Module 2: The Commun Axiom Software Package
Written by the DSH coordination team* in collaboration with Luc Véronneau, Solutions Architect
Une version française de ce billet a été publiée ici.
This post is the second in a short series presenting each of the three components of the solution proposed by the Centre for Interdisciplinary Research on Montréal (CIRM) to the partners of the Data for Society Hub (DSH), i.e., Centraide of Greater Montréal, the Montréal Regional Director of Public Health, the Department of Diversity and Social Inclusion — City of Montréal, and the Montréal — Métropole en santé. These modules address challenges in terms of stakeholder engagement, identification and production of datasets, types of analysis, and ways of sharing data that respect privacy principles. The related blog posts will outline the means by which we intend to experiment in order to achieve the goals of the DSH community; they present a specific state of the project and solution, bound to evolve. You are invited to take part in the discussion, to share your thoughts and comments, and to contact us if you wish to be involved in this process, which will run until 2025. Read the previous blog post.
The Data for Society Hub’s (DSH) data-sharing objectives hinge on the development of a multi-functional sharing platform: the Commun Axiom software package. This software package is designed to meet our partners’ needs in matters of data indexing, transformation, sharing, and governance so as to provide a better understanding of Montréal’s communities.
Data sharing guidelines
The DSH Partners Committee has met on a monthly basis since Spring 2020 to co-develop the parameters for the proposed technological solution. Discussions between university, public, and para-public sector representatives have highlighted the following four fundamental issues:
- security — automatic cleansing of the data provided by organizations through a locally performed extract, transform, load (ETL) process;
- indexing — a repository of available datasets from which to submit access requests and foster collaborations around common interests;
- intellectual property and impact measurement — ability to view a complete history of the origin and use of each dataset shared on the platform;
- lightweight infrastructure — hosted on a Raspberry Pi nanocomputer that operates at low cost and with reduced power consumption.
Based on these guidelines, the Commun Axiom software package offers three main applications: Let’s Agree, Commons, and Ledger.
Let’s Agree: Establishing the terms for data sharing
The sensitive nature of certain data complicates their dissemination and accessibility. The application Let’s Agree allows users who own or receive data to negotiate the related anonymity and security criteria. This involves establishing the transformations to be performed before sharing a dataset, such as replacing specific postal codes with larger geographic areas or swapping birth dates between multiple individuals.
All these conditions or “metaconstraints” form a final agreement protocol that can be signed digitally. The signing of the protocol generates a Ledger entry (see below) to seal the agreement and formalize the split between the sending and receiving parties.
Commons: The hub of data pooling
The Commons application ensures the security required for the transfer of data between users. It allows for the automatic preprocessing of data according to the criteria set by the agreement protocol. In this way, data owners ensure that the ETL process occurs prior to the transmission. This is achieved through asymmetric end-to-end encryption to prevent the data from being read by a third party.
The Commons application includes a portfolio that lists and describes available datasets to facilitate research and discovery. The orchestrator relays access requests from users to the owners, while notifications are issued for any new dataset made available. After a download, the local storage function keeps the transformed versions of the data within an encrypted space.
Ledger: Tracking data as it is shared
The Ledger application is based on blockchain technology for the purpose of archiving all transactions made in Commun Axiom. It allows in particular for the registration of published datasets and the signing of user agreements.
It also ensures that the authors and current owners of the datasets are traceable, even after various overlaps and combinations. This has two advantages in terms of intellectual property and impact measurement: 1) to credit and acknowledge the work of data-providing organizations; 2) to allow data-providing organizations to identify the users and scope of the content they make available.
Upcoming steps in the development of Commun Axiom
The technology infrastructure team is currently working to implement the data sharing structure. The goal is to have the Commons application up and running, though not completed, by the end of 2021. It will then be possible to test the data transfer and transformation tools within a prototype version. Developers will then experiment with the rest of the features, including the management of metaconstraints, through the creation of Let’s Agree.
Eventually, we want to decentralize the hosting of the data sharing infrastructure, which is the largest component of the platform. Our teams will call on volunteer organizations with sufficient resources in their servers to set aside the equivalent of a workstation acting as a relay for dataset transfers. The maintenance of user accounts, authentication, and permissions will remain centralized within a standalone monolithic application.
The Commun Axiom software package is the brainchild of Luc Véronneau (Véronneau Techno. Conseil inc) made possible by the DSH technology infrastructure and development team. The team is currently composed of Luc Véronneau (Solutions Architect), Mohammad Moshawrab, Arina Vincter, and Ada Yetis (Developers).
*Copywriting: Julie Levasseur; content editing: Karolyne Arseneault and Luc Véronneau; translation: Angelina Mazza; translation editing: Julie Levasseur and Luc Véronneau
