Re: DNSSEC HW Support

2010-03-16 Thread Gary Wallis

I'd like to get your feedback on the following thoughts regarding DNSSEC HW 
support.

Any layer 2 or 3 devices forwarding frames or packets should not be affected by 
the implementation of DNSSEC regardless of the type of protocol (TCP/UDP) or 
the query size (large or small).

Layer 4 devices (smart switches) should not be affected by the implementation 
of DNSSEC using the same logic.

My thoughts are these products simply forward data based on an frame, IP 
address, or protocol and should not be affected by the implementation of 
DNSSEC.  Would you agree?

Thanks in advance.



I think you are basically correct except for one very important caveat:

DNS BGP anycasting (in wide spread use by many large operations,) where 
you might need to sign zones on the fly with special crypto hardware.

___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users


Re: DNSSEC HW Support

2010-03-16 Thread prock...@yahoo.com
  I'd like to get your feedback on
 the following thoughts regarding DNSSEC HW support.
  
  Any layer 2 or 3 devices forwarding frames or packets
 should not be affected by the implementation of DNSSEC
 regardless of the type of protocol (TCP/UDP) or the query
 size (large or small).
  
  Layer 4 devices (smart switches) should not be
 affected by the implementation of DNSSEC using the same
 logic.
  
  My thoughts are these products simply forward data
 based on an frame, IP address, or protocol and should not be
 affected by the implementation of DNSSEC.  Would you
 agree?
  
  Thanks in advance.
  
 
 I think you are basically correct except for one very
 important caveat:
 
 DNS BGP anycasting (in wide spread use by many large
 operations,) where you might need to sign zones on the fly
 with special crypto hardware.

So if I'm testing a router for DNSSEC compliance, you'd recommend I run a test 
using RIP or OSPF, then a separate test for BGP.  Is that correct?

I'm trying to figure out how many tests I need to run for an individual product 
(layer 2, 3, 4, and 7) before I can say it is completely DNSSEC compliant.


  
___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users


Re: DNSSEC HW Support

2010-03-16 Thread Niobos
On 2010-03-16 15:57, prock...@yahoo.com wrote:
 I'm trying to figure out how many tests I need to run for an
 individual product (layer 2, 3, 4, and 7) before I can say it is
 completely DNSSEC compliant.
By definition, any layer 2, 3 and 4 product is DNSSEC-agnostic: DNS with
or without SEC-extension is considered payload. If a L2,3 or 4 devices
does work with DNS and doesn't work with DNSSEC, it's broken and needs
replacement. For completeness: switches and routers are layer 2 and 3
respectively.

Layer 7 devices might be affected, since they may preform extensive
checking on the DNS-content itself.

To answer your question: 0 tests for layer 2, 3 and 4. To be completely
compliant, you'd need to run an infinite number of tests for layer 7
devices. I'd test the different algorithms, including some very recent
(RSASHA512) and different security statuses (bogus, insecure, secure).

Niobos
___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users


Re: DNSSEC HW Support

2010-03-16 Thread Warren Kumari


On Mar 16, 2010, at 11:39 AM, Niobos wrote:


On 2010-03-16 15:57, prock...@yahoo.com wrote:

I'm trying to figure out how many tests I need to run for an
individual product (layer 2, 3, 4, and 7) before I can say it is
completely DNSSEC compliant.

By definition, any layer 2, 3 and 4 product is DNSSEC-agnostic:


Well, yes, kinda.

Unfortunately there are a large number of things like firewalls and  
consumer CPE that folks think of as layer 3/4 devices, but that do  
silly things like assume DNS is only UDP, or max out at 512 bytes or  
force DNS proxy mode.


While we could argue for hours abut whether they are really only l3/l4  
devices, it wouldn't change the fact that folks think of them as  
routers.


ICANN SSAC / CORE released a report a while back: http://www.icann.org/en/committees/security/sac035.pdf 
 and I know that I have seen a bunch of other more recent tests.


W


DNS with
or without SEC-extension is considered payload. If a L2,3 or 4 devices
does work with DNS and doesn't work with DNSSEC, it's broken and needs
replacement. For completeness: switches and routers are layer 2 and 3
respectively.

Layer 7 devices might be affected, since they may preform extensive
checking on the DNS-content itself.

To answer your question: 0 tests for layer 2, 3 and 4. To be  
completely

compliant, you'd need to run an infinite number of tests for layer 7
devices. I'd test the different algorithms, including some very recent
(RSASHA512) and different security statuses (bogus, insecure, secure).

Niobos
___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users


--
Beware that the most effective way for someone to decrypt your data  
may be with rubber hose. --- SSH 1.2.12 README



___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users