Sorry for the lack of answer. Yes, I have some expertise in iptables and
firewalling in general.
One thing you should need is multiple IPs on each server. Then, via
iptables and cron you can change your source ip address every minute or so.
Give me some time and I can post examples.

On Mon, May 21, 2012 at 7:07 PM, <fmende...@terra.com> wrote:

>
> Hello Natalio,
>
> do you have a precise example on how to implement this?
>
>
> Thanks.
>
>
> On lun 21/05/12 4:35 PM , Natalio Gatti nga...@gmail.com sent:
>
> I can only think in one solution. Via iptables and src-nat. Not so-random,
> but you can change your outbound IP address every minute. And AFAIK, once a
> connection has been established, the nat table mantains the translation.
>
> On Mon, May 21, 2012 at 5:42 PM, <fmende...@terra.com> wrote:
>
>> Hello everyone
>>
>>
>> I am the owner of a growing hosting enterprise in my country (PerĂº), and
>> we are facing big rise on our client number.
>>
>> As an efect of this we are seeying a rise in mail outbound in our
>> servers. Even thoug we put limits to hourly sending, having more than 9k
>> clients, all delivering through the same cluster, it lacks of efectiveness
>> because each server in cluster uses only one ip for sending tasks. We are
>> now seeying blocking issues because of the many clents generated traffic.
>>
>> We talked to some people at godaddy and hostgator, as we know they use a
>> cluster system that includes on each server a list of IPs that rotates in a
>> random fashion, so even with high demand quality service on mail delivery
>> from client accounts is always achieved.
>>
>> I would like to ask for some guidance and help to this comunity on how
>> can we could implement such solution to rotate in a random or other way the
>> IPs for sending clients mails.
>>
>> I hope you people can see my situation and can help me with this. We used
>> to work with exim, but since we changed to QMT it was the best desition we
>> ever made on this matters. Now we need to push it to a next level.
>>
>>
>>
>> Thanks a lot.
>>
>>
>>
>
>

Reply via email to