Cost of traffic data access?

[Ian Mason]

On 14 Aug 2010, at 16:22, Roland Perry wrote:

> In article <4C66A0A6.3030508 at zen.co.uk>, Peter Fairbrother  
> <zenadsl6186 at zen.co.uk> writes
>
>> Roland, do you have even a very rough range for the cost? Thanks,
>
> There are perhaps two extremes punted around long ago, where it's  
> been alleged some reverse-DQ requests cost £100 each, versus some  
> requesters only prepared to pay £15/hr for proven effort sorting  
> out the answers. But where we are today, I don't know.
> -- 
> Roland Perry
>

£15 a hour wouldn't even represent cost recovery of salary, let alone  
overheads, for ANY engineer I've employed in the last 10 years. A  
realistic minimum charge would be in the order of £35/hour just on a  
cost recovery basis for low level engineering staff extending to £100/ 
hour for senior staff on the same basis.

As to the particular data Peter is asking about NO sane ISP keeps  
those records, I doubt any insane one does either. To do so, even as  
a one off, would incur significant engineering effort and involve  
setup costs in the thousands if the existing ISP network was suitably  
structured to make it a possibility. For many ISPs it might not be  
possible to do without network re-engineering across an entire  
network potentially involving effort in the six figure region.

The nearest to what Peter's asking for that anybody routinely gathers  
is netflow data and that is analysed and discarded quite quickly.  
Also it is strictly statistical rather than accurate, in the sense  
that there is no guarantee that events have ALL been logged. Even  
when recorded netflow data would usually be at a much coarser  
granularity than Peter's envisioning - e.g. destination AS number  
rather than destination address.

Ian