Thank you for your quick response.  I do see your points about Fast Fail 
and see why it could hamper a spammer.  But how are the benchmarks for
servers like this though.  Would an ISP that accepts mail for thousands 
of domains be able to use this in production though?

If I was to make a for loop (from i to 1,000,000) and 
send a 10Mb file to people with email address 
[EMAIL PROTECTED] (steve1, steve2, steve3, etc).

James would accept all of that bandwidth (all of those 10 Mb files first).  
So it seems like whoever has a larger bandwidth pipe wins.  If you are 
getting billed at 95% though this could hurt you.  
Of course a system administrator should catch this though

In the link that was sent it did mention "basic message-rate cap".  
Something like a simple iptable rule do you mean?  Or is there something extra in 
James.

Thanks,
-Steve

On Thu, 7 Aug 2003, Serge Knystautas wrote:

> Steven Job wrote:
> > How does James work for large enterprise mail systems.
> > I was reading the following faq item.
> > http://james.apache.org/FAQ.html#2
> > 
> > It seems like this could be a problem on larger systems.
> > 1) All spam will be delivered without verification upfront.  This can 
> > cause problems in that all of this data will be downloaded first.  
> > Couldn't this cause bandwidth limitations?
> > 
> > 2) Couldn't people easily give you a DOS attack by sending to large files 
> > to millions of people in your same domain?  James will accept all the mail 
> > and then see if it should deliver it.  This could be a lot of bandwidth 
> > and disk room for email messages that never should be delivered.
> > 
> > I might be making a mountain out of nothing at all.  But I was just 
> > wondering if these are actual limitations and if others had problems.
> 
> http://nagoya.apache.org/wiki/apachewiki.cgi?NoFastFail
> 
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to