[Collisions] source prefix analysis of DITL traffic

[Warren Kumari]

On Aug 24, 2013, at 6:34 AM, Jim Reid <jim at rfc1035.com> wrote:

> On 23 Aug 2013, at 21:09, Warren Kumari <warren at kumari.net> wrote:
> 
>> I realize that we cannot put the current DITL data into this (because of the data sharing agreements), but if this sounds like something that would be very useful I can look into getting some quota donated and working with folk to get data contributed. 
> 
> Warren, that's a kind offer - thanks. I suspect though the layer-9+ problems would take longer to resolve than running the data through the new iron that OARC has (or soon will have) available.
> 

Oh, yeah, no doubt. This was more of a longer term thing, for the next time that someone wants to actually *do* something with the data.

Being able to run a test query across a few hundred TB in a couple of seconds lets you find all sorts of interesting stuff.

W


> It should be straightforward to look at the distributions of prefix counts across all the gTLDs. My guess is it'll take perhaps a day or theresabouts once someone gets hold of that summary data to write a script, feed its output into (say) a spreadsheet and spit out some results. A database might be overkill for this task IMO. Database people may well disagree.
> 
> Interisle's team would have done that analysis for all the new gTLDs if we'd more time and the hardware that was available hadn't already been fully committed to other parts of the project. For instance a power-law distribution analysis of all the prefix data for 2012 would have taken time away from crunching the 2013 DITL data that only became available ~1 week before our report's deadline. Chewing the 2013 DITL pcaps obviously had to take priority. So we did the best we could under the circumstances: a bit of sampling and presented those findings in the report.
> 

--
The above email is neither interesting or relevant, but at least it provided no new information.