nanog mailing list archives

Re: [nznog] Web Servers: Dual-homing or DNAT/Port Forwarding?


From: "cb.list6" <cb.list6 () gmail com>
Date: Tue, 10 Dec 2013 17:27:23 -0800

On Dec 10, 2013 2:32 PM, "Geraint Jones" <geraint () koding com> wrote:

On 11/12/13 10:13 am, "Alex White-Robinson" <alexwr () gmail com> wrote:


Wotcha,

Number 1 gets you thinking along the IPv6 route (no pun, and imho :) )
since you have to treat each boxes as if it was public.

I see this kind of statement surprisingly often. Having a public address
doesn't make a device public.

Yes it does, it makes end to end connectivity work again. NAT broke that
(and its one of the best things about v6). People have been relying on the
fact that you need rules to get through a NAT to reach a box - thereby
having NAT work as an inbound firewall. NAT != Security.

But yes having a public address means your box is public, you have to do
something to STOP traffic getting to it. With NAT you have to do something
to ENABLE traffic to get to it.


Correct. IPv6 correctly supports the end to end model.

Firewalls can be scalably implemented on host, not middle boxes.

The firewall mindset is locked in from  the win2k days, NAT reinforced
that, and it is worth re-evaluated removing firewalls with ipv6

Question: are nanog meeting networks stateful firewalled?  Follow up
question -- if there is no firewall, do folks experience a higher degree of
malware infection after the meeting ?

CB

I don't really see a drive to have devices exposed to the internet
without
a stateful device in front of them in IPv6 world. People shouldn't allow
unsolicited connections to hit your internal workstation on any address
scheme.

Cheers,
Alex.


Date: Tue, 10 Dec 2013 05:56:41 +1300
From: Pieter De Wit <pieter () insync za net>
To: nznog () list waikato ac nz
Subject: Re: [nznog] Web Servers: Dual-homing or DNAT/Port Forwarding?
Message-ID: <52A5F649.7070904 () insync za net>
Content-Type: text/plain; charset="iso-8859-1"; Format="flowed"

Hi,

I normally use a combination of "1" and "2". I prefer 1 for weird and
"not nat friendly" protocols, like SIP or some other application. The
general rule of thumb is to use number 2 in other cases. In both setups,
remember to deploy local firewalls as well. This will help for the case
when a box on the subnet is hacked.

My other twist is to deploy "1" without the private NIC, along with
local firewalls (and as you said, dedicated FW).

Number 1 gets you thinking along the IPv6 route (no pun, and imho :) )
since you have to treat each boxes as if it was public.

Cheers,

Pieter





Current thread: