Cloud-based SaaS Platform for identity retrieval and matching
Business Requirements
The client needed a platform that would allow them to query and obtain detailed and private information about individuals and entities. They required access to a large dataset of approximately 600 million records and growing.
The obtained identity information was to be utilized in various domains such as targeted marketing and securing high-value transactions in the financial sector.
Nuage Solution
The solution provides a platform to the client to query and obtain detailed, sensitive, and private information about individuals and entities. We enabled the client to obtain the required information using our cloud-based APIs, which were implemented using a server-less and micro-service-based architecture accessing very large datasets (circa 600 million records and growing). The identity information so obtained has been put to use by the client in domains ranging from targeted marketing to protecting high-value transactions in the financial sector.
For near real-time processing, a fan-out design pattern for Lambda execution was used. While for batch processing, we used AWS Step Functions to invoke Lambda in the Service Call-back design pattern. Also, AWS Aurora – PostgreSQL was used as a primary data store along with API Gateway with Lambda Auth function for authenticating and authorizing API requests. The team used a micro-service-based serverless architecture based on AWS Lambda for the implementation of APIs. Which were then stitched together using an in-house developed orchestration engine. Also, the team used Jasper Reports to get end-user reports that were required for billing and utilization. While a configurable GUI was produced with HTML5/CSS along with using Angular.
A key challenge in uniquely identifying a person or entity based on non-deterministic elements like name and address information is because of the use of nicknames, spelling variations or mistakes as well as the use of abbreviations. We have built a solution based on a customized multi-tier approach for fuzzy matching implemented by leveraging various text search capabilities provided by the PostgreSQL database. We have implemented both Soundex and Levenshtein algorithms for matching.
We have also implemented dictionaries that can be customized as well to perform many filtering tasks that can be leveraged in conjunction with the tsvector approach for matching the free text address lines containing apartment numbers, street names, etc.
The solution was implemented as a cloud-based software-as-a-service (SaaS) by the client.
Technology Stack
Frontend: HTML5, CSS, Angular 6
Backend: Java (for AWS Lambda functions)
Database: AWS Aurora – PostgreSQL
Cloud Infrastructure: AWS, AWS Lambda, AWS API Gateway, AWS Step Functions
Reporting: Jasper Reports
Fuzzy Matching Algorithms: Levenshtein, Soundex
For more such analysis and insights, click here- https://nuagebiz.tech/case-studies/
For more details and personalized assistance, reach out to info@nuagebiz.tech