[Openpacket-devel] Discussion about standardization of anonymization profiles

[Christoph M. <ma...@tm...>]

Hi OpenPacket Developers,

my name is Christoph Mayer and I am the developer of the PktAnon network 
trace anonymization tool (www.tm.uka.de/pktanon). We have had already 
some discussion about PktAnon on this list mid last year. Now, I want to 
push this further and hope to start a discussion about the possibility 
to bring PktAnon into OpenPacket.

I have some points that I want to throw in and hope that you kick in to 
the discussion on them:

1. Currently there is no standarized anonymization on OpenPacket or 
other trace archives. In my opinion there should be a standardized 
anonymization process. Therefore a defined set of anonymization profiles 
should be worked on that users can easily use. PktAnon, e.g., uses 
XML-based profiles that define how a trace is anonymized. Having 
community-standardized profiles (de-facto standards) of such profiles 
would in my opinion be a great benefit. Users (or the server-side) could 
anonymize using a default profile and the trace published with 
meta-information which profile was used. PktAnon currently contains 
three default profiles which we came up. What do you think of these 
profiles? Are they adequate? Is it in your opinion at all possible to 
define such a set of standard profiles for a system like OpenPacket?

2. How should a system for anonymization in OpenPacket look like? Should 
users anonymize the data, or the server system? What is adequate and 
safe for the user? What is adequate and safe for the server?

3. How can validation of a tool for anonymization look like? Code 
reviews, statistical tests on the anonymized traces? Should there be 
some kind of validation directly after the anonymization of the trace?

I would be very happy to hear about your opinions! I think OpenPacket is 
the right way for a packet repository to work. Privacy issues are in my 
opinion the reason why there are still not many people submitting their 
traces. Having a standardized process here would help for sure!

Best Regards,
Chris
-- 
Dipl.-Inform. Christoph P. Mayer
Institute of Telematics, University of Karlsruhe (TH)
Zirkel 2, 76128 Karlsruhe, Germany
Phone: +49 721 608 6415, Email: ma...@tm...
Web: http://www.tm.uka.de/~mayer/