Malcolm on 18 Oct 2007 13:36:03 -0000 |
On Wednesday 17 October 2007 4:23:12 pm Matt Ayres wrote: > Not exactly. The bandwidth a home user can get these days is due to the > fact that the broadband companies have done a lot to ensure no servers > are running and are oversubscribing the capacity of their networks, but > not enough that you notice. Servers use a lot of bandwidth so by not > allowing them the broadband providers can afford to oversubscribe more > and therefore offer more to be competitive. That depends very much on the server. Mine (which was living in my closet until I moved and lost my DCA dsl and had to use Verizon) runs a webserver, mailserver (postfix/postgrey/amavis/spamassissin stack, imap, pop), hosts half a dozen users and a small community website as well as my personal stuff, etc. and has barely any traffic - it was running fine on a low bandwidth DSL line. On Thursday 18 October 2007 12:04:24 am Brent Saner wrote: > you're vastly focusing on the business model here. > residental (and even some small businesses) mail servers don't get the > amount of spam to where it becomes unmanageable (trust me on that. used to > work for a little tech shop that AFAIK didn't have any filtering at the ISP > level. the admin set up a bayesian and spam assassin and we rarely saw more > than one spam every two days if that.) My server gets around 15,000 a day. Most of that is dictionary spam so gets dropped at HELO. 20-50 a day (for all users) makes it through the spam filters (it's gone up recently, and I haven't had time to trac down any improvements yet). > it all has to do with being smart about it. publicly displaying a direct > e-mail is a no-no. using an e-mail addy you don't want spam to get to for a > mailing list is a no-no (because many mailing lists, such as PLUG, publicly > archive- as they should. but crawlers find e-mail addys from this). This is where address extensions come in very useful. It lets you identify the spam sources, and start dropping emails. In a couple of months when enough spam to this address gets through the filters, I'll change my subscription address and add this one to the reject list. As it's a public list, it'll get a human-friendly reject message so real people trying to reach me will get an error with a working address. That's cut down on most of my spam from publicly archive places. (Now if only bug trackers would provide an easy way of either masking the email from crawlers or changing your email on every ticket, that would get almost all the rest). The address extensions also make it easy to filter mail, and to train bayesian spam filters if there's an address you need to leave open (like a mailing list or bug tracker) without biasing the rest of your email. ___________________________________________________________________________ Philadelphia Linux Users Group -- http://www.phillylinux.org Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce General Discussion -- http://lists.phillylinux.org/mailman/listinfo/plug
|
|