[NCAP-Discuss] Privacy preserving query infrastructure (SMCQL)

[Eric Osterweil]

Hi Matt,

The setting we talked about is that we have DNS data providers who may want to participate in NCAP analyses, but have understandable concerns about sharing/exposing their data, and more general privacy concerns when they do.

<describing the paper (not my work, no ownership/proprietary feelings from me>
It proposes a secure/private federated database.  In regards to data providers wanting to protect their data, but still trying to share (i.e. DNS providers that have name collisions data but are understandably worried about having to expose it, while still trying to enable NCAP analyses).  As reported in the paper: SMCQL would facilitate the creation of a federated database (whereby different data owners would let some analyses be done on their data without letting it leave their networks).  The execution of queries would then be done in a way that matches the security/privacy wishes of the data holder.  The architecture (in Figure 1) describes untrusted end-users, semi trusted brokers, and secure data holders.
</describing the paper (not my work, no ownership/proprietary feelings from me>

Eric

> On May 22, 2019, at 1:29 PM, Matt Larson <matt.larson at icann.org> wrote:
> 
> Eric,
> 
> How are you thinking this work applies to NCAP?
> 
> Thanks,
> 
> Matt
> 
> 
>> On May 22, 2019, at 11:18 AM, Eric Osterweil <lists at osterweil.net <mailto:lists at osterweil.net>> wrote:
>> 
>> I wanted to bring to the group’s attention some work that I have become aware of, which seems to me to be worth considering for NCAP.  In 2017, a paper was published in VLDB (the very top-tier of large-scale data infrastructure research) about a tool for ``mutually distrustful parties’’ to allow untrusted clients to query and join datasets.  I (personally) found the work to be inspired.  It uses techniques like secure circuits and Oblivious RAM (ORAM) to implement a Secure Multiparty Computation (SMC) extensions/optimizations for SQL-like quires (called SMCQL).  The paper can be found here:
>> https://urldefense.proofpoint.com/v2/url?u=https-3A__dl.acm.org_citation.cfm-3Fid-3D3055334&d=DwIGaQ&c=FmY1u3PJp6wrcrwll3mSVzgfkbPSS6sJms7xcl4I5cM&r=xhCX8vQGcsNMzNMbgIokNle9Mpt6sQ45tM98iwh4H0w&m=3NzPJbEUwAeHXxHyJF9WYbEoPnbkjGDmpB-OQyDj7k8&s=HcSy267oTgllN1J2LCJjMAC2nFTSweg7qlXXJlRnzsM&e= <https://urldefense.proofpoint.com/v2/url?u=https-3A__dl.acm.org_citation.cfm-3Fid-3D3055334&d=DwIGaQ&c=FmY1u3PJp6wrcrwll3mSVzgfkbPSS6sJms7xcl4I5cM&r=xhCX8vQGcsNMzNMbgIokNle9Mpt6sQ45tM98iwh4H0w&m=3NzPJbEUwAeHXxHyJF9WYbEoPnbkjGDmpB-OQyDj7k8&s=HcSy267oTgllN1J2LCJjMAC2nFTSweg7qlXXJlRnzsM&e=>
>> 
>> I have reached out to the authors and begun reaching out to others that I know who work in this area for other references.  Ahead of that, however, I thought I would share early-and-often with the group, here.  I am happy to share any more of my thoughts.
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/ncap-discuss/attachments/20190522/3a6d6822/attachment-0001.html>