ORIGO

Contents

• Introduction
• Disclaimer
• Structure and Development
• Command Line Toolkit
• Tutorials
• Evaluation
• Limitations
• Citing

Introduction

ORIGO is a decentralized oracle that doesn't rely on the trusted hardware and two-party computation. It allows users to prove the provenance of TLS records from a specific API and, in the meanwhile, prove the dynamic content fulfills a value constraint as pre-defined in a public policy in zero-knowledge. It achieves confidentiality and integrity under certain honesty assumptions (e.g. honestly behaving proxy) which we define in the associated paper ORIGO. ORIGO relies on a proxy sitting between the client user and the server hosting the API. The TLS stack running on the server does not require any modifications, making ORIGO extendable to various existent APIs and web resources.

Disclaimer

ORIGO is a research project and its implementation has not undergone any security audit. Do not use ORIGO in a production system or use it to process sensitive confidential data.

Structure and Development

The structure of the project is as follows:

    ├── certs		# self-signed certificates for local development
    ├── commands	# command line toolkit logic
    ├── dependencies	# contains additional git submodules
    ├── docker		# containerization
    ├── docs		# documentation files
    ├── ledger_policy	# public policies defining API and constraint specifications
    ├── prover		# logic of http client to perform an API requests and prove policy compliance of response value
    ├── proxy		# tcp tunneling service for tls handshake and record layer traffic capturing and policy verification
    ├── server		# local https instance to serve sample traffic
    ├── transpiler	# policy to snark circuit generator translation

where we provide further details of sub-structures respectively in the prover, proxy, and server folders.

We quickly outline the key ideas behind the struture of the repository. For local development, developers can run the script certs/cert.sh to automatically generate self-signed root, server, and client certificates. The plan is to eventually seperate the folders server, proxy, and prover such that these folders can be executed individually on different machines for different deployment scenarios. The main idea behind the ORIGO repository structure is that policies, circuits, and other configurations should be managed outside the folders server, proxy, and prover. Only strictly necessary files are supposed to be copied into individual prover, proxy, and server folders such that these folders can be copied and executed on different machines. Folders to manage policies or snark circuits are kept in separate locations (e.g. dependencies, ledger_policy, transpiler, etc.). The commands folder maps modular functionalities from generating transcript data, parsing and extracting policy specific data, transpiling policies, compiling snark circuits, and generating proofs, to the ORIGO command line binary such that developers can easily play around with different computation sequences and configurations. The server folder and server-commands are only required ORIGO is executed entirely locally.

Command Line Toolkit

ORIGO is a command-line toolkit which allows users to (i) transpile policies into SNARK circuits, (ii) generate input data for policy compliant data provenance proofs from private APIs, and (iii) prove policy-compliant proofs in zero-knowledge. The high-level workflow of the ORIGO command-line toolkit is as follows:

1. (proxy) start the proxy service
2. (server) start the server service (optional, required for local testing)
3. (prover) transpile public policy into a circuit generator
4. (prover) collect circuit input data by
   • refreshing API credentials (optional, if local deployment tutorial is used)
   • request API defined in policy and post-process tls and record layer traffic
5. (prover) generate snark-circuit in arithmetic representation and compute witness data
6. (prover) compute setup and generate zkp
7. (proxy) postprocess captured traffic transcript and collect zkp public input
8. (proxy) verify zkp

Tutorials

We provide all details of how to execute ORIGO in different deployments in our tutorial guidelines. Before you start running the tutorials, please follow our installation instructions to correctly set up the repository and its dependencies.

Evaluation

To reproduce results provided in the research paper ORIGO, we provide an evaluation script evaluation.sh. The evaluation script can be executed by calling ./evaluation.sh in the root location of this repository after installing the repository with the ./installation.sh script as described here.

Limitations

• ORIGO currently supports off-chain zkp verification only and we plan to address on-chain zkp verification in the future.
• The constraint system currently supported by ORIGO public policies is very limited with value proofs of float greater than GT, float less than LT and string equality EQ and will be extended in the future.
• Adding flags for cmd toolkit input parameters.

Citing

We welcome you to cite our research paper if you are using ORIGO in academic research.

@inproceedings{2022origo,
    author = {},
    title = {ORIGO: Proving Provenance of Sensitive Data},
    year = {2022},
    publisher = {},
    booktitle = {},
    location = {},
    series = {}
}