1 |
On 08/03/2013 15:40, Michael Mol wrote: |
2 |
>> IPv6 is wonderfully easy to use client-side and reasonably easy to plug |
3 |
>> > into an existing network (the routers mostly know what to do already). |
4 |
>> > The fun starts when you need to write an app that tracks and does range |
5 |
>> > allocations at ISP scale, all while keeping the PTRs in line too. Sadly |
6 |
>> > for me, my team works in that area and such a magic app is one of our |
7 |
>> > deliverables |
8 |
> My mouth is watering... |
9 |
> |
10 |
>> > |
11 |
>> > One day when I've climbed down off the walls and my fingernails have |
12 |
>> > grown back, I might be up to relating what it is taking to get that |
13 |
>> > done.... :-) |
14 |
> I don't suppose you knew I'm a huge IPv6 advocate, and travel around my |
15 |
> state giving free training sessions... |
16 |
> |
17 |
> I would absolutely love to hear about the problems you're facing. |
18 |
> Further, I'd love to help you get past them...and can put you in touch |
19 |
> with experts who might also be able to help. |
20 |
> |
21 |
> |
22 |
|
23 |
I too am a huge fan of IPv6. It will all at once fix two things: |
24 |
|
25 |
1. The craziness of trying to conserve IPv4 space |
26 |
2. NAT. Finally, a good solid techical reason to make NAT just go away |
27 |
and stay away. Permanently. Forever. |
28 |
|
29 |
Without getting into huge amounts of boring detail on the list, the |
30 |
major problem is finding software that can fulfil our backend needs. |
31 |
|
32 |
We're an ISP and we sell address space. We also allocate space |
33 |
internally to our own customers and need to keep accurate records as to |
34 |
who has what. There are some good packages on the market that function |
35 |
as a DNS backend where you enter the relevant IPv6 data and it takes |
36 |
care of the forward and reverse entries. We need something a bit more |
37 |
than that, we need the software to do block allocation. So it must not |
38 |
be a mere record of stuff, it must be the authoritative source of |
39 |
reference for that stuff, and it must understand the people involved and |
40 |
their roles (what they may allocate etc). |
41 |
|
42 |
The best package we found so far is one from Cricket Liu's company (I |
43 |
forget the app name for the moment, it was a while ago) but it doesn't |
44 |
do allocation and the backend db isn't really designed for that. |
45 |
|
46 |
And I nearly forgot to mention, I also need magic post-allocation hooks |
47 |
in the code. I'm determined to get to a point where unallocated blocks |
48 |
will be black-holed by the core routers. They must only start passing |
49 |
the traffic once a properly assigned block has been allocated and |
50 |
recorded as such. If you ever worked at ISP level, you will immediately |
51 |
understand why that is rather important :-) |
52 |
|
53 |
Speak later. |
54 |
|
55 |
|
56 |
-- |
57 |
Alan McKinnon |
58 |
alan.mckinnon@×××××.com |