Lohika team has started with reviewing the functional and non-functional requirements. Which helped to crystallize the high-level vision of our client into the low-level architecture that met all criteria. The new architecture design was event-driven with the three main architecture qualities in mind:
- Performance and overall system efficiency
- Extendibility to minimize the degree of manual work incorporated
The first component that Lohika has built was a rule-based engine with a custom DSL, developed specifically for our client. This engine was the heart of the system, created to block potentially fraudulent transactions via a set of rules. Our solution allowed for the creation of new fraud-detection rules – without involving developers. This fraud-detection solution ensured the company was capable of reacting to a new type of threats faster than most of its competitors.
While the team was working on the rule-based engine, they also built a CI/CD pipeline using CircleCI to automatically build, test, and deploy the solution in AWS.
All services were dockerized and running inside ElasticBeanstalk environments leveraging SQS as a communication channel. The use of AWS services eliminated a lot of DevOps challenges related to infrastructure deployment, configuration, and support.
Lohika extended the integration with Skrill payment gateway, removed the architectural limitation for the number of payment providers per country, and by integrated other payment gateways including Stride, Adyen, and iDEAL.
Lohika integrated Trulioo to speed up the process of identity verification. This step allowed our client to reduce manual work performed during the uploading and analyzing of personal data.
There was a requirement for the platform to be PCI DSS compliant. In the scope of this goal Lohika team implemented data encryption with secret key rotation and secure logging adapter that prevented sensitive data leakage.