On 12/19/17, 8:50 PM, "NANOG on behalf of Owen DeLong" <nanog-bounces@nanog.org on behalf of owen@delong.com> wrote:
On Dec 19, 2017, at 07:39 , Livingood, Jason <Jason_Livingood@comcast.com> wrote:
On 12/18/17, 2:36 PM, "NANOG on behalf of Harald Koch" <nanog-bounces@nanog.org on behalf of chk@pobox.com> wrote:
They could use IPv6. I mean, if the mobile phone companies can figure it out, surely an ISP can...
Except for cases when it is impossible or impractical to update software on a great number of legacy devices…
JL
Yeah, in those cases, they should use IPv6 + NAT64 or similar mechanism.
I’m a fan of IPv6-only plus translation, but not in this case. If I have a functioning management network that’s mostly in IPv6 and partly in rfc1918 space (or even squatted space), I don’t get much out of NAT64. Renumbering the servers that actually touch/manage devices gets, what, a /29 of IPv4 addresses? Better to focus on evolving to whatever will replace those legacy devices. Lee
Owen