[panda-users] Language & Format Agnostic Data Schema

[Kenneth Adam Miller]

Hello,

I was thinking there are quite a few dynamic analysis utilities out there.
I think the two most flexible and prominent are Intel PIN and PANDA,
because they work on most OS's and architectures.

Assuming you had different analysis requirements & goals, you might have a
pin tool and a PANDA tool that each capture very similar data. Having
exterior tools that can process the data captured in agnostic of which
analysis tool was used is ideal, especially if there were something that
could also cross language boundaries really well (e.g., you have some post
dynamic analysis data tools that are written in python and others that
might be written in java, c, Ocaml or whatever).

Piqi would facilitate this extremely well; you could create a data schema
specification that would allow any type of data gathered by PIN or PANDA to
be serialized to a format (JSON, XML, protocol buffers), and it also
generates language interfaces (OCaml, Erlang, Python, C & Java) for parsing
the serialized data back out.

My question is, has anybody already done something in this direction (any
kind of data schema spec in protocol buffers, piqi, capn'proto, ect) in
order that I don't reproduce any work?

Also, if no one has given this any thought, what is the interest level in
such an endeavor?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/panda-users/attachments/20141207/d3b524ca/attachment.htm