On Tue, 4 Dec 2001, Hesham Soliman (ERA) wrote:
> > This is something that I was getting at.  If BU processing is not
>   > > implemented, then wouldn't all packets be routed via the HA?  
>   > 
>   > Not necessarily, AFAICS (if HAO-option is used).  Depends 
>   > on how strict
>   > your requirement on mobility (in this context, requirement for
>   > non-breaking connections if IP changes; depends a lot on 
>   > how architecture
>   > is designed) is.
>   > 
>   > Do you see any flaws in my reasoning (this wasn't commented 
>   > on) -- I think 
>   > this should answer some questions..
> 
> => Your understanding is correct, but I disagree with the 
> conclusion (if the above is a conclusion). It's clear
> to me (and many) that there are security hazards associated
> with the HAO (thanks to your draft). But rather than redefining
> mobility, or relaxing the requirements on mobility, I think 
> we should work on something that fixes the problem. 
> So my point is, let's fix the problem instead of redefining 
> the original goal. Breaking connections was always a no no !

You're right, I that wasn't a conclusion.  I only tried to enumerate the
dependencies and needs for BU and HAO.  A conclusion seems to be that
without HAO, you cannot get real mobility, and there would be no use for
BU.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords



--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to