On 10/7/11 2:41 PM, Lorenzo Colitti wrote:

    Seems like we'd like to be able to introduce IPv6 support without
    NATs into such a home network without breaking IPv4.


This approach doesn't break IPv4, but it will work regardless of whether
IPv4 is working or not. So if/when IPv4 goes away you can plug any port
into any other port, even in topologies that IPv4+NAT can't handle, and
it will still work.

I don't think "doesn't work if IPv4 doesn't work" should be a goal.

    You seem to focus on a clean slate approach.


More than clean slate it's ships in the night.

So if we take a topology below (if it doesn't display correctly, look at figure 7 in http://tools.ietf.org/html/draft-chakrabarti-homenet-prefix-alloc-00)

                   +------+--------+                    \
                   |     IPv6      |                     \
                   | Customer Edge |                      \
                   |    Router 1   |                      |
                   +------+--------+      +------------+  |
       Network A          |               |            |  |
     +-------------+------+-------+-------+-----+      |  |
     |             |      |       |             |      |  |
+----+-----+ +-----+----+ |  +----+-----+ +-----+----+ |  |
|IPv6 Host | |IPv6 Host | |  |   IPv6   | |IPv6 Host | |  |
|          | |          | |  |  Router 2| |          | |  |
+----------+ +----------+ |  +----+-----+ +----------+ |  |
                          |       |                    |  |
                          |       +-------------+      |  |
                          |       | Network B   |      |  |
                          |       |             |      |  |
                          |  +----+-----+ +-----+----+ |  |
                          |  |   IPv6   | |IPv6 Host | |  |
                          |  |  Router 3| |          | |  |
                          |  +----+-----+ +----------+ |  |
                          |       |                    |  |
                          |       +--------------------+  |
                          |         Network C/A           |
                   +------+--------+                      | End-User
                   |     IPv6      |                      | networks
                   |    Router 4   |                      |
                   +------+--------+                      |
       Network D          |                               |
     +-------------+------+--------+---------+            |
     |             |               |         |            |
+----+-----+ +-----+----+     +----+-----+ +-+------+-+   |
|IPv6 Host | |IPv6 Host |     | IPv6 Host| |IPv6 Host |   |
|          | |          |     |          | |          |   /
+----------+ +----------+     +----------+ +----------+  /


If this was an IPv4/NAT network it wouldn't reliably work, since there is a NAT loop involving R3 and R2. Thus in IPv4 the link out of the bottom of R3 would actually not exist.

I think what you are saying is that it is fine if the deployment of IPv6 encourages folks to create the R3/R2 loop above, even if that makes IPv4 fail.

I find such an approach rather unrealistic; it might be reasonable in greenfield IPv6-only home networks but I would never recommend that as a way to move current IPv4 home networks forward.

I thought we wanted the solution to apply to the average Joe, and not just experts like ourselves.

Regards,
    Erik
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to