O Canada! > Canadian > TekSavvy > what causes difference in usage total?

what causes difference in usage total?

Are both measuring in base 2 or base 10?

I.e. 1024 vs 1000

In base 2, 1GB (Can be noted as GB or GiB) is 1024 MB, 1048576 KB
In base 10, 1GB is 1000 MB, 1000000 KB
That's almost a 5% difference.

both are measured in base 2 so that does not account for the difference.

Under tomato my time is set for EST timezone, and I suppose videotron will do likewise. So that cannot be the cause either.

reply to vientito1
Perhaps there is overhead?

reply to vientito1
I have the same question. My router with OpenWRT always reports a usage lower than what gets reported by Videotron.

I use a software called vnstat to measure and create visualizations of my usage.

The difference can be substantial as well. Like 10-20 percent.

I must admit though that I never checked if it was base 2 or 10. I'll check that later.

it's never consistent.. sometimes my router measure is higher and other times it is lower.

reply to vientito1
when i was on DSL i found that Tomato always reported around 10% less than the tracker at TSI

differences in measurement were too far off to account for it

TBH i suspect that there's some kind of rounding or BS that inflates the usage, but i couldn't ever figure out what it was

and the incumbents aren't exactly transparent in how they measure, they just insist that its correct (even though many times its been proven wrong)

reply to Funky_
I use software as well!

reply to vientito1
Get UNLIMITED and stop worrying about it.

reply to rjbrake

reply to vientito1
depends on your software that's measuring it. Most likely it doesn't actually count bits, it estimates them based on throughput speeds.

For instance, PFsense RRD graphs take a sample of the speed at certain times. my numbers may be off, but it gives you the concept. For the last day it checks it once every minute, takes the 2 points and gives you an average speed over that time which can then be used to approximate usage. But as time goes on (my router is 3 years old so that would be a lot of data), it drops some of it and goes to a 5 minute average. Well a lot can happen in those 5 minutes. The result is that the "max" throughput changes and the averages decrease. If the through put is a function of those, it would then go down aswell.

Other things to consider are at what point you're measuring.
Lan traffic? - would have inter lan traffic included
lan-wan traffic? would just include requests and no over head from the router
Wan traffic - should be pretty close.

Basically it comes down to how it calculates or estimates the throughput. almost all usage counters are estimates as actually counting them is highly inefficient.