[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: ocean: do not boil



I inherently agree with avoiding NAT (that is what DSTM is all about).  But I would like to hear discussion on Tony's mail to Rob.  I think we have a serious issue to reach a group resolution on here and Tony has stated it very well.  I believe we need to work on NAT and will work on it even though I don't like it and will avoid recommending it to any user I have to support.  But I don't like what I am hearing from some here nor that all we care about are ISPs which is what I think I am hearing from some.

Why is it some are saying "don't do as I do but do as I say".  We were told that we needed the scenarios (and ISP is only one of them) to complete analysis.  Yet I am hearing same proponents tell us we should select a default before we have them all worked out?

This is not good folks.  Or maybe I am missing intentions or assumptions?

/jim

> -----Original Message-----
> From: itojun@iijlab.net [mailto:itojun@iijlab.net]
> Sent: Wednesday, September 25, 2002 1:36 AM
> To: Hesham Soliman (EAB)
> Cc: v6ops@ops.ietf.org
> Subject: Re: ocean: do not boil 
> 
> 
> >=> I think we're drifting from the main issue 
> >a little bit. Here is a scenario that I'd like
> >to understand how it would work:
> >What do we recommend a new small 3g operator who only
> >has 10 million subscribers (just a small town
> >in china!) and wants to allow people to run p2p apps:
> >
> >a. Run a dual network with 1918 addresses
> >b. Just run an IPv6 network
> >
> >If people want to do a), do you really think
> >this would be simpler than b) ? Easier to manage
> >and operate? 
> 
> 	"10 million subscriber" scenario is nontrivial by nature.
> 
> 	for (a), 10.0.0.0/8 space only servers 17 million subscribers
> 	(theoretical maximum), and since address alloation 
> efficiency is not
> 	that good, you can't have a single NAT cloud.  you need to split
> 	customers into multiple NAT clouds (also I don't think 
> any single
> 	NAT box can take traffic from 10 million subscribers).
> 
> 	even for (b), I don't think single NAT-PT box (Alain, 
> forget about
> 	DNSSEC for now) can take traffic from 10 million 
> subscribers, so we
> 	at least need to split NAT-PT load to multiple NAT-PT 
> box somehow.
> 	(i assume the customers will want to use IPv4-only 
> yahoo.com.cn, right?)
> 
> 	so if we need to take a pick, it is between (a) IPv4 
> multiple NAT cloudm	and global IPv6, or (b) global IPv6 and 
> multiple NAT-PT.
> 
> itojun
> 
>