Solved Am I right that 'Internet speed' is actually 'Bandwidth'?

wiganken said:

The thread on 'Show off your internet speed' intrigues me. Broadband speed test results can be very high but is this achievable in actual daily usage? I believe that when it comes to the actual download speed achieved this depends on the upload speed of the site that you are downloading from?
Microsoft's Windows update, for example, has only so much bandwidth which has to be shared out amongst all people downloading at that time so it probably throttles the upload speed so the end user will get nowhere near their bandwidth capability.
Am I understanding this correctly?

Click to expand...

Yes. You test your internet speed with the Ookla speedtest app and their website with, if available, your ISP's servers to check that you have the correct speed for your subscription (I have exactly the speed of my subscription).
However, in actual use, speeds can vary depending on the servers involved.

wiganken said:

The thread on 'Show off your internet speed' intrigues me. Broadband speed test results can be very high but is this achievable in actual daily usage? I believe that when it comes to the actual download speed achieved this depends on the upload speed of the site that you are downloading from?
Microsoft's Windows update, for example, has only so much bandwidth which has to be shared out amongst all people downloading at that time so it probably throttles the upload speed so the end user will get nowhere near their bandwidth capability.
Am I understanding this correctly?

Click to expand...

please check in settings
Windows update > Advanced options

in this window
delivery optimisation

i have download set to
Absolute bandwidth

and upload options
set to Disabled (off)

you check your settings
best of luck Steve ..

XxXxX said:

i have download set to
Absolute bandwidth

Click to expand...

Okay, thanks. I just did that so I'll see if anything changes.

Just remember, Internet speed tests like Ookla represent a theoretical maximum for blasting network data from one end to another. It doesn't take into account real-world overhead like file access times.

For example, you want to download a large MP4 or a Windows ISO. The server must load a giant file partially into memory from its own disk, begin a network download session, and your PC has to accept that data and write the file locally. Now your real-world "speed" has dropped.

What Ookla is doing is sending a raw stream of bytes, and your test client is accepting that data and immediately throwing it away. There's no processing done on the sent data (which is dummy data any way). The speed test is more of a bragging rights thing.

To your other question, large Internet services will partner with other companies to create Content Delivery Networks. A CDN is a pool of servers scattered across the Internet at strategic locations. The idea is instead of everyone ending up on the same server, and bottlenecking traffic, your download request is silently redirected to a parallel "local server". Now "local" may still be far away from your location, but hopefully it's better than the worst location.

Using CDN's, companies like Amazon and Netflix can stream movies more efficiently and each location doesn't have to be overbuilt in network capacity. When a CDN fails, you get random connection errors or slow downloads, and it's up to the CDN provider to go fix themselves.

This is how Windows Update can scale up whenever Patch Tuesday or a new Windows release comes out.

New network speed feature in the Network icon is used instead of the "classic" www.speedtest.net by Ookla. I am not sure, but I think there is a little difference between the two. I think this is because they use different gateways to connect and measure speed or because they use a different method.

For example, Windows 11 network speed measured my download at 243.35 Mbps and my upload at 154.09 Mbps while a few seconds later Ookla measured my download at 244.58 Mbps and my upload at 147.80 Mbps. No big deal, but which one you consider more reliable?

I block data spam and still eat real spam. aka Special Ham.

Perfect fried like this, but also perfect mixed with scrambled eggs in an omelet with or without French fries inside. In Greece we even add some feta cheese (white salty goat cheese) to make it even tastier...

kelper said:

I agree with OP. I reckon that low upload speed can affect download speed. Perhaps this is due to error checking by the source? It anyone is able to cap only their upload speed, this would be an interesting experiment.

Click to expand...

Yes. As part of transferring data in one direction, there is protocol data transferred in both directions between the source and destination hosts. This usually does not affect the transfer speed unless there is a lot of congestion. This might show up in asymmetrical service (like over copper cable where upload speed can be much less than download).

There are two communication protocols. One is TCP (Transmission Control Protocol) with handshake between transfers and error checking. It takes more time due to the handshake and the resending of lost packages, typically used in a download. The other is UDP without handshake, just packet sending and without error checking. When some packages are lost they are not resent. This is typically used in video streaming where lost packets result in artifacts or frozen picture. In extreme cases (such as very bad reception of DTV over the air) even the sound has glitches. Of course it is faster because there is constant package sending without handshakes or error checking and is used in situations that package loss is not critical, such as in voice communication or video streaming. So this is another factor affecting internet speed, the communication protocol, TCP or UDP with the second being the faster.