps). Latency (put simply, the speed at which messages are sent) is not advertised. But the lower the latency, the more real-time and complex services become possible.

Online gaming, teleconferencing, and forex/share trading depends on the signal being sent, received and responded to as fast as possible. Fibre’s latency is unbeatable thanks to the speed of light, offering 1–20ms connections to local servers compared to 20–150ms connections with FTTN, and 200–1000ms over wireless.

Over the last few years, Amazon and Google have both published some remarkable statistics on latency. Both found that every 100ms of latency cost them 1% in sales or conversions. They produced examples showing that one second (1000ms) of extra delay resulted in around 9% reduction in sales and conversions. As we transition to an e-commerce focused digital economy and online always-on retail and services, speed is king – but latency is queen.

Fibre-only models resolve many issues telecommunication engineers face on a daily basis, yet the Turnbull government continues.
The points raised above are all well known, widely understood, and should come as no surprise to anybody inside the NBNco.

References

2 February 2008 - Fibre to the Node: At what price? (internode.on.net)
29 November 2009 - Optimising ADSL2 Service Performance (Simon Hackett)
6 April 2014 - The Relative Cost of Bandwidth Around the World (cloudflare.com)
11 February 2016 - Our ‘woeful FTTN NBN’: an open user letter (itwire.com)
6 March 2016 - "Internet Australia calls for NBN Re-Think" by Internet Society of Australia (internet.org.au)
6 April 2016 - Australia's broadband policy is a flimsy, cynical House of Cards (Mark Pesce via theregister.co.uk)
28 February 2017 - Fast, Faster, Fastest - The changing goalposts of High Speed Internet Access by Andrew Cox.