Annotated edit history of
AAAAvsA6 version 4, including all changes.
View license author blame.
| Rev |
Author |
# |
Line |
| 4 |
AristotlePagaltzis |
1 |
For [IPv6] there are two types of [DNS] ResourceRecord, <tt>AAAA</tt> and <tt>A6</tt>. |
| 1 |
PerryLorier |
2 |
|
| 4 |
AristotlePagaltzis |
3 |
<tt>AAAA</tt> is the simple one -- it works just like an <tt>A</tt> record in [IPv4], but with more bits. |
| 1 |
PerryLorier |
4 |
|
| 4 |
AristotlePagaltzis |
5 |
<tt>A6</tt> supports fragments of [IPv6] addresses. You can say <tt>example.com</tt> has the fragment <tt>2002:c000:0201</tt>, <tt>engineering.example.com</tt> has "the example.com fragment" + <tt>:0001</tt>, and <tt>foo.engineering.example.com</tt> has "the engineering prefix" + <tt>:0000:0000:0000:0001</tt>. This means that when <tt>example.com</tt> renumbers, they change just the one ResourceRecord for <tt>example.com</tt> and all of their machines have been renumbered. |
| 1 |
PerryLorier |
6 |
|
| 4 |
AristotlePagaltzis |
7 |
Some people complain that <tt>A6</tt> is too complicated, and prone to problems: |
| |
|
8 |
* What is the [TTL] on the response if the fragments all have different lifetimes? |
| |
|
9 |
* What's the status of it if some of the entries (but not all of them) are signed with [DNSSEC]? |
| |
|
10 |
* What protection is there against [DoS]? |
| 1 |
PerryLorier |
11 |
|
| 4 |
AristotlePagaltzis |
12 |
The <tt>A6</tt> people say that it's necessary for networks that regularly change [IP] address (eg: networks that use a dynamic 6to4 address). |
| |
|
13 |
|
| |
|
14 |
<tt>A6</tt> in general is the newer standard, and <tt>AAAA</tt> is the older standard, however, there doesn't yet appear to be much momentum for <tt>A6</tt> support. |
| 1 |
PerryLorier |
15 |
|
| 2 |
CraigBox |
16 |
----- |
| 3 |
MikeBeattie |
17 |
CategoryDns |
