BBC does IPv6 ;) (Was: large multi-site enterprises and PI

Sun Nov 28 02:21:42 UTC 2004

At 12:25 PM 11/27/2004, Iljitsch van Beijnum wrote:

>On 27-nov-04, at 17:43, Paul Vixie wrote:
>
>>those of us who prefer static assignment + dhcp6 over EUI64 find a /64 to
>>be an obscene waste of address space on a per-lan (or per-vlan) basis, but
>>sadly there are already some cool wireless gadgets whose idea of ipv6 does
>>not include either static or dhcp6 addressing, so there's some tension.
>
>>my house has four vlans (core, family, guest colo, and wireless) -- so in
>>the cisco home networking model i'd need a /62 for fewer than 50 hosts?
>
>While IPv6 is still IP, it's not just IPv4 with bigger addresses. We have 
>128 bits, so we should make good use of them. One way to do this is to 
>make all subnets and 99% of end-user assignements the same size. Yes, this 
>wastes bits, but the bits are there anyway so not wasting them really 
>doesn't buy you anything at this point. The advantage of having a fixed 
>/64 per subnet is that one size fits all: there is no need to worry about 
>the subnet size when designing the network, whatever happens, all hosts 
>that you'll ever want to put in this subnet will fit in it. Always having 
>a /48 has a similar benefit: if you ever need to renumber, you only need 
>to do a search and replace on the top 48 bits, the internal addressing 
>structure can remain the same.
>
>In a parallel universe IPv6 could have 64 bit addresses, saving 16 bytes 
>of overhead per packet (so the additional overhead re IPv4 would only be 4 
>bytes rather than 20), but here in the universe we're all most familiar 
>with, this ship has sailed a long time ago.
>
>>i life fred's reasoning.  companies with size and qualifications like
>>cisco's should qualify for an ASN and for PI space.  all the world's not
>>a home-DSL or home-cable or isp-colo network.  routing shouldn't always
>>follow addressing.  we'll need to discover a workable equilibrium unless
>>we want to encourage NAT in IPv6 the same way we (passively) encouraged it
>>in IPv4.
>
>All I hear is how this company or that enterprise "should qualify" for PI 
>space. What I don't hear is what's going to happen when the routing tables 
>grow too large, or how to prevent this. I think just about anyone "should 
>qualify", but ONLY if there is some form of aggregation possible. PI in 
>IPv6 without aggregation would be a bigger mistake than all other IPv6 
>mistakes so far.

First we built routers for IPv4 and hoped they'd have enough memory and 
performance to handle the future.

When it looked like routers weren't going to scale, we pushed MPLS on the 
assumption that we had to use higher performance "switches" in the core, 
and that edge systems would need to pick the routes so that core systems 
didn't have to worry about routing packets. After all, there wasn't going 
to be any way for core routers to handle wire speed forwarding. Well, a 
bunch of companies proved that wrong.

Routing table growth size has been another doomsday statement. There'd be 
no way to build equipment with enough memory. Or the memory would cost too 
much. Or the lookups would be too slow.

Time and again the "end of world" predictions have been proven false. 
Memory prices in the mid 1990's were in the $35/MB range, and memory sizing 
was a serious concern. Today's memory prices and CPU power far exceed what 
we had to work with then.

I have to agree with Fred. Anyone who today qualifies for an AS number will 
need to be allowed a block in IPv6 space if we ever hope to have v6 survive.