broadband speeds wireless/wired

Can somebody please explain to me as if to a braindead moron....

If you have wireless usb adapter, what/how does the speed of it work/matter?

I have one adapter that was reporting (on the little network connector icon) that it was connected at 150 Mbps. The one I've got in at the moment is connected at 72.0 Mbps. Should that make a difference if you do an online "speed test", or is the speed test testing what the 4G device is doing, and not what your computer is doing? (one done with the 150 came at around 15, but but the speed test with this adapter came at 18, how can that be????, shouldn't it be like 7???? Shouldn't the faster speed one be faster? )

I was looking at getting a wired connection but they tell me they don't have to give you any better than 3000 (something). That'd be worse than I've got, wouldn't it? It's just 3 Mbps when I've got 15-ish?

(Why do they have to mix up kbps Mbps M-something else, blah blah blah? Why can't they just stick with the same measure?)

thanks

 

Connectivity speed is always measured in terms of bits per second (bps) but that number can get unnecessarily large so we use the SI prefixes kilo (K, thousand), mega (M, million), giga (G, thousand million or billion), etc. Bits are too small a unit to be useful to the common person though, so the average person will prefer to use bytes per second, particularly megabytes per second as file sizes are typically around there. Nowadays, it's pretty much accepted that there are 8 bits for every byte but that wasn't always the case.

Now, the speed of a USB wireless adapter is just potential, based on how many antennas it uses (and other factors, but let's not get into that). The only thing you need to know is that it's an up to figure. Higher rated devices are generally faster, but there's no guarantee.

Online speedtests don't test your internal network speed (the speed of the wireless adapter), it tests your internet speed as delivered by your ISP, in your case over a 4G device. Getting different results from your adapters is not a big issue as wireless is highly variable and it depends on your environment.

A wired internet connection is typically delivered over the copper telephone lines and the minimum standard is pretty much that you have a working telephone service as they cannot guarantee anything else. There are online checkers that can estimate what you'll get but the only way to really know is to sign up and test afterwards. This can mean that a landline connection is slower than a wireless one.

 

Thanks, searching online still returns the info that wireless is worse than adsl, but from what I've seen/have got, I didn't think that was right. It's a bit sad that I'd be losing out to get a fixed line connection (otherwise than being able to go mad with downloading).

So is the speed of the adapter more likely to have an impact if you are moving files rather than just connecting to the internet? Maybe if you're streaming video or something?

 

Under best conditions, wired is always better. With the mixed conditions of Australian internet, you have to test at each individual location. For example, my uncle gets about 4Mbps over ADSL, but gets ~20Mbps through fixed wireless... all because the tower is down the road whereas the DSLAM is pretty far away.

Yup. Transferring files between computers within your local network, including streaming, will be limited by the speed of the adapter(s). If you're streaming over the internet, you'll be obviously limited by the 4G device before anything else.

 

wired is about making the best conditions for file transfers.. wireless is about convenience then EoP stuck it's head in the door.. works well especially in rentals where cabling up is generally not done unless your landlord is ok with it before and you pay for it....

 

any "speed test" (which should more accurately be called a throughput test) will be limited by the speed of the narrowest bottleneck in the data link from client device and data source. with modern link electronics, chances are often good the bottleneck is elsewhere.

two reasons:

(1) bits is a measure of connectivity speed, bytes of throughput. it's the same reason the engine in your car has separate measures for RPM and horsepower, they are not describing the same thing. I have an engine here somewhere that will easily do 30,000RPM but its horsepower is very low indeed.

I absolutely detest "speed tests" which display results as xbps because what they are actually doing is taking throughput, reverse-engineering it back to a measure used to describe connectivity speed, and then that result doesn't match what Joe Average can see in his connectivity device, and Joe doesn't understand that going from layer 1 to layer 4 is not a direct translation - there are a whole bunch of variables around encapsulation, fragmentation, error control, transmission overheads etc that aren't constant from moment to moment on the same connection, let alone from one connection to the other.

(2) for the same reason we don't measure all distances in millimetres, it would rapidly become inconvenient. a scale of units of measurement is required.

for instance, the distance from Melbourne to Sydney is 874,000,000 millimetres, but 874km is more relevant to the scale. similarly, a 1TB hard drive is less usefully described a 8,796,093,022,208 bit drive. I wouldn't want to be doing cabinetry using kilometres as a scale though.

I've got gear at work that runs from 9,600bps through to 1,000,000,000,000bps but I tend to refer to the latter as 1Tb.

 

speed tests services test every part of the connectivity between your host computer and the remote server, and finds the lowest common denominator of all of them, if that's a congested link in the core network it finds that, or your local wireless, or your 4g modem.

Come on we all know a bigger number is better and faster for marketing terms, unless you put in a si unit that nobody knows and then it's cool that your the first to use that unit.

Marketing sucks


Back to the op thou, wireless is unpredictable and unreliable, it is also in most cases half duplex meaning that only 1 direction of data transmissions happens at any one time unit (fractions of a second) all these different queues and buffers and transmissions technologies mess with how TCP actually performs.

I tend to think of those internet speed sights a bit like home loan comparison rates, they are valid only in the conditions that are specified in the fine print but it does not mean that you will get that result.


Marketing screws with things like wireless.
Everyone thinks
they should get 5 bars all the time on there mobile or wifi at home
that it should always be connected at the highest bitrate possible
that they should get perfect coverage in their home all the time
that more power always equals better
that bigger numbers are always better



I think when you have been working in the industry awhile some things just really start to grind your gears when you understand how it works and it's just not possible all the time

 

Yes, but isn't the speed at which the adapter is connected a factor? If it's 72, how can you get a result of 150/180? (or am I doing the sums wrong)

 

72 and 150 are both greater than the 15 or 18 you're seeing on the speed test. Could just be variance, or your borderline for 150 and your adapter is jumping up and down in speed (wireless changes dynamically)

It can be that the adapter that sites at 72 stays at 72 for your location.

The adapter that reports 150 might swap between 72 and 150, or it could be dropping packets.


(i can't be bothered looking up the rates that it could jump between)

 

except the one that comes after the $...

what grinds mine is people that won't take the time to learn a new concept and actually make a conscious decision to remain ignorant, and then complain about the subject.

it seems a fairly base requirement that someone be able to explain how something should be working, before complaining that it's not. I don't mean being able to demonstrate the physics or mathematics, I mean just describe, in plain language applicable to the subject, the expected and perceived behaviors.

in terms of an internet connection, that means taking the minimal time required to understand things like bits vs bytes. if the person can't take the time to do so then complaints of "it's slow" don't get a lot of time from me.

a complaint of "my modem shows a 10,000Kbps connection, and I normally get a download speed of about 1MB/s. for some reason I'm only getting half that now but the modem hasn't changed, why?" is a question that indicates some intelligence.