As the level of “gigabit capable” UK broadband ISP network coverage continues to rapidly expand then more and more people are now finding out what those working within telecoms and computer networking have known for a long time –being able to get or harness the full 1000Mbps (1Gbps) isn’t always possible.. yet.
One of the most fundamentally challenging issues that consumers face, when trying to attain a good broadband speed today, is centred on the limitations of the line itself. Twisted pair copper lines are notorious for issues related to signal degradation, both over distance and caused by other factors, such as weather damage or poor home wiring etc. The result is often woefully poor speeds.NOTE: At the time of writing around 34% of the UK are within reach of a 1Gbps capable network (here).
Thankfully the new generation of Gigabit-capable “full fibre” (FTTP) services, as well as DOCSIS 3.1 based Hybrid Fibre Coax (HFC) networks, resolve most of this and bring with them the option of top tier consumer packages, which can offer download speeds of 1Gbps (Gigabit per second) or even better.
Naturally many people will be attracted to spend a bit extra on the fastest 1Gbps packages and can thus end up being surprised when the top speeds prove harder to attain than expected. None of this is particularly new territory for those of us educated in the ways of computer networking, but as more people adopt such packages then it’s worth highlighting why achieving what is advertised can still be a challenge, even on full fibre.
For the purpose of this article we’ll explore a situation where somebody orders a 1Gbps package but, after conducting a broadband speedtest or downloading a file, finds that they’re “only” able to achieve a few hundred Megabits or less. On the one hand most people would be more than happy with such a speed, but it still begs the question, why can’t you harness all of the speed? We’ll try to explore some of the possibilities below.NOTE: Most 1Gbps packages will be advertised alongside an average speed (as measured at peak time) of 900Mbps+, as per the Advertising (ASA) watchdog’s rules.
Most ISPs will do their best to build networks that broadly deliver on what they promise and, as per the ASA rule mentioned above, if they can’t then it should show in the average (median) speeds that they’re required to promote. Nevertheless an average is a figure that stems from a balance of both good and poor results, with some people always being on the losing side of that equation.
Consumer broadband networks need to be affordable and as such they share their data capacity between many users in a given area. For example, in some locations an ISP might be using a 10Gbps capacity link to serve tens of customers (i.e. 10, 40, 50 etc.) or more .
For the most part a single user in that area probably wouldn’t even notice this because of how few would be using their connection at the same time and, even if they were, then they’ll probably only actually NEED a much smaller slice of that speed at any given moment (e.g. somebody with a 1Gbps line but who is only using 24Mbps by viewing a single 4K video stream).
Lest we also forget that others in the area might have chosen a slower package (e.g. 100Mbps, 500Mbps etc.) and so would have a lower level of peak demand. Providers generally create models of usage to help cost and balance how much capacity they’ll need to buy, while also allowing some overhead for peaks.
Nevertheless some ISPs may be more budget conscious, which can result in them oversubscribing their network in a particular area (i.e. too many users for the given capacity supply). The impact of this is more likely to show up at peak (evening) times, when more people are online and making active use of their connection, as opposed to off-peak (daytime) periods with lighter levels of network load.
Complaints from end-users are usually enough to resolve this, although it can take time. Equally you might be able to resolve it by changing ISP. One other solution is to consider buying a dedicated business line of a similar speed (e.g. leased line), although those are often significantly more expensive – reflecting the naturally higher cost of offering a truly uncontended service.
At this point it goes without saying that wireless networks, even those that use the very latest 802.11ax (WiFi v6) technology or access to the new 6GHz band, are far from perfect. Indeed a typically low powered WiFi signal is extremely weak and rapidly starts to degrade over a fairly short distance. On top of that any objects or structures in their way (varies – depending upon the materials used) can further hamper or even stop the signal.
Due to this you tend to get the best WiFi performance when within just a handful of feet of your router – inside the same room – and this is also dependent upon the device you’re using (Smartphone, Laptop, Tablet etc.) having a similar or better capability than that of your router (e.g. the WiFi chips and antenna inside many Smartphones may be inferior to those inside high-end laptops etc.).
Long story short, even if the WiFi network claims it can deliver peak speeds of 1Gbps, or faster, then the reality is that you’ll often receive significantly less. As such it’s very UNWISE to test a 1Gbps broadband line over a WiFi link and you’ll often be better off conducting such testing via a wired connection to one of the Ethernet based Local Area Network (LAN) ports on the back of your router.
Admittedly there are also other ways to extend and improve the performance of your WiFi service (see our Wi-Fi Tips Article), such as by using a Mesh (repeater) system to effectively replace your router’s own WiFi, but these still run into some of the same problems and speeds will vary.
Assuming you’ve connected to the LAN port on your router in order to test a 1Gbps broadband line then you’re immediately likely to run into a couple of problems. Firstly, most broadband routers today only ship with 1Gbps Gigabit Ethernet capable ports, which should be fine. However, once you account for various overheads, the maximum real-world data transfer rate ends up being less than 1000Mbps.
Depending on how you calculate this and what overheads are involved (we’d need a separate article to explain all this), the top usable speed of a single 1Gbps Ethernet LAN port is likely to be around 987Mbps (older setups without jumbo frames support might expect even lower – c.930Mbps). In other words, you’ve already lost some Megabits and that’s before you’ve even got anything else running.
We should clarify this doesn’t mean your router is completely capped to a sub-1Gbps speed. The example above only reflects a single user and wired connection, although all routers are designed to serve multiple LAN ports and users over WiFi at the same time.
In an ideal world router manufacturers should really be sticking even faster 2.5Gbps or better LAN ports on their devices, although there’s always been a historic problem with recognising the demand for such things and implementing that (not long ago almost every router was still coming out with 100Mbps ports on the back, even when WiFi was going much faster).
One thing that many people either forget to consider or simply aren’t able to check (it’s a hard thing to pin down) is the measure of load on your home network. For example, a lot of software applications (e.g. Microsoft Windows 10) and devices will be constantly checking for and running automatic updates in the background, some of which could be several Gigabits in size. Likewise there may be other people using services on the network.
Flick over to the next page (2) in order to continue reading..