http://www.mail-archive.com/[email protected]/msg97598.html

Dev

Wednesday, June 1, 2005, 12:58:09 PM, you wrote:

MTS> That interface and features are EXACTLY what
MTS> we should have in Imail bynow. Who's mail server
MTS> software is that?

MTS> Dev wrote:  My bad for not being
MTS> hyper-accurate with my
MTS> terminology.

MTS> True, "tarpitting" is the ability to trigger a
MTS> "time-out" period against originating IP's that exceed
MTS> a specified limit of bad recipients.

MTS> Your criticism of resource usage with "slowdown"
MTS> tarpitting is certainly valid. However, any decent
MTS> tarpit implementation also allows you to change that by
MTS> specifying "close connection", which I do. Note the
MTS> screenshot of the new server's tarpit interface. See?
MTS> No TCP sockets were harmed or wasted during the filming
MTS> of this picture. :-)

MTS> We finally abandoned Imail because, at the time, they
MTS> didn't have even a rudimentary dictionary-attack
MTS> solution, (not to mention the geriatric webmail.)
MTS> Attacks simply chewed up CPU cycles as Imail chased
MTS> it's own tail.

MTS> Now, the new registry-enabled "hang up" in Imail 8.2 is
MTS> a start--it's crude but effective. In the next version,
MTS> I trust that Ipswitch will enhance it to be world-class
MTS> and fully-featured!


MTS> Dev


MTS> Thursday, May 26, 2005, 10:01:52 AM, you wrote: 

MTS>       My understanding of the definition of
MTS> tarpitting is slowing down the
MTS> delivery of messages to multiple
MTS> recipients...If my understanding is
MTS> correct see "delay between recipients" in
MTS> the SMTP Advanced tab in IMail
MTS> Admin. If not please enlighten me to the proper mening

  

  

MTS>   LC> SMTP tar-pitting means that the SMTP server
LC>> artificially injects delays 
LC>> into its side of the SMTP dialogue
LC>> (specifically, before returning SMTP 
LC>> response codes) with the SMTP client.  The
LC>> delays can be 10's of seconds or 
LC>> many minutes.

LC>> tar-pitting is an attempt, by one MX, to
LC>> "hurt" the millions of IPs that 
LC>> abuse MXs by slowing the abusive IPs and
LC>> "wasting" their tcp sockets and 
LC>> SMTP sessions.

LC>> In adverse effect, the most hurt is done to
LC>> the MX machine server itself 
LC>> since it must also waste a tcp socket and a
LC>> SMTP session on the MX side for 
LC>> eacd tarpitting action.  In this
LC>> shoot-MX-in-foot battle, which side has 
LC>> the most tcp/smtp resources to waste?  A
LC>> single MX or the aggregate of 
LC>> millions of abusvie IPs attacking it?

LC>> I always recommend against tar-pitting in
LC>> general.  And from the large 
LC>> number of IMGate/Imail users have added IMGate
LC>> because IMail SMTP server 
LC>> was unstable and incapable of supporting
LC>> high-volumes of simultaneous SMTP 
LC>> connects, I strongly recommend against any
LC>> tar-pitting by the Imail server.

LC>> The great benefit of IMGate to IMail users is off-loading the
LC>> DNS+SMTP-to-Internet dialogue from Imail to
LC>> IMGate.  Tar-pitting goes 
LC>> exactly in the opposite direction, loading up
LC>> IMail SMTPD with more demands 
LC>> for tcp/smtp resources, and insanely, with no
LC>> noticeable advantage, and NO 
LC>> pain for the attackers.

LC>> Len

LC>> _____________________________________________________________________
LC>> http://IMGate.MEIway.com : free anti-spam
LC>> gateway, runs on 1000's of sites



  


  
  







To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/

Reply via email to