Why your data is really disappearing

[jannievanzyl]

SauRoNZA pretty much summed it up but important to reiterate that the network systems will likely measure lower down the stack than an end-user device which will measure at the Network Layer and not lower down.

Then you have the zero-rating of many protocols, not just sites, for example ICMP.

All, in all, this means the number will not always (if ever) align exactly.

 

[Swa]

The data usage you measure on your end wouldn’t be accurate for the lack of network overhead that applies which isn’t visible to you.

Also the “formula” you are thinking of is probably along the lines of the minimum increment at which a call record is cut.

There however is still no loss and everything gets charged for that should be charged for.

You do realise the networks are made of essentially routers that count bytes right?

Only much bigger and better ones worth millions of rands compared to your consumer device that only sits on your side of the connection and therefore can’t possibly measure overall usage accurately.

More than likely the very means by which you are measuring is in fact exactly what adds the extra data you think you are using that they are charging you for, because unless setup correctly your monitoring will have a data cost.

No if I send or receive a TCP packet it gets counted payload and header, or overhead as you call it. Explain why you think my measurements are inaccurate. This sounds like the excuse MTN was using for charging people for their own network overheads and not just data actually used. Now if the networks can't measure it the same and instead measure their own network traffic it just reaffirms it to be an estimate and not an actual count. Then they apply a formula to ensure they don't overcharge. Cell C is better in this regard charging only between 95-96% when I still measured it. If they were able to measure up to the byte data used it wouldn't be necessary to under bill.

 

[SauRoNZA]

No if I send or receive a TCP packet it gets counted payload and header, or overhead as you call it. Explain why you think my measurements are inaccurate. This sounds like the excuse MTN was using for charging people for their own network overheads and not just data actually used. Now if the networks can't measure it the same and instead measure their own network traffic it just reaffirms it to be an estimate and not an actual count. Then they apply a formula to ensure they don't overcharge. Cell C is better in this regard charging only between 95-96% when I still measured it. If they were able to measure up to the byte data used it wouldn't be necessary to under bill.

1. There are no estimates.

2. They can and do measure up to the byte level.

3. Your measurements are wrong.

You haven’t answered the question of HOW you measure it.

I literally support the software that Jannie (and everyone else pretty much) use to do their billing, I can tell you as a point of fact there are no estimates or formulas to reduce billing.

 

[Next Thing]

1. There are no estimates.

2. They can and do measure up to the byte level.

3. Your measurements are wrong.

You haven’t answered the question of HOW you measure it.

I literally support the software that Jannie (and everyone else pretty much) use to do their billing, I can tell you as a point of fact there are no estimates or formulas to reduce billing.

I agree

Depending on where you measure it will define what figure you get from a user perspective.

If you measure it on the ethernet port that plugs into your home network you will get one figure that does not include any of the dodgy/misc traffic that does not get through your device's firewalling.

If you measure it on the port the ppp based connection is going via to the isp you may be including usage that the isp is not counting in from an overhead point of view.

If you are measuring it on the ppp based connection interface you will probably get the most accurate figure disregarding some traffic that may be zero rated.

Which figure will probably most accurately reflect the ISP figure. The 3rd one. But there are caveats in that as well.

It all depends on where you are getting your figures from and where exactly the isp is measuring from. And most consumer devices do not allow you to be granular enough

 

[Swa]

1. There are no estimates.

2. They can and do measure up to the byte level.

3. Your measurements are wrong.

You haven’t answered the question of HOW you measure it.

I literally support the software that Jannie (and everyone else pretty much) use to do their billing, I can tell you as a point of fact there are no estimates or formulas to reduce billing.

1. According to what was stated here there is.
2. Not my actual usage.
3. No they aren't.

As I stated previously I used to use software, Netmeter. Now I do not care how they measure it. From my perspective my usage is that which I send and receive. The network's traffic and overhead isn't my traffic. Now it appears they can't measure this and measure something else instead and then apply a formula to reduce the usage to not overcharge. Some networks reduce it more than others. That is the definition of estimating.

 

[SauRoNZA]

1. According to what was stated here there is.
2. Not my actual usage.
3. No they aren't.

As I stated previously I used to use software, Netmeter. Now I do not care how they measure it. From my perspective my usage is that which I send and receive. The network's traffic and overhead isn't my traffic. Now it appears they can't measure this and measure something else instead and then apply a formula to reduce the usage to not overcharge. Some networks reduce it more than others. That is the definition of estimating.

They don’t.

 

[graybeard]

No-one is disputing that there can be differences between a consumer's central router logging and the service provider's. A couple of megs I can understand, but a difference of 400 Mb to as much as 1.5 Gb on a 20 Gb bundle? C'mon! ESPECIALLY when other service providers are pretty close to the consumer's expected, but one service provider in particular has massive discrepancies in their favour!

Somewhat telling is that they will avoid giving a reasonable explanation (actually, any kind of explanation) when challenged on central, point of egress/ingress, logging.

 

[Swa]

No-one is disputing that there can be differences between a consumer's central router logging and the service provider's. A couple of megs I can understand, but a difference of 400 Mb to as much as 1.5 Gb on a 20 Gb bundle? C'mon! ESPECIALLY when other service providers are pretty close to the consumer's expected, but one service provider in particular has massive discrepancies in their favour!

Somewhat telling is that they will avoid giving a reasonable explanation (actually, any kind of explanation) when challenged on central, point of egress/ingress, logging.

Exactly. But as usual Jannie is dodging and flat out denying the issue.