![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
> From: Richard Shockey <rshockey at ix.netcom.com> > ... > The simple fact is that E-Mail attachments are only going to get bigger and > more frequent, for no other reason than RFC2305 and RFC2532 The Internet > Fax standards that use SMTP. ... I don't understand that reasoning. Don't classic fax machines use a rather low rate, less than 28 kbit/sec? 9600 bit/sec sounds right, but I'm probably wrong. Isn't the Internet fax protocol intended to be used not only over T3's, but also over SMTP/TCP/IP/PPP/v.34 modems? If so, and unless Internet FAX switches to an amazingly bad encoding, most Internet FAX's will not be much larger than they are now. How often does your current old FAX machine spend an hour receiving a single FAX? One hour at 28 kbit/sec is less than 13 MByte. Thus, it seems unlikely that RFC 2305 and RFC 2532 will stress the current de facto limits on SMTP message sizes. ] From: "Martin Djernaes" <martin.djernaes at kecam-han.de> ] I know that the internet were not build for "general use", but it is the ] life of the net today, at it should be the goal for the people ] implementing it (us?). Let us get away from the idea that it should ] always be used the way we implemented it and change our view to how ] would the user like to use it. There is no magic. Users have been demanding the impossible from SMTP on the grounds that they don't understand why not for decades. The same people who wouldn't try to load 10 ton of gravel in the back of their light duty pickup truck expect email to be 99.9999% reliable, have end-to-end delays of less than 15 seconds 99.9999% of the time, and to move arbitrarily large files in at most 15 seconds. Just because people want something does not mean that it is possible or even desirable. It's actually worse than that. Instead of demanding a way to move large files, they demand a way to move large email messages, and they cannot understand the distinction. ] If the user would like to transfer 28 MB we should make it possible ] (there is always somebody who is in front of the big group, so I do not ] say that just because one person wants it we have to make it possible). There are scaling problems. Consider how many disk drives an ISP would need to dedicate to mail directories to support many users moving 28 MByte messages. A 32 GByte drive can hold only about 1000 such messages. Consider an ISP with 1,000,000 customers. Remember the bad old days before the ESMTP extension when a second SMTP hop would run out of space in its spool directory, and the message would try to bounce. ] If the mail service from today isn't the right solution, we should ] invent a better. If all providers would allow a user to place XX MB via ] FTP on a server, ... Didn't I see mention of something called an "external body"? That notion avoids the scaling problems of requiring that all spool directories and all destination mail directories be large enough when many people decide to forward a 28 MByte Good Times virus. Vernon Schryver vjs at rhyolite.com
Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.