[IPv6crawler-wg] IPv6 Crawling project info

[Olivier MJ Crepin-Leblond]

 Hello Christian,

thank you very much for your kind reply. I really value your feedback
and look forward to receiving feedback (and Frequently Asked Questions
to populate the online FAQ) from others on the distro above.

Le 02/09/2010 10:25, Christian de Larrinaga a écrit :
>
>
>  
> 2020media are not members of LINX nor LONAP. So you are having to organise direct peers I suppose? As the installation is in Telehouse might it not be worth contacting John Souter to see if they will sponsor a port through LINX and give LONAP a nudge? If 2020 don't need to peer directly perhaps the project / ISOC England could.  I'd be happy to negotiate this on your behalf. 

I'll need to discuss this with 2020Media. I am unsure as to what peering
directly would entice and will await feedback from them before deciding
on this.

> In 3.3.1
> It is good to see data will come through for www. dns. mx. and even popping the downstream servers. It would also be interesting to see use of sip. and possibly naptr as these are critical for voice services which are increasingly dependent on IP. 

The trouble we're hitting is how to test for it without generating too
much unusual traffic which would trigger firewalls. Also - it appears
that use of naptr & sip is not so widespread when compared to the other
services.

> In 3.3.4.1
> In testing for IPv6 connectivity, and with no agreed standards or methodology, we have found that the results we have gathered are somehow subjective and that more accurate trends are more likely to appear as more results are gathered over time.
>
> Yes this is not unexpected. It might be useful to add a process to get this to happen (help standardise methodologies). It is worth talking to NANOG and UKNOF both to let people know about the project but also to get feedback and input. 

Would you recommend that we speak to UKNOF before validating our
results? I know that the chaps at 2020Media are regular attendees at
their meetings.

> In 3.5.3 Mistakes made
> I am very impressed you have managed to find a way through the maze. It was clear after the domain spring clean project the chapter worked on in 2000 that DNS suffers entropy. 

All credits go to Dr. Al Ansary and his team for this... and we're
learning about more potential mistakes every day. I've now come across
DNS records which contain commas! I didn't even know BIND could work
these out...

> In 3.5.4 Future work
> In its current status, the project has laid the foundation for wide-scale crawling and we are sitting on top of a mine of valuable data. The project time and resources were sufficient to do primary analysis of this valuable data. The logical next step is to start a sequel project to curate, filter and analyze this data.
>
> Might it be helpful to have an API so that other projects can interface to the data directly?

The API exists, it just needs to be explained, which is what I'll try
and do in the second part of the report. Effectively, the Web site and
its click-on filters generate an automated query with the back end
running on port 4444 of the machine, and which responds to SQL queries.
So you can format SQL queries in an HTML line and get results accordingly.

> I've taken a brief glance over the website at 212.124.204.162 and www.ipv6matrix.com and it is a great start. 
> I think this is brilliant Olivier. Well done to you and the partners you enthused to get involved. This has clearly been a lot of work. 
>

Thank you. ISOC's feedback so far points to the same direction, although
they have asked me to populate the FAQ part of the Web site, which is
what I am doing every time I can think of a question to put in there.
Any other suggestions?

Kind regards,

Olivier

-- 
Olivier MJ Crépin-Leblond, PhD
http://www.gih.com/ocl.html

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://salsa.gih.co.uk/pipermail/ipv6crawler-wg/attachments/20100908/2b4dc81c/attachment.html>