Designing IPv6 network guidelines?

[itojun@iijlab.net]

> * Is routing between 6bone and production addresses in use/reliable yet?
>Does 6to4 work in practise?  Are there 6to4 prefixes being announced to
>the net at large? (Looking for a setup where IPv6 enabled LAN segments
>would connect via ipv6 enabled gateways to a central IPv6 router, from
>which would provide the connectivity)

	to stabilize, try to remove tunnels and move to IPv6 over leased
	line (or IPv4/v6 dual stack connectivity over leased line).  it works
	for me.

>1) Some RFC recommended using site-local addresses for point to point
>links.  Won't this break traceroute?

	you don't even need site-locals for point to point links.  they just
	work fine with link-local address.  all routing protocols should run
	fine with p2p with link-local address only.

>3) RIRs are allocating /35 prefixes for ISPs and the like.  Organisations
>get a /48 from them.
>Now, let's take an example about how academic networks in North Europe are
>built (rough estimate):
> * a country-wide operator entitled to a /35
>  * university or the like which would get (some?) /48's
>   * depertment which would get a /48 [?] or multiple /64
>    * lab or LAN's which might get multiple /64's
>
>I read from www.jp.ipv6forum.com that ISP's should allocate /48's to
>households etc.  I fail to see how /35 wouldn't run out very quickly (It
>_is_ only 8096 addresses) with this practice.  Also, as demonstrated from
>the above, if you have several organisatorial levels, it might be
>difficult to design a clean network if only elements you can use are /35,
>/48 and /64.

	if /35 runs out, country-wide operator can ask for more to RIR.
	/35 was assigned from RIR to facilitate slow-start.  country-wide
	operator should be able to grow into /29 sub TLA without renumber,
	then become a proper TLA (/16) - this is what I understand at this
	moment.

	if you use something other than /48, you will make renumber harder
	for customers (universities).

itojun