Dynamo gravitated toward the data sharing and data availability problem that was mentioned by multiple panelists during the hack-a-thon's kickoff. While other participants may be focusing on (really smart) algorithms, Dynamo wanted to leverage its cross-platform, cross-technology team to create a broad-reaching Data Platform built on multiple Blockchains, ElasticSearch, and modern web technologies.
What it does
Dynamo focused on addressing the following problems:
- Creating a Trusted, yet Trustless, Data Platform that can act as the source of truth
- Providing mechanisms by which agencies, organizations, and individuals can contribute to the Data Platform
- Exposing hooks to allow for data ingesting, processing, tagging, and scoring to classify Data better for algorithmic use
- Allowing Open Access to permissioned data, and providing the tools to allow participants to share
- Architecting a scalable and resilient Data Platform that can be rebuilt with a high guarantee of data integrity
The above Platform required implementing, integrating, and coding with a number of different Open Source technologies. Additionally, the team created two real world scenarios that showcases how the Platform could apply to human trafficking.
Trusted Data in Human Trafficking Workflow
Scenario: Dynamo has setup multiple Hyperledger channels representing the private data networks of three organizations: Department of Homeland Security, Global Fund to End Modern Slavery, Fairfax County Police Department, as well as a globally available data network built on Ethereum.
An organization's data is private until they decide to publish data to the global Ethereum network. In this case, local law enforcement may decide to mark a person in their system as missing, thus promoting their data globally.
The Ethereum network represents the System of Record and creates a chain that will:
- Log / audit the promotion of data.
- Kick off an indexing job into ElasticSearch.
- Create a unique IPFS Address that can be used to correlate data locally, globally, and within ElasticSearch.
Dynamo was originally going to use Amazon Lambdas to perform the data processing and indexing, but due to the distributed nature of IPFS and the lack of support on Lambdas, the indexing work had to be performed on a physical instance.
Once the data is promoted to the Global Data Network, the UI / UX representing the list of missing persons will be updated and show that the data has been record. Because data in the global domain is in a Blockchain network, the data will be globally distributed and could be made available to devices on the edge. Depending on the data provided during the promotion of the data, biometric data such as fingerprints, facial data, etc. would be made available for on-demand processing.
Dynamo's team had originally created biometric focused screens for a registration system, but pivoted focus based on the panel discussion. Screens of the initial proof of concept can be found here.
How we built it
Dynamo's solution includes the following technical deployments:
- 3x IBM's HyperLedger deployed in Amazon Web Services (AWS) 1, 2, 3
- Ethereum deployed in AWS
- AWS ElasticSearch
- Distributed Application (Ethereum) to interact locally with Ethereum
- React front end to demonstrate a human trafficking data sharing scenario
- Data analysis example pulling data from AWS ElasticSearch
Challenges we ran into
Dynamo encountered the following non-trivial challenges when building out the Trusted Data Management:
- HyperLedger deployment issues setting up the appropriate BC nodes
- Ethereum / Solidity implementation issues and JS dependencies
- The sheer number of systems were had to integrate with.
Accomplishments that we're proud of
The Dynamo Team is especially proud of the scope of the project as well as how we were able to incorporate cross-discipline expertise to attack a problem from many different angles. While we had an idea prior to the hack-a-thon, our ideas were solidified during discussions with mentors at the hack-a-thon, and we essentially built most of everything from scratch.
What we learned
While the team has a good amount of experience using and developing for each of the individual technologies, having to integrate them with each other has proven much more challenging. We also recognize that the problem set is difficult, and that success is just as (if not more) dependent on stakeholder buy-in, than any specific underlying technology.
What's next for Dynamo Trusted Data Initiative
Our goal is to create an ecosystem by which data scientists, government agencies, world organizations, and NGOs can participate and be more productive. We want to make our solution more robust, and begin pilot programs with stakeholders to address the many flaws that our POC has. We ultimately hope that by creating a system that people want to participate in, we can get better data, quicker, which ultimately will allow the really smart people out their respond better to the problem of human trafficking.