scitt-community / scitt-api-emulator

SCITT API Emulator
https://scitt-community.github.io/scitt-api-emulator
MIT License
10 stars 7 forks source link

SCITT API Interoperability Client

This repository contains the source code for the SCITT API interoperability client and sample emulator.

It is meant to allow experimenting with SCITT APIs and formats and proving interoperability of implementations.

Note the SCITT standards are not yet fully published and are subject to change. This repository aims to keep up with changes to the WG output as faithfully as possible but in the event of inconsistencies between this and the IETF WG documents, the IETF documents are primary.

Prerequisites

The emulator assumes a Linux environment with Python 3.10 or higher. On Ubuntu, run the following to install Python:

sudo apt install python3.10-venv

Clone the Emulator

  1. Clone the scitt-api-emulator repository and change into the scitt-api-emulator folder:

    git clone https://github.com/scitt-community/scitt-api-emulator.git
  2. Move into the emulator director to utilize the local commands

    cd scitt-api-emulator

Optional Dependencies

If you want to use conda, first install it:

You can get things setup with the following:

conda env create -f environment.yml
conda activate scitt

Start the Proxy Server

The proxy server supports 2 options currently:

Note: the emulator is for experimentation only and not recommended for production use.

Start a Fake Emulated SCITT Service

  1. Start the service, under the /workspace directory, using CCF

    ./scitt-emulator.sh server --workspace workspace/ --tree-alg CCF
  2. The server is running at http://localhost:8000/ and uses the workspace/ folder to store the service parameters and service state
    Note: The default port is 8000 but can be changed with the --port argument.

  3. Start another shell to run the test scripts, leaving the above shell for diagnostic output

  4. Skip to Create Signed Claims

Start an RKVST SCITT Proxy Service

  1. Start the service, under the /workspace directory, using RKVST
    The default port is 8000 but can be changed with the --port argument.

    ./scitt-emulator.sh server --workspace workspace/ --tree-alg RKVST

Executing Commands

The service has the following REST API:

Note: The submit-claim and retrieve-claim commands use the default service URL http://127.0.0.1:8000 which can be changed with the --url argument. They can be used with the built-in server or an external service implementation.

Create Signed Claims

  1. Create a signed json claim with the payload: {"sun": "yellow"}, saving the formatted output to claim.cose

    ./scitt-emulator.sh client create-claim \
        --content-type application/json \
        --subject 'solar' \
        --payload '{"sun": "yellow"}' \
        --out claim.cose

    _Note: The emulator generates an ad-hoc key pair to sign the claim if --issuer and --public-key-pem are not given. See Registration Policies docs for more deatiled examples_

  2. View the signed claim by uploading claim.cose to one of the CBOR or COSE Debugging Tools

Submit Claims and Retrieve Receipts

  1. Submit the Signed Claim

    ./scitt-emulator.sh client submit-claim \
        --claim claim.cose \
        --out claim.receipt.cbor
  2. View the response, noting the Entry ID value

    Claim Registered:
        json:     {'entryId': '1'}
        Entry ID: 1
        Receipt:  ./claim.receipt.cbor
  3. Create an environment variable (ENTRY_ID) for the value of entryId above:

    ENTRY_ID=<entryId>

Retrieve Statements and Receipts

  1. Retrieve the Statement, based on the ENTRY_ID set from the submit-claim command above

    ./scitt-emulator.sh client retrieve-claim \
    --entry-id $ENTRY_ID \
    --out claim.cose

    retrieve-claim sends the following request:
    GET /entries/<entry_id> to retrieve the claim.

  2. Retrieve the Receipt, based on the ENTRY_ID set from the submit-claim command above

    ./scitt-emulator.sh client retrieve-receipt \
        --entry-id $ENTRY_ID \
        --out receipt.cbor

    retrieve-receipt sends the following request:
    GET /entries/<entry_id>/receipt to retrieve the receipt.

Validate Receipts

./scitt-emulator.sh client verify-receipt \
    --claim claim.cose \
    --receipt claim.receipt.cbor \
    --service-parameters workspace/service_parameters.json

The verify-receipt command verifies a SCITT receipt given a SCITT claim and a service parameters file. This command can be used to verify receipts generated by other implementations.

The workspace/service_parameters.json file gets created when starting a service using ./scitt-emulator.sh server. The format of this file is not standardized and is currently:

{
    "serviceId": "emulator",
    "treeAlgorithm": "CCF",
    "signatureAlgorithm": "ES256",
    "insertPolicy": "*",
    "serviceCertificate": "-----BEGIN CERTIFICATE-----..."
}

"signatureAlgorithm" and "serviceCertificate" are additional parameters specific to the CCF tree algorithm.

To view the file:

cat workspace/service_parameters.json | jq

COSE and CBOR Debugging

The following websites can be used to inspect COSE and CBOR files:

Code Structure

scitt_emulator/scitt.py contains the core SCITT algorithms that are agnostic of a specific tree algorithm.

scitt_emulator/ccf.py is the implementation of the CCF tree algorithm. For each claim, a receipt is generated using a fake but valid Merkle tree that is independent of other submitted claims. A real CCF service would maintain a single Merkle tree covering all submitted claims and auxiliary entries.

scitt_emulator/rkvst.py is a simple REST proxy that takes SCITT standard API calls and routes them through to the RKVST production SaaS service. Each claim is stored in a Merkle tree underpinning a Quorum blockchain and receipts contain valid, verifiable inclusion proofs for the claim in that Merkle proof. More docs on receipts here.

scitt_emulator/server.py is a simple Flask server that acts as a SCITT transparency service.

scitt_emulator/client.py is a CLI that supports creating claims, submitting claims to and retrieving receipts from the server, and verifying receipts.

In order to add a new tree algorithm, a file like scitt_emulator/ccf.py must be created and the containing class be added in scitt_emulator/tree_algs.py.

Run Tests

./run-tests.sh

Contributing

This project welcomes contributions and suggestions. Please see the Contribution guidelines.