[AusNOG] Current "Best Practice" WRT email size

[Mark Newton]

On Nov 26, 2015, at 1:10 PM, Ross Wheeler <ausnog at rossw.net> wrote:
> I know email is being constantly asked to take ongoing abuse and to become the defacto file-transport-and-archive system of choice, particularly by the technically incompetent, but how far does it go?

It goes to where the users demand.

There’s no specific reason why email can’t be a defacto file-transport-and-archive system of choice. It’s carried by TCP just like every other file-transport-and-archive system, and everyone has clients for it. If the users want to use it for that, what’s wrong with it?

> Case in point: earlier this week, I had a call from a customer "needing" me to increase our mail size. (I thought we were 'reasonably generous' in current global terms, at 16MB per message). I asked what he considered it needed to be, his response was that "right now" he needs 50-60MB, but that he thought it shouldn't have any limit - but if it had to, that 300-500MB per message would "probably do for now”.

Go back to the question, “Why does the limit exist?”

SMTP servers used to have the limit because large files took a long time to send, bandwidth was expensive, storage space also cost a lot of money, and if a message was too big the client would probably crap its dacks when it tried to receive it anyway. 

Even a decade ago, 10 Mbyte limits were the norm. You’re currently happy with 16 Mbytes, even though all the resources which were extant when the limits were first envisaged have scaled by, what, a factor of 1000 in the right direction?

You currently have a limit which prevents your users from sending a RAW format image off their digital camera as a file attachment. That seems unusually small to me.

Is there any specific reason why you couldn’t update it to 1 Gbyte?  If you had allowed it to grow at the same rate as bandwidth and Mbytes-per-dollar storage costs over the last ten years it’d probably be closer to 10 Gbytes by now, so setting it to 1 Gbyte is an order of magnitude less than organic growth.

  - mark