Originally Posted by Spudstr

The only thing ipv6 has done is make network engineers job a bit harder. Mainly because of the 128bit address space. Its simply larger than ipv4 address space.

Lets put it this way. cisco and juniper just recently released router gear that supports 1mil-4million prefixes.

Lets look at todays data http://bgp.potaroo.net/as2.0/bgp-active.html

There are 330k ipv4 address prefixes being announced across the net. Get a couple providers sending you full routes and your maxing out your shiny 10-20k route processor card. Sucks doesn't it?

So lets put this into perspective, a /24 is the smallest netblock that is allowed to be announced to the world. Everyone knows how big a /24 is. There are 172k /24's being announced independently of other network address space. Thats a LOT.

Now lets look at IPv6. Given that ARIN is allocating out /32's worth of IPv6 address space. Which is in essence every every single person the equivalent of 4billion IPs. Yes they are giving out the entire ipv4 routing table equivalent to everyone who wants it as the minimum.

Now the SMALLEST IPv6 address prefix that you can announce as the current "standard" is a /48. There are 64,000 /48's inside of a /32. So yes this means. it will take less than 5 networks announcing every single /48's individually to fill the same amount of router memory as todays IPv4 network space. This isn't even taking consideration of the amount of memory each route uses since ipv4 is 32bit and ipv6 is 128bit, so each route entery is going to take more memory.

So with this being said, the routers that power the backbones are going to have a hard time keeping up with the prefixes, hell they currently are having this problem. Let alone your servers firewall filtering stuff. You have not seen anything yet.