Internet Speed Overkill

We use the internet more and more. Of course we want this experience to be as quick and convenient as possible.  But much of any “slowness” we might encounter comes from elsewhere on the internet, and increasing our access speed won’t make the internet any faster for most of what we do.

We’ve just upgraded our internet connection speed to 500 Mbit/second for both uploading and downloading.

We have vivid memories of the various connection speeds we’ve had in the past, stretching all the way back to our first ever dedicated connection (56kb frame relay – a stunning 9,000 times slower, but in the mid/late 1990s, seemingly fast enough for our office of 15+ people and webserver all to share) and the dialup modem speeds we had prior to then – all the way back, not only to the classic Hayes Smartmodem with its 1200 baud (ie bps) speed and at-the-time revolutionary AT command set, but further back to 300/75 bit acoustic coupled modems.  (Our related article, “The Exciting Potential of (and Need For) Universal Fast Internet Service, Everywhere“, includes a ‘blast from the past’ link to the sound of a modem dialing/handshaking/connecting – very familiar to older readers, but not so much to younger ones.)

No doubt there are a few readers who can remember even slower modem speeds.

Luxuriating in our almost impossible-to-comprehend 500 Mbit speed, we found ourselves strangely disappointed.  Which got us to wondering – is there such a thing as “too fast”?  The question is valid, because for most of us, the faster the connection speed, the higher the monthly fee.

For a long time, there’s been a clear benefit to getting a faster internet connection.  Web pages would load perceptibly more quickly.  Audio files would play more smoothly.  Voice over internet calling became possible.  Video would stream without pausing to rebuffer, and at higher quality resolutions.  Video calling and conferencing became possible.  Files would download in fewer seconds.  And so, for some of us, we’d eagerly upgrade to a faster speed on a regular basis.

But, to mix metaphors, on the “internet superhighway”, is there such a thing as a “super car” – a vehicle with stunning horsepower and top speed, but unable to use either due to traffic congestion and speed limits?  Surprisingly, the answer to that seems to be “Yes, there is”.

Sure, just like the super car will be able to move away from the lights more quickly than the car in the lane alongside it, there are some situations where faster speed might be beneficial, but for most of us, most of the time, there will be no perceptible difference between 500 Mbps and 50 Mbps.  Let’s look at why that is, and also consider how to get the most out of your internet speed, and what applications might require high/higher speeds.

What Internet Speed Really Means and Measures

The whole concept of internet “speed” is poorly defined and much misunderstood.  Think again of vehicular traffic.  You have two vehicles traveling down the freeway.  They are both going at 70 mph.  One is a regular car, the other is a bus.  The car can carry four people and their luggage.  The other is a bus which can carry 50 people, their luggage, and some extra freight, too.

In internet terms, the “megabits per second” (ie Mb/sec or Mbps) speed is not a measure of the “miles per hour” speed of the internet connection.  That is because, essentially, all internet connections travel at the same speed.  This speed is a bit less than the speed of light (depending on if it is a radio signal, going through a wire, or through glass fiber), and is in the order of 120,000 – 150,000 miles per second, slowed down only when the connection travels through nodes and routers and connections on its path from your computer to the other computer it is connected to.  Data could flow down that connection, other than for switching delays, so fast that it could go around the world in 0.15 seconds.

The Mbps doesn’t measure the internet equivalent of vehicular miles per hour.  Instead, it measures the amount of “carrying capacity” of the vehicle – the number of passengers and freight that can be transported.

Of course, if you have need to move 20 people, it will be quicker to have them all travel in one large bus than to have a passenger car make two or three roundtrips to shuttle them all to their destination.  And so, on the internet, when you’re moving any appreciable quantity of data, it does make a difference in the time it takes if the connection “speed” is greater.

A better term for internet “speed” is internet bandwidth – how much data it can transport per second.  All other things being equal (warning – they are not!), a 500 Mbps bandwidth connection will move ten times as much data as a 50 Mbps connection.

A Common Error in Measuring Internet Speeds

Note that we’ve been referring to Mb/sec – to spell it out, megabits.  This is totally different to MB/sec – something you’ll often see mentioned, for example, in hard disk or USB transfer speeds.  There is a world of difference between a lower case and upper case “b”/”B”.  The “B” signifies a Byte, which in simple terms is 8 bits.  So – and again in simple terms – as a rule of thumb, 1 MB = 8 Mb.

