[AusNOG] [LINK] [ISOC-AU-mems] Happy Birthday ... AARNet

Wed Mar 18 12:03:26 EST 2009

Inside an enduser's home network there's not a workable solution for  
coping with dynamically assigned prefixes from a provider.   The  
current combination of AutoConf and DHCPv6 combined with what's  
mandatory and what's not leads to this - you'll end up with PCs/Macs/ 
etc being very confused about what prefix to use and no connectivity  
at IPv6 level.

The PD features are available at the lowend from Cisco for LNS/LAC  
(eg. 72xx/73xx), but not the useful sized boxes necessary to do this  
at scale running SB train.

I'm very keen to have a workable, complete solution.  But it's not  
there yet despite what people keep claiming.  (If it was all done then  
I'd have it all running!)

MMC

On 18/03/2009, at 9:29 AM, Mark Smith wrote:

> Matthew Moyle-Croft wrote:
>> Not sure about your argument.
>> The main problem is that at the moment that the standards that  
>> deliver v6 broadband in a general sense are still all draft and, at  
>> the moment, don't quite work together.  (Feel free to tell me I'm  
>> wrong, but ONLY if you can actually send me a complete set of  
>> receipes to do it as at least one Broadband Forum member has told  
>> me it can't be done yet).  The main sticking point is prefix  
>> delegation and how that works in an end-user's network.
>
> What is the particular issue? Cisco routers support DHCP based  
> prefix delegation, including automatically configuring downstream  
> interfaces with the announced /48. It's a solved problem.
>
>
>> Once this is fixed and people stop having pissing matches about who  
>> wins (AutoConf, DHCPv6 etc) we'll be sweet and the CPE vendors can  
>> finish their work.
>
> DHCPv6 is considered the most likely deployment model. RAs (i.e.  
> AutoConf) are generally considered only to bootstrap basic IPv6  
> networking, for any thing else e.g. NTP config, DNS servers, prefix- 
> delegation etc. DHCPv6 is considered to be the advanced end-node  
> configuration protocol.
>
> There is some lobbying to "bloat" up RAs with these types of  
> options, however I think it is probably because people are a bit set  
> in their view that if you use DHCPv6 methods it means you have to  
> set up a separate server to the router. Of course, DHCP doesn't  
> require that, as plenty of ADSL CPE shows.
>
>> (Yes, you can do this in the simple case with static ranges etc,  
>> but that doesn't scale and doesn't work for normal people like my  
>> parents).
>
> What do you mean by doesn't scale and doesn't work for you parents?
>
>> MMC
>> dasmo wrote:
>>> Seems to me the problem is cash. ISPs won't eat it. Probably need  
>>> it  subsidised by the government. Some transit providers still  
>>> aren't ipv6  compatible, there's customer equiptment that needs to  
>>> be replaced and  there's no authority setting a deadline like the  
>>> digital tv system.  Plus, it's hard to explain the benefits to an  
>>> end user who will most  likely see the issues now rather than a  
>>> solution to an issue from the  future.
>>>
>>> Would be a better use of money than that stupid filter though.
>>>
>>> On 17/03/2009, at 16:57, Mark Smith <marksmith at adam.com.au> wrote:
>>>
>>>
>>>> Geoff Huston wrote:
>>>>
>>>>>> I specifically remember a slip connection to Hawaii growing  
>>>>>> from  1200
>>>>>> bps to 2400 bps preceeding the 56Kb frame relay connection.
>>>>>>
>>>> <snip>
>>>>
>>>>
>>>>> Yawn. That was years ago. On to today's problems. What are we  
>>>>> going  to
>>>>> do given that noone is doing anything remotely serious in IPv6  
>>>>> and  the
>>>>> crunch time of IPv4 address exhaustion is getting ever closer?  
>>>>> If we
>>>>> can't manage to preserve some level of protocol coherence across  
>>>>> the
>>>>> network in the coming few years then we may end up not much  
>>>>> better  off
>>>>> than the situation on 20 years ago. Or do we say goodbye to all  
>>>>> this
>>>>> end-to-end IP stuff and just run client sever over http and forget
>>>>> than anything else was ever possible?
>>>>>
>>>>>
>>>> I don't think Internet end-users are aware of the problem, let a  
>>>> alone
>>>> what it is, why its occurring, and what the consequences will be.  
>>>> They
>>>> haven't been told what it is, and they don't know to ask for it.
>>>>
>>>> That seems to me to be a marketing problem. We need to get the  
>>>> message
>>>> to the Internet end-user market that the Internet is heading  
>>>> towards a
>>>> wall, and needs to be upgraded. We need to explain in very  
>>>> simple  terms,
>>>> what the problem is - "The Internet is running out of phone  
>>>> numbers!"
>>>> (and then explain that public Internet addresses are like phone   
>>>> numbers)
>>>> - I think should be a simple enough place to start.
>>>>
>>>> Who should run this campaign? ISOC or the IPv6 Forum (or both) I   
>>>> reckon.
>>>>
>>>>
>>>> Regards,
>>>> Mark.
>>>> _______________________________________________
>>>> AusNOG mailing list
>>>> AusNOG at lists.ausnog.net
>>>> http://lists.ausnog.net/mailman/listinfo/ausnog
>>>>
>>> _______________________________________________
>>> AusNOG mailing list
>>> AusNOG at lists.ausnog.net
>>> http://lists.ausnog.net/mailman/listinfo/ausnog
>>>
>

-- 
Matthew Moyle-Croft Internode/Agile Peering and Core Networks
Level 5, 162 Grenfell Street, Adelaide, SA 5000 Australia
Email: mmc at internode.com.au    Web: http://www.on.net
Direct: +61-8-8228-2909		     Mobile: +61-419-900-366
Reception: +61-8-8228-2999        Fax: +61-8-8235-6909

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ausnog.net/pipermail/ausnog/attachments/20090318/6e3bb53f/attachment.html>