mike-wsm

I use a data counter widget to keep track of wifi usage because it's right there on-screen, and can be set to display daily usage and monthly usage from my start-of-month which is the 28th.

There is a BT counter available but I have to log in and then allow for it starting on the first of the month, and being a day or two behind, so I tend not to use it.

But occasionally I check both and it seems that the BT count is up to twenty per cent lighter.

Is there, I wonder, some measure of data compression on ADSL which is unwrapped in the router before entering wifi-space?

mixture

Nope, no compression on ADSL.

BT use packet shaping on their network and may also be serving up data from proxies they host.

Therefore for a given time period, the data transferred may be different as your traffic throughput gets artificially skewed by their network systems.

Also, are you counting the router throughput or your wifi throughput ? It may well be you're picking up non-internet traffic (e.g. sending files to print) if you're counting the wrong thing.

Gertrude the Wombat

Nobody can give you a definitive answer without knowing how each counter does the sums - there are plenty of different plausible ways of doing the counting, it wouldn't be surprising if they counted different things even if they were counting the same traffic.

Milo Minderbinder

A number of points here
1) the wifi monitor on the PC is only going to monitor just that traffic between the specific PC and the router
2) if you have a BT Home Hub, then that actively looks for firmware updates from the BT servers and automatically updates itself. It may do this a couple of times a month. Thats extra data
3) again, if you have a home hub, third parties may be loggging into your open BTWiFi or BTFon channels. Thats more extra data
4) there will always be a degree of redundancy in the data sent over the internet due to error checking and correction. Depending on the router, phone line and max packet size this can be quite significant

mike-wsm

Milo - Thanks!
1) It's just one tablet on wifi, nothing else on my part of the router.
2) Understood, but that's more data on the BT counter, not less.
3) Third party data is not charged to the hub owner.
4) Understood, but that's more data on the BT counter, not less.

Gertrude - Thanks!
I really don't understand that. They are selling me 10GB a month but you say this is a variable feast. Ooh-er! Have learned never to argue with cantab even when they're wrong (last time it was failure to recognise an inverse twelfth power law, which I, in my ever-simple approach to life, regarded as significant)

Mixture - Thanks!
Skewing: wouldn't this make it fatter, not slimmer?
Peripherals: nothing on this router, only the tablet. Sorry, I live a very simple life.

mixture

Spread it out a bit like butter on toast.

At least that's the analogy that comes to mind after a glass or two of new years champers....

mike-wsm

mixture - okay, understood, so not going to make a lot of difference over a day or a month, I guess

mixture

You might want to experiment with different measurement tools, see whether they yield different results.

I doubt it but if your router happens to support SNMP then that would be a good source of data to use for measurements.

bnt

Don't forget about HTTP compression - which operates between web server and browser. A counter would count the compressed data size. If the server tries to send it but the browser can't handle it, the data can be resent, skewing the results.

Gertrude the Wombat

Are they charging you for bits down the wire? Or payload bytes in TCP and UDP packets (ie not charging for transport protocol layer overheads, ICMP, ARP, etc etc)? Or payload traffic through their web (etc) proxies (ie only charging for stuff that the end user would recognise as data they've downloaded)? Or ...?

This could be a complete red herring, for all I know there's a standard way of accounting and charging and everybody does it the same way ... but if so I don't know what it is!!

mike-wsm

Gertrude - Thanks, good word, payload. I rather assumed adsl and wifi are both counting payload, but it sounds rather as though bt is counting payload and that the tablet counts gross wifi traffic. That would account for a difference, though ten to twenty per cent is a lot. Difficult to quantify exactly because the bt count is 24-48h late. The app takes its count from the tablet, which is gingerbread.

mixture

My guess would be volume billing of anything, if they're doing it correctly it should be on a 95th percentile 5 minute basis but although that's the way its done wholesale, they might just be using total cumulative volume (or something else) rather than 95th percentile..... time to check their T&Cs me thinks !

mike-wsm

mixture - Thanks! I really haven't a clue. I just need to know why my tablet counter is pessimistic and it would be useful to have a fudge factor, for example telling me that I can run to 12GB on my tablet counter without exceeding my isp's 10GB.