Sadly, some people, including some internet companies that should know better, don’t understand this eight-fold difference in measurement, and because they are so used to seeing Bytes in most other computer contexts, they “correct” a lower case “b” or automatically use an upper case “B”, without realizing their error.

The Impact of Speed on Web Page Loading

You might think that doubling your internet speed will halve the time it takes to download a webpage.  This is not so.

The average web page these days seems to contain about 3MB of data, although there is a wide variation in this figure.  The major component of most webpages are the images on it, then programming and script files, and only a very small part being the actual text/words you read.  If this is of interest, here are – one  two – a couple of interesting pages about this.

Some of that data may be already cached on your computer.  But to consider the simple issue of how long it takes to download 3MB, if you have what today would be considered a slow internet connection (5 Mbps) it would take about 4.8 seconds to download that webpage.  If your connection was 50 Mbps, it would be just under one half a second (0.48 seconds), and therefore, at 500 Mbps, it becomes 0.048 seconds.

This points to a couple of issues.  The first is that by today’s standards, 5 seconds is unacceptably slow for a web page to load, and so you definitely get an improved browsing experience by pushing your speed up from 5 Mbps.

The second is that few of us would notice or care about a difference between half a second and something ten times faster.  We’re in the realm of vanishing returns.

But there’s another factor that becomes bigger when evaluating web page load times at high speeds.  The way a web page loads is not a simple case of your computer sending a request to the website saying “please send me this page” and the website saying back to you “okay here it is” – it is not a single simple request and response.

The website computer will send you multiple files – one for each image, one (or more) for the text, and a series of additional files with programming, scripts, and other clever things.

We just checked our own blog site and we see that viewing its home page involved 109 requests and responses, with information coming in from over a dozen different webservers.  It took about 2.5 seconds for the page to be completely downloaded.  But – and this is the thing – almost all of the 2.5 seconds was spent waiting for the various websites to respond, not time actually taken for the downloading.  The downloading represented the smallest part of the 2.5 seconds, and no matter how fast the internet connection was, the remote computers won’t respond any faster.  In our case, our home page is about 2.7 MB, and so at 500 Mbps you might think it would only take 1/25th of a second, but in actual reality, it took 60 times longer.  The internet connection speed is only a very small part of the total time.

Here’s a screen shot that shows some but not all of these requests and how the total time to create the page on a computer is spent.  It is only the little grey patches at the right of each bar that is time spent downloading.  For example, over 0.4 seconds is spent at the start just waiting for the web server to work out how to respond and what to send back in response to the page request.

There is another delay not shown in these tests – that is the time it takes for your computer to decode the information it receives and create a nicely formatted webpage on your screen.  That can add some further small amount of additional time.

So the point here is that, in round figures, it is not uncommon to find a web page that takes 0.5 seconds or less to download, but which has an additional 2 seconds (or more) of delay due to the time it takes the webserver(s) to respond to the page request and send your computer the data.  Maybe a faster connection will reduce the 0.5 second part of the total process, but nothing you do can vary the other 2 seconds.

So, does it really then matter if the total time to view a page is 2.5 seconds or 2.1 seconds?  The “feel” of interactivity will be the same either way.

What About Internet/Online Games

If you have teenagers in your house, they might spend a lot of time playing online games.  These don’t use as much bandwidth as you might think – usually it seems in the order of 3 Mbps is plenty.

But what games do require is fast latency, and we’ll talk about that next.

A Little Quoted Measure of True Internet Speed

Let’s go back to the analogy of cars and buses or trucks on a freeway.  We know that the car and bus are driving parallel to each other on the freeway, their speed is the same.  What varies though is the amount of freight and passengers they each can carry – their bandwidth, which in internet terms is expressed by Mbps.

There’s another important factor too.  What say the car has to drive 200 miles to get to its destination, and the truck only has to drive 100 miles.  It will finish its job sooner, even though it is driving at the same speed.

Or, what say, the car and truck, while driving at the same maximum speed when traffic allows, are taking two different routes.  One route is wide open on a freeway.  The other route is over surface streets, with congestion, stop lights, road construction and traffic accidents.

