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

Re: Discussion of the Home/SOHO environment



On Jan 3, 2008, at 20:47, Brian E Carpenter wrote:
On 2008-01-04 16:00, james woodyatt wrote:
On Jan 3, 2008, at 18:25, David Miles wrote:

While IPv6 supports it, it would be resource intensive for edge routers to support transitioning from one prefix to another with multiple prefixes simultaneously active.
Actually, no it isn't. The IPv6-enabled Internet gateway my employer sells today does this. Every time the public IPv4 address changes, the IPv6 6to4 interior prefix moves.

That's not quite the same as running with several simultaneous
prefixes, though. I agree that this ought to work painlessly too,
since it's a design feature of IPv6, but the CPE will have to
actually support it, and so will all the consumer devices.

It's basically the same thing if you count the private IPv4 subnet and the IPv6 link-local prefix along with the globally reachable IPv6 prefix.

Looked at this way, Apple's home gateway products have supported multiple Internet subnet/prefixes for several generations now. We went with IPv6 link-local a long time ago, mainly because multihoming in IPv4 is a huge pain in the neck, and just about every machine we sell can (and frequently does) operate multihomed on IEEE 802.3 and IEEE 802.11 at the same time. Having more than one IPv6 prefix on the same interface is just a wrinkle on the more general problem of having more than one Internet protocol address and/or multiple network interfaces.

From the perspective of the social contract to which service providers and consumers subscribe, there is nothing really new here. Consumers are already used to having their operating systems (and consumer devices) manage the multihoming problem for them, and they really shouldn't be any more inconvenienced by having their IPv6 prefix change out from under them than they are by having their IPv4 address changed out from under them. ISP's are doing the latter all the time, and I have no reason to believe they won't find reasons to do the former as well— if, for no other reason than to protect their carefully traffic-engineered networks from growing congested because their customers attempt to run unauthorized servers.


--
james woodyatt <jhw@apple.com>
member of technical staff, communications engineering