We have the same secure concern.

So typically NVE auto-discovery shall be used with some authentication 
measures and it owns some security supporting features, for example 
Session-ID field.

Presentation time limited, pls refer to the draft itself for further info.

Thanks!

Zhongyu









Anoop Ghanwani <[email protected]> 
发件人:  "nvo3" <[email protected]>
2014-07-25 07:48

收件人
"Joe Pelissier (jopeliss)" <[email protected]>
抄送
"[email protected]" <[email protected]>
主题
Re: [nvo3] Comment on draft-gu-nvo3-virt-edge-auto-discovery






I had the same comment/concern after listening to the presentation.

One thing that I found significantly different from the VDP approach 
(which I am not sure that I can agree with as being the right approach, 
simply because it doesn't seem secure) is that, with this scheme, the VM 
directly discovers/registers with the NVE; in the case of the VDP 
approach, it is the hypervisor which does the registration.

Anoop


On Thu, Jul 24, 2014 at 11:22 AM, Joe Pelissier (jopeliss) <
[email protected]> wrote:
Greetings:
It appears that the protocol introduced in 
draft-gu-nvo3-virt-edge-auto-discovery largely duplicates the work that 
was done in IEEE with VDP.  I would like to solicit the authors’ view of 
potentially using VDP for this purpose.
 
Thanks!
Joe Pelissier

_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3

_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3


_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3

Reply via email to