This applies to the internet as well.  Maybe your internet connection is at the far end of the equivalent of a winding hilly one lane dirt road.  Or maybe it is right next to a freeway on-ramp.  One measure of this is how quickly your connection can get to the main internet “super-highway”, which is something measured by a “ping” test and is expressed as a measure of “latency” in milliseconds.

This is the time it takes for your computer to reach out to another computer on the internet (ie to “ping” it) and have it simply and almost instantly respond back with an acknowledgement.

Usually you should hope for the ping data, as shown for example on the speedtest.net site and service, to be in the order of 5 – 10 msecs maximum.  In my case, I’m usually experiencing 3 – 4 msec of latency.

This same latency measure can be applied not just for the time it takes to reach “an internet onramp” but also for the time it takes to get all the way to a webserver you wish to access.  This depends on the distance the message has to travel, and the number of devices it needs to get switched through en route.

For example, a quick test reveals it is taking a consistent 68 msec for my computer to ping the webserver somewhere in the mid-west of the US on which The Travel Insider Blog is hosted.

Looking a bit further it is possible to see all the 13 different segments of the path it travels via a Traceroute (Tracert in Windows) command

If you’d like to know more about traceroute information, this is probably the definitive guide.

One more traceroute image which is interesting to look at.  I tested a server in New Zealand.  Note how the ping times jump on lines 9 (when the signal has to travel from Portland over to Salt Lake City), and again on line 13 when it heads over to New Zealand.  Even nearly the speed of light starts to result in measurable delays when traveling very long distances.

So, to summarize this, there are simple delays in your internet connection to an outside website that will not change if you speed up your bandwidth, because the delays are either in other people’s equipment or unavoidable due to the limit of how fast the data travels through a line.

How Much Bandwidth Can You Use?

This question was formerly irrelevant.  When internet speeds were way below 1 Mbps, and Wi-Fi connections were that fast or faster, and ethernet cabled networks were quoting 10 Mbps, the internet connection was the slowest part of anyone’s network.

But now, the opposite may be true.

There are a number of different types of Wi-Fi.  There are two different frequencies (2.4GHz and 5GHz) and different Wi-Fi standards such as 802.11a, b, g, n, and ac.  There are other standards less commonly found in consumer grade electronics, too.

The most common 802.11b type of Wi-Fi has a maximum speed of 11 Mbps and often will be lower.

802.11a and 802.11g go up to 54 Mbps if it has a wide 20 MHz channel, and also will usually be much slower.

The newer 802.11n is faster again and can go up to 600 Mbps.  The 802.11ac can be faster still, going all the way up to 3.467 Gb/sec, but is rarely found over 800 Mbps.

Although I’ve a nearly new Wi-Fi router that claims maximum speeds of up to 1750 Mbps, I’ve never got faster than 152 Mbps down and 206 Mbps up from devices connecting to it when they are close to the router.  And when I move to another room, speeds plunge down to an unpredictable 25 Mbps or less down (I did two tests in a row just now, one with half the connection speed of the test immediately prior) and 20 Mbps or less up.

To get the very best Wi-Fi speeds, you need both a state of the art router/Wi-Fi unit, and also state of the art products that you are trying to connect to the router with.  And those large multiple antennas on the router – they’re not just for show.  They actually do help get faster data transmission; but without similar antennas on your user devices, you’ll always be struggling to get close to the advertised maximum speeds.

So, for most of us, with most devices, your Wi-Fi connection speeds are these days unlikely to be faster than your internet connection.  Your limitation is not your internet speed, it is your Wi-Fi connection.

What about good old fashioned ethernet style wired internet?  Well, that used to be limited to 10 Mbit, then speeds increased to 100 Mbits, and more recently, there has been a growth of 1000 Mbit/sec ethernet.  That sounds good in theory, but there are a couple of considerations.  The first is to be sure that all the devices you are connecting will be capable of that speed, and the second is to be sure that your wiring and device is also capable of that speed.

You need “Cat 5e” or a higher number (ie 6 or 7) type cabling to reliably experience 1000 Mbit speeds; earlier Cat 5 cable might work, especially for shorter distances, but might not.  Additionally, if your computer’s ethernet port is connected to a PCI bus, it might experience congestion inside the computer and not be able to accept the full speed data flow.

There’s one more consideration with wired ethernet.  While, sure, it is quoted as 10/100/1000 Mbps, in reality, the actual real-world speeds are lower – more like a maximum of 800 Mbps and often somewhat less.

So while some regions and providers will even offer 1000 Mbps internet connections, for a regular domestic or small business environment, it is hard to see how you’d take advantage of more than some small share of this.

Upload and Download Speeds

Most internet connections have two different speeds.  One is the speed at which data is sent to you – this is called the download speed.  The other is the speed at which you can send data to the internet – the upload speed.

Most of the time, we are predominantly downloading – receiving web pages, watching video, downloading files.  It is very rare we upload anything of any size, and often that uploading is done “in the background” (like sending an email) so we don’t even realize how long the upload is taking.

It used to be common to provide what was termed “asymmetrical” internet access, with different upload and download speeds.  These days, sometimes that remains the case, sometimes you have a symmetrical connection with the same speed both ways.  For most of us, our major focus should be on the download not upload speed.

Real World Connection Speed

Now for a difficult point.  If you read the fine print on an internet contract, you’ll probably see that the provider is only promising speeds “up to” the quoted speed.  You can be sure your speed will seldom or never exceed the promised speed, but what you can’t be sure of is whether it might often be slower than the (not) promised speed.

Some types of internet connection are more likely to suffer from congestion, because you are sharing more of the “last mile” connection than you are with other types of connection.  Cable based connections used to be notorious for this, with a slowdown every evening, but these days they are better than they formerly were.

However, you should occasionally check your connection speed – use the speedtest.net site to do this.  If your speed is more than perhaps 10% below your promised speed, and particularly if it remains low for more than an hour or so, call to your provider and report the problem.

In our case, we’re not quite getting half the promised speed.  Here’s a summary of tests we’ve done over the past five days.

But there’s so little value in getting our full rated 500 Mbps speed that we can’t even motivate ourselves to phone Frontier, wait up to an hour on hold, and then go through a lengthy and pointless set of meaningless “troubleshooting” steps – all of which we’ve already done – and then be given some platitude of an excuse with likely no further action taken.  We’re more than happy with “only” half the promised speed.

Real World Internet Speeds

Think again of our road analogy.  So you’re on the freeway in Montana and the sign says 80 mph and you’ve a fast powerful car that can easily exceed that by a comfortable margin.  You pass by the sign at the full posted speed.  But then you catch up to a bunch of traffic that is crawling along at 50 mph and blocking all three lanes.  It doesn’t matter what the speed limit is, and it doesn’t matter how fast or powerful your car is.  You’re stuck.

The average download speed on the internet has been suggested to possibly be in the order of 72 Mbps.  No matter how fast your connection is, if the computer somewhere else on the internet can’t send you the information at the speed you can accept it, or if somewhere on the path from that computer to your computer there is some congestion, you’re stuck at the slower speed.

So from that perspective, it is very rare that you’ll ever get to use the full capacity of your fast internet connection.

How Much Bandwidth (“Speed”) Do You Need?

The answer to this question depends of course on what you use your internet connection for and how many people you share it with.

For most of us, the really big need for bandwidth comes when we’re streaming video from Netflix, or Amazon Prime Video, or whoever/whatever.  And if there is a second person in your residence who is also streaming a different video at the same time, using the same internet connection, then yes, the bandwidth required does indeed double.

As you’ll see on their page, Netflix advises that we will need 5Mb/sec for each video stream we are watching at the same time.  If you’re streaming in Ultra HD – which is their non-intuitive way of saying 4K and HDR, you’ll need 25 Mbits/sec.

Note that some other services suggest and possibly require different amounts of bandwidth to stream video (some services suggest only 15 Mbps for 4K video but that is probably because they are not also adding the extra HDR data to the signal).

You probably have computers, phones and other devices more or less permanently connected to the internet, and even when you are doing nothing on them, they are still carrying on various “conversations” and connections through the internet, and that is likely consuming on average 1 Mb/sec – sometimes a bit more, often a bit less.

If you are surfing the internet, each web page you open up will take measurable bandwidth, but just for a fraction of a second, and if the bandwidth isn’t available, it will simply load more slowly.  This can affect your video streaming, if there isn’t plenty of remaining bandwidth in your connection over and above the amount being used for the video streaming.

