[Top] [All Lists]

Re: [ietf-smtp] Draft for compressing SMTP streams

2016-08-18 03:07:23
On 18/08/2016 08:06, Arnt Gulbrandsen wrote:
Fwiw I find the arguments below unpersuasive, but here they are and maybe we can do this round very quickly:

Compressing can let mobile clients upload faster to submission servers, achieving a better approximation of atomicity. I find this weak because they do not seem to pipeline and hardly anyone seems to have shown any interest in lemonade, not even to consider and reject it.

Compressing can help ESPs and big receivers. Well, perhaps, but both of those have enough bandwidth to do it the simple, reliable way.

Compressing can help spammers avoid detection when on botnets by using less bandwidth 😅

I think points against it are:

- it will take several years before any new option is widely implemented enough to make any difference. By then it is probable that mobile clients will have better access to fast Internet, so the percieved benefit to them will be reduced, reducing incentive to implement the option. - compression will give little real gain for small messages. The biggest potential gains would be for big messages. However, big messages will usually contain attachments. Many common attachments are already compressed (eg JPG, PNG, DOCX etc) so further attempts to compress those will gain little. (AIUI base64 encoding, even of uncompressed data, also reduces compressability).

ietf-smtp mailing list