Full Duplex Gigabit speeds over HFC

Some interesting news from CableLabs – a US-based R&D consortium that focusses on cable broadband: they have achieved full duplex speeds over a hybrid fibre-coaxial (HFC) cable. In the lab they have achieved a symmetrical connection of 1Gbps, using the DOCSIS 3.1 standard.

Why is this exciting for us? Guess that needs some background.

As you no doubt are aware, the current iteration of the NBN – which once stood for National Broadband Network, but now doesn’t – uses a mixture of different technologies to deliver broadband to the Australian population. This is collectively termed an MTM approach, meaning multi-technology mix.

Some of these technologies are very good, such as Fibre to the Premises (FttP). Some of them, such as Fibre to the Node (FttN), are demonstrably inferior but cheaper in the short term (though not in the long term).

Some, such as HFC, are legacy technology that the NBN company bought from Telstra and Optus, an investment which has so far looked questionable. Eventually, about 34% of premises will be serviced by HFC, with NBN services beginning to come online in the next few months. The main advantage is that it is already installed.

Unfortunately, much of the cable looks to be unfit for purpose (especially the Optus stuff). The problems are that it is too-slow, in poor repair, and over-subscribed.

The breakthrough from CableLabs addresses the first of these issues (too slow). They have achieved gigabit speeds, using Full Duplexing. Furthermore, they anticipated eventual speeds of up to 10Gbps (bear in mind that this is a theoretical speed, and a fraction of fibre’s theoretical terabit top speed). More background!

In order to send information both ways over the same connection (i.e. uploading and downloading at the same time, which most modern internet applications require), the usual technique is to use either Frequency Division Duplexing (FDD) or Time Division Duplexing (TDD).

With FDD, two different parts of the bandwidth spectrum are reserved, one for download and the other for upload. With TDD, the same bandwidth can be used, but the upstream and downstream packets of information are staggered, such that they take turns travelling along the connection (with the downloads generally getting more turns). TDD is generally preferred, since spectrum is a finite and valuable commodity.

With Full Duplex however, simultaneous uploads and downloads can occur on the same bandwidth, giving you the footprint of TDD, but with much faster speeds.

The long and short of it is that, assuming the other outstanding issues with the HFC networks can be addressed, they could actually provide the kind of speeds demanded by the internet’s next generation. Eventually it’ll all likely have to be replaced with fibre, anyway – at considerable cost – but at least that day might be delayed.

You can read more about the DOCSIS 3.1 breakthrough at the CableLabs blog. Just be warned, you’ll have to wade through self-congratulatory guff like this: “A Full Duplex DOCSIS network is a prime example of the CableLabs 2.0 vision using its rich innovation funnel of technologies to transform the industry.” 

Golly