[BEHAVE] 答复: Using DNS64 to load balance to NAT64

> -----邮件原件-----
> 发件人: behave-bounces@ietf.org [mailto:behave-bounces@ietf.org] 代表
> marcelo bagnulo braun
> 发送时间: 2009年10月21日 15:40
> 收件人: Iljitsch van Beijnum
> 抄送: Cameron Byrne; behave@ietf.org
> 主题: Re: [BEHAVE] Using DNS64 to load balance to NAT64
> 
> Iljitsch van Beijnum escribió:
> > On 20 okt 2009, at 08:01, Cameron Byrne wrote:
> >
> >> Additionally, the NAT64 and the DNS64 should be able to communicate in
> >> such a way that the NAT64 can signal an overloading situation such
> >> that the DNS64 drops the overloaded NAT64 out of the round-robin
> >> prefix64 synthesis for some period of time.
> >
> > I have one problem here in addition to Andrew's objections:
> >
> > If make a given host move to a different NAT64, that means that for
> > some time, there will be an inconsistency between the external IPv4
> > addresses associated with different sessions. This makes protocols
> > like STUN fail, and as such, should be avoided.
> but you could solve that problem by making the pref64 used by the dns64
> dependent on the source address of the query, right?

If I understood it correctly, the external address for the internal host
learnt from STUN is used by the external host to initiate communication
session. So would it cause any problem when allocating prefix according to
destinations (i.e. A record contained in DNS reply)?

Xiaohu

> I mean, in that way, you could make sure that all the DNS queries coming
> to the dns64 server from the same source IP address get the same Pref64
> bt different sources get different Pref64
> 
> This would result in per source host load sharing and woudl address this
> particular issue i think
> 
> Regards, marcelo
> 
> 
> >
> > You really don't want to run so close to your translator capacity that
> > systems like this become useful.
> > _______________________________________________
> > Behave mailing list
> > Behave@ietf.org
> > https://www.ietf.org/mailman/listinfo/behave
> >
> 
> _______________________________________________
> Behave mailing list
> Behave@ietf.org
> https://www.ietf.org/mailman/listinfo/behave