So, our suggestion is that you take the maximum number of concurrent (non 4K) video streams and allow 5 Mbps for each, and then add another 5 Mbps on top of that for other people and devices using the connection at the same time, and there’s one measure of how much bandwidth you “need”.

So, in truth, there’s no point in having 500 Mbit/sec internet.  You’re unlikely to see any difference in actual performance at any speeds over 50 Mbit/sec.  The only reason for going over 50 Mbit/sec would be if you have multiple people all sharing the same internet connection.

An Easy Way to Speed Up Web Browsing

If you’re using a Chrome web browser (and most people do) you can switch on an option to pre-load web pages that Chrome thinks you might be about to load next.

Go to Chrome’s settings via the three dots in the top right corner, select Advanced, then select Privacy and security.  On that page, turn on the option that says “Preload pages for faster browsing and searching”.  A similar feature is available in Firefox.

That means that if Google or Firefox guess right, when you click the link to go to the next page, some or all of it has already been downloaded and is ready on your machine, waiting for you.  This feature will speed up much of your web browsing more than any internet speed increase.

Summary

In our case, we upgraded from 35/15 Mbps to 50/50 and noticed very little difference.  We then upgraded from 50/50 to notionally 500/500 (actually more like 225/250) and noticed no difference whatsoever – if anything, it seems that sometimes now our computer struggles to keep up with the data flow into it.

Your download speed is much more important than your upload speed.

A single user is unlikely to experience any benefit in his internet use experience by increasing his connection speed up from 50 Mbits/sec.  By all means upgrade from your current speed to 50 Mbps if the cost to do so is modest, but be wary of paying significant amounts more for faster speeds.

If there are other people sharing your internet connection, and particularly if you wish to do 4K video streaming, then there might be value in going over 50 Mbps.

So, you might ask, why did we get a 500 Mbps feed?  Because – after needing to get the WA Attorney General to help encourage Frontier to make good on their promise extended by them in good faith and absolutely not in error, it ended up costing us a mere $40/month (this is an advertised package promotion they offer in some areas).  That’s a rate less than we were paying for the 35/15 service just a couple of months earlier.

2 thoughts on “Internet Speed Overkill”

  1. In reality – whether you have a 50, 500mb or Gig-E internet connection – these are only testing to for speed testing. The real world speed generally maxes out at closer to 10mb. Speed test sites are for beginner admins, and looking glasses are for network professionals. This is because the ISP’s connection a website provider to most handoffs is somewhere between 1 and 10gb, and at a certain point they may become saturated as they are all oversubscribed. It’s the nature of the business.

    Like at the ISP if you want a faster, more reliable connection purchase your internet from multiple providers. This way when one goes down or becomes saturated it does not impact you.

    Video is a modern format, and can buffer when speed becomes choppy. Voice cannot, and with VOIP (which requires solid but little connectivity) – if it cannot reach the destination quickly (<300ms) the call drops.

    1. Hi, Dan

      That’s a more pessimistic view than my own. Do you have anything you can refer to as support for your suggestion that in the real world, speeds generally max out at closer to 10Mb? Not saying you’re wrong, just inviting you to support it with something so one can see what assumptions and measurements were used to create the claim.

Leave a Reply

Scroll to Top
Scroll to Top

Free Weekly Emailed Newsletter

Usually weekly, since 2001, we publish a roundup of travel and travel related technology developments, and often a feature article too.

You’ll stay up to date with the latest and greatest (and cautioned about the worst) developments.  You’ll get information to help you choose and become a better informed traveler and consumer, how to best use new technologies, and at times, will learn of things that might entertain, amuse, annoy or even outrage you.

We’re very politically incorrect and love to point out the unrebutted hypocrisies and unfairnesses out there.

This is all entirely free (but you’re welcome to voluntarily contribute!), and should you wish to, easy to cancel.

We’re not about to spam you any which way and as you can see, we don’t ask for any information except your email address and how often you want to receive our newsletters.

Newsletter Signup - Welcome!

Thanks for choosing to receive our newsletters.  We hope you’ll enjoy them and become a long-term reader, and maybe on occasion, add comments and thoughts of your own to the newsletters and articles we publish.

We’ll send you a confirmation email some time in the next few days to confirm your email address, and when you reply to that, you’ll then be on the list.

All the very best for now, and welcome to the growing “Travel Insider family”.






David.