Hacker News new | ask | show | jobs
by michaelt 4039 days ago
Well, presumably if blocks of IP addresses get radically smaller, routing tables will get large enough that not every system can handle them. Given how sluggish people are about upgrading to IPv6 people obviously aren't upgrading often, so there must be a lot of legacy equipment out there.

So it's not like I can just buy an one or two IP addresses, like I could in a liquid marketplace.

2 comments

If you are starting a new cloud provider and offering something like containers (say LXD), you have a big problem on your hands. IPv6 and v4 aren't really interchangeable from an end user perspective. Also, techniques are NAT64 are shocking when you look at them in detail (my jaw dropped at least ... not all protocols can be supported with NAT64, proxies need to be stateful etc.) We are in for a mess.
Weren't there already problems last year because the BGP tables got too big for some ISP equipment to handle? http://www.zdnet.com/article/internet-hiccups-today-youre-no...
Heaven forbid you use a router that was actually designed to be a router, rather than a switch with a henky supervisor card thrown into it.

Also help us if the routing manufacturers actually put more than 2GB of memory on the routing engine.