| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8
Comments:
<0> absolutely
<1> cisco just wants to pump it's prices
<1> and it's stock
<1> is it possible to boot Mouring from console?
<2> the dichotomy between cisco hate and cisco adoration here is amusing
<1> just like we love and hate unix.
<3> cisco is a staple in network solutions. what hated is there buisness practices
<0> cisco's not that great
<0> there are better
<2> uh oh
<0> their service is bad as well :D
<4> heh
<2> here comes some juniper love
<4> cisco has some good products, as does juniper
<4> i use both :)
<5> **** that, ascend! ;)
<4> cisco switch, juniper firewall
<3> when i toured a datacenter there where brand names BFG routers, wam termination and the likes
<6> how effective is spam filtering nowadays?
<0> haha
<3> that i didn't know existed
<3> cisco who?
<6> i'm getting about 20 spam emails per day. ***uming someone is doing their job filtering email, what's an acceptable # of spam emails?
<2> that depends, omnix
<3> that would depend more on the person who writes the mail filters
<2> i've had the same email address for ~15 years, and i made the mistake of posting to mailing lists/usenet with my real address before spam existed
<5> omnix: if done by the right product, quite effective...
<3> if one gets thru, add another mail filter
<5> 20 is NOT acceptible
<2> i have excellent spam filters but i still have ~100/day that i look at and make decisions about
<2> my false positive rate is zero though
<5> i have about 2 spams i look at / month
<6> the answer i've been getting from the mail admins are, "nothing more we can do about spam"
<2> sure there is
<5> thats bs
<2> but when you get overly aggressive you start getting false positives
<5> omnix: do you run your own business/mail server? or just an end user?
<2> 20 spams/day should take about 1 minute to delete
<6> we run our own
<6> i run the network so i'm actually considering putting a postfix server in front of the mail server
<2> so you have to ask yourself if filtered legitimate emails would cost you more than a minute's worth of your time each day if you try to eliminate those 20
<5> then you need what im developing, which is an email filter :)
<6> we have about 200 people who also have the same problem
<1> i've made the mistake of posting to the Monastery :(
<5> but while i finish it, i would recommend you use a postfix + mailscanner frontend
<2> effective spam filtering cannot be done entirely algorithmically
<5> if you use the right tools and logic, it can, but it will never be perfect, and some input is always needed from the user.
<2> no it can't
<5> well, thats according to you...
<2> that's why postini are so great - they have teams of people who scan incoming mail to help tweak the filters
<2> nobody has proven so far that it can
<2> in fact you yourself just stated that it can't
<2> "it will never be perfect, and some input is always needed from the user"
<5> yes, it cant be perfect, but it can damn close....
<2> the perfect auto-pilot spam filter is the holy grail of apps these days
<5> im sure postini is not perfect
<2> no, of course not
<2> but they come closer than anything i've ever seen
<5> if human monkeys look at it, perhaps so
<2> they toss about 3k-5k messages per day destined for me with zero false positives
<2> ~200 squeak through, and my own filters catch about half of those
<5> have you considered your privacy?
<2> leaving me with ~100 to toss per day - about 3 minutes' work
<2> privacy?
<2> you have expectations of privacy when you write on an electronic postcard?
<2> if i want privacy in email i use pgp
<2> because i ***ume it's not private otherwise, postini or not
<5> and how do they inspect that?
<2> they don
<2> 't
<2> i've yet to receive a pgp-hashed spam message
<5> still, not that many people actually use pgp
<2> that doesn't make non-pgp emails any more private than if it were more widespread
<5> i'd rather have a local on-site filtering system that i can monitor and adjust
<2> i do have that
<2> but i'd rather let someone else handle the lion's share of it
<2> it conserves on bandwidth really well, too
<5> so youre just being extra cautious basically...
<2> when postini are catching 70-80% of incoming emails that's a lot of bandwidth i don't need to waste tossing things locally
<2> a lot of cpu i don't need to churn either
<5> as long as we can agree that there are uses for both solutions..
<2> we find it very cost effective
<2> yes, and we use both
<2> we're a postini reseller and it is like free money - it sells itself
<5> so what do you use for the other 30%-20%?
<2> on the rare occasions that folks are hesitant i give them a free month
<2> they always start paying after that
<5> what kind of interface and control do people have to this postini?
<2> i use rbls, local access tables, procmail recipes
<2> most of that other 20-30% is legitimate
<5> so basically, nothing rally effecitve
<2> we still toss some stuff locally, but not a heck of a lot
<2> again, almost all of what gets through postini is legit
<5> i see
<5> how much does it cost?
<2> mainly i'll throw up a filter if we have a user who is being bothered by someone they don't want to deal with
<2> the enterprise version, where you operate your own mail server, is about $1.25/user/month
<2> users are real bodies
<2> aliases are free
<2> you point your mx records to postini
<2> you acl your smtp server to only accept incoming from postini
<5> right...
<2> you have a web interface to manage your users - there is a batch upload/download feature
<5> another relay based filering service
<5> im wondering what that interface offers and looks like
<2> there are several defined spam "types", such as "make money fast", "racially insensitive", "xxx", etc.
<5> can you view reports on all emails? including filters?
<2> you can set filters for least/most aggressive on each of them on a per user or a per org basis
<2> you can see reports on the number/percentage of what they filter
<5> hrm. not sure i understand
<2> things that they're 100% sure is spam you will never have any more access to
<5> you have perhaps a screenshot?
<2> things they are pretty sure are spam but not 100% go in a quarantine box which you can manage
<2> ask me later when it's not midnight
<5> ok, i understand that
<5> heh, if i remember :P
<2> i'll take some screenshots when i'm at work
<2> anyway
<2> they also do virus scanning
<5> of course...
<2> and if you configure it so that your users cannot download the stuff in the virus quarantine they will give you money if a virus gets through
<2> users can look at the headers of stuff in the virus quarantine but they can't download the mail unless you let them
<2> as far as the spam, you can have postini mark up some x-headers for you and deliver it anyway so that you can make your own determination based on the score
<2> or you can just have them use the quarantine, which is what most of our customers do, as they don't want to waste the bandwidth receiving spam that they'll toss anyway
<2> we have customers that get 90-95% spam and tossing all that before it arrives really conserves their bandwidth
<5> yeah, but you can never trust headers
<5> The reason im asking all these questions is for research, im working on developing an email filtering system for businesses that host their own mail servers.
<2> so what's the difference between having postini give you the x-spam headers or, say, spam******in doing so?
<6> wettoast, so you're just the MTA and u send to the business's MDAs?
<5> and it has bee working very well for them
<2> nobody has ever done it well except fairly large companies, wettoast
<5> omnix: not sure what you mean there...
<2> i'm absolutely convinced that it can't be done effectively on a small scale like that
<6> u know what MTAs and MDAs are?
<5> roycroft: well, maybe you would be surprised....
<2> the bsdi mail appliance was a good idea but that didn't work out well because bsdi didn't have enough people maintaining the filters
<2> i'd like to be surprised
<2> but i'm dubious - having dealt with the spam problem for over a decade now
<6> roycroft, does postini reject mail from domains that are missing an SPF record?
<5> omnix: MTAs are Mail Transport Agents, MDAs, not ****ing clue you tell me....
<2> and i've yet to see you say anything negative about postini
<2> i hope not, omnix
<6> MDA = mail delivery agents
<2> spf is Evil(tm)
<6> like a pop server
<2> i know they don't as a matter of fact
<2> if they did most legitimate email would not get through
<2> spf is an aol marketing ploy
<5> MDA = Muscular Dystrophy ***ociation
Return to
#unixhelp
or
Go to some related
logs:
#computers
#delphi
#politics
#computers
#gentoo
oggy-you
teridhair
#nintendo
#beginner
#hardware
|
|