Three years ago, Jiri Kram, a solution architect who studied Fintech at MIT and specializes in cloud computing and blockchain, designed a Smart Cities solution to help a Spanish retailer get shoppers to the right stores while safeguarding their privacy. Last month, after seeing the situation in his beloved Madrid and eager to help combat the COVID-19 scourge, Jiri rebuilt his solution as a Smart Quarantine tool to help people avoid areas within cities with high incidences of COVID-19 cases.
This is Jiri’s first-person account of how did it. This article is based on the original proposal he submitted to the COVID-19 Global Hackathon.
Inspiration for the Solution
This Smart Quarantine solution is inspired by the situation in Madrid. I was thinking about how technology could relieve the pressure on the healthcare system, so that people don’t die in hospital corridors. Based on WHO and Gates Foundation information, I realized we need to slow down the spread of the virus. So I decided to repurpose a solution I originally designed in 2017 to get people into retail stores.
I did this as a volunteer effort for the COVID-19 Global Hackathon, where I was one of 16,000 attendees of the online event. The requirements for submissions were created by the WHO and other health organizations, and supporting partners of the event included AWS, Microsoft, Salesforce, Twitter, Facebook, Slack and others.
What the Solution Does
Smart Quarantine provides services to citizens and city governments.
For Citizens
- Provide proactive real-time alerts when they are in proximity of a zone (e.g. part of the city) or a location (e.g. a shopping mall, train station, etc.) where the risk of COVID-19 is increased
- Protect their identity in alignment with EU GDPR law, permitting use of the app without risk of “Big Brother”-like surveillance
- Simplify the onboarding process and minimize the volume of information that must be collected via device and or sent via network
For City Governments
- Create a social graph that provides visibility into how a city operates in real-time (e.g., how people move between each area)
- Enable cities to set, update and deploy real-time policies across a range of areas: city-wide, area-wide and location-wide (e.g., for a shopping mall or train station)
- Build machine-learning models that will (in a GDPR-compliant way) predict the movement of people between parts of the city
- Develop an early warning/prevention system that can send notifications and alerts (e.g., when a person is moving to an area with high numbers of confirmed COVID-19 cases)
- Ensure the solution can be used in European cities where GDPR regulation requires special architecture consideration
How I Built It
The original solution, built in 2017, was designed for a large retailer to enable them to get people to their stores based on the availability of the goods in stock per each store. It could also direct people to a new location when a store runs out stock.
Given that the COVID-19 situation requires a similar approach—directing people as they move inside the city—the architecture was repurposed to fit the new use case of enabling city governments to direct people (without stalking them!) from dangerous locations to safe ones.
The solution is GDPR-compliant out of the box, because it was designed specifically for use in places like Madrid, Barcelona, Milan and other large European cities. As it happens, each of those three cites has been hit very hard by COVID-19.
Therefore, I kept the same infrastructure, designed in compliance with TM Forum standards, to enable it to run blockchain ledger as part of an extension of an existing BSS. The solution also allows telcos to deploy GDPR-compliant solutions as value-added services to an existing subscriber base.
The solution also leverages the fact that the major cloud providers like AWS and Azure are now deeply integrated with telco infrastructure, which means that edge locations are now extensions of the cloud service. This has allowed us to orchestrate cloud and edge with low latency and thereby pave the way for driving real-time decisions via machine-learning engines.
And with the proliferation of modern mobile devices (primarily iOS and Android), it’s easy to use existing cloud services and send proactive notifications to users’ devices.
Lastly, GDPR compliance is enabled by characteristics of modern private blockchain solutions like Hyperledger Fabric, Corda or AWS QLDB, which is what I used in this case. We are creating an identity hash that can be used by APIs.
Challenges I Ran Into
The biggest challenge was finding the right team members. Not many people have the high level of expertise required for some of the technologies in this solution. For example, it is still very difficult to find people who truly understand how to use blockchain and machine learning together.
In part, this is because there is still a tendency to see blockchain as public (decentralized), and AI as centralized. Therefore, many people don’t understand the reason why a blockchain network should be run as private instead of public. Another issue is the need to find people with deep knowledge of why GDPR compliance is critical for a solution like this.
Accomplishments I’m Proud Of
While the solution was originally designed for Europe it was successfully tested in Dubai, with one undisclosed government agency. Therefore I know this is not just a pipe-dream.
What I Learned
A solution developed for one purpose can become the foundation for solving a completely different use case. In this situation, the original purpose was to protect the identity of people while allowing the retailer to send them recommendations and get them to the right store. And now, this solution is repurposed for protecting healthy people entering COVID-19 risk areas and locations. It’s the same principle of a smart city, but I’ve repurposed it to save lives.
What’s Next for Smart Quarantine
The next step is to form a partnership with a CSP (Communication Service Provider). The solution is ready but can be only deployed by the telecommunications provider.
Author’s Note: This picture, which depicts the Madrid Bear hugging a nurse, was posted by the CEO of Telefonica, Jose Maria Alvarez-Pallete. For American readers, this would be like the Statue of Liberty leaving her pedestal and hugging a nurse. When I saw this picture, I decided I must do something, and the global COVID-19 Hackthon gave me chance to contribute. For which I am grateful.
Jiri Kram is a solution architect who studied Fintech at MIT and specializes in cloud computing and blockchain. You can reach him on LinkedIn.
Subscribe to the free Cloud Wars Newsletter for in-depth analysis of the major cloud vendors, including our ongoing coverage of tech’s response to COVID-19.
