Case Studies

Read how Integra has helped Diglossia, literacy and language testing platform be reliant, scalable and fault tolerant, and how it caters to hundreds of thousands of simultaneous test takers.

Diglossia’s standards-based Arabic assessments provide actionable data to target specific skills, improve student outcomes, and measure progress over time. Diagnostic data allows teachers to provide meaningful formative feedback and develop personalized, evidence-based intervention strategies.

Diglossia stands out as the premier provider of a comprehensive program that encompasses assessment, training, strategic planning, data analysis, and reporting across all grade levels from