[bisq-network/proposals] Investigate Oracle solutions (#253)

[chimp1984]

Oracles have been a underestimated (because over-hyped) feature. But I think it is a quite powerful concept which can add a lot of security and possibilities inside its trust limits.
Oraclize (now provable.xyz) has been working on oracle ideas based on the TLS Notary idea since many years and extended their solutions. I don't think that a commercial product would fit into Bisq, but re-using their concepts and run our own Oracles would be an interesting option.

There are basically 2 models [1]:
- Using Amazon webservice infrastructure 
- Using Google's Android remote attestation technology

I guess the third option using Ledger wallet is too limited for any use case of Bisq.

The trust/security limitations are to trust Amazon and/or Google to not compromise their infrastructure to hack the setup. If we can use both models it would gain even higher trust/security as both need to be compromised (depends on use case).
I think this is good enough for quite a few use cases in Bisq. Basically all those which are based on bonded roles, like Seed nodes, XMR-proof service, price providers, notification relay,...

I would suggest that we start with the easier solution first (likely the AWS) and get a full PoC running to gather experience. If that works out well we could try to get the Google/Android model working as well and start to use Oracles for the most critical nodes.

I assume it will require some deterministic build system like Gitian Build [2]. So that will be a considerable challenge as well.

Any dev who is experienced with the required technologies/operational skill set interested to work on that?

I think we should break it down to a first investigation/research package to see how feasible it is for Bisq's use cases. If that is successful get a PoC working. After that get one node running as Oracle. We might also use the services offered by provable.xyz for basic research.

Beside the existing use cases for infrastructure nodes, there might be new use cases which we have not thought of as the feature was not considered to be feasible. I guess for new types of trade protocols it might be an interesting tool as well.
Also for hosted Bisq instances with a trust minimized setup it might be interesting....

[1] https://docs.provable.xyz/#security-deep-dive
[2] https://github.com/devrandom/gitian-builder

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/bisq-network/proposals/issues/253
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bisq.network/pipermail/bisq-github/attachments/20200823/725bbb85/attachment.html>