Am 29.07.2012 11:48, schrieb Mark Alan:
>> if you do not trust you OUTGOING traffic the only valid
>> reason is that you doubt your machine is comprimised
> 
> [The problem, as said in another email, is (mostly) solved]
> 
> - I do not trust anything connected 24h to the Internet

then shut it down
do not run things you do not trust

> - I do not trust anything in a Xen VPS that sits in a datacenter
> owned / managed / maintained by I do not know exactly who

so get a root server instead let things maintain from
people you do not trust

> - I do not trust any software, open source or otherwise, that has a
>   level of complexity high enough to not be fully understood by the
>   installer, maintainer, user, etc.

a well made postfix setup is 100% understandable
if your mailserver setup is NOT understodd by you SHUT IT DOWN

>   [ Just google for "OpenSSH FBI backdoor".  Its IPSEC stack was a
>   relatively small but nevertheless highly sensitive piece of software.
>   Look how it managed to elude, for so many years, so many security
>   conscious people, including most of the more security conscious
>   developers around: the developers of the OpenBSD - the "Ultra-Secure
>   Operating System". ]

completly off-topic

> This 'thing' just become so complex and with so many variables, that
> it became impossible to know them all and to account for them all.
> We can only reduce the size of the target and make it a little more
> difficult to break in.

you make it not difficult to break in with
OUTGPING rate controls

> And that is why we keep an eye on syslog and cousins and ask for help
> here on this list when we start to see firewall drop outs related with
> Postfix.

keep your eye in intrusion events
instead cripple down your network stack

>> and NO a synflood will never come in the OUTPUT stream
>> except your machine is compromised, but if so shut it down
> 
> I am afraid that time will show you otherwise

i am afraid this will not happen because my machines are
not partly out of control like yours

> These systems are not 'simple', not even 'complicated', they are real 
> 'complex systems'

not really

> And, worse, with so many knowledgeable people with time and resources to
> invest into breaking these systems, these are now real 'complex adaptive
> systems'

if knowledgeable people maintain them you are mostly safe
not knowledgeable people should better not run public servers
because in the case of a intrusion or conig mistake the are
doing damage also on thir parties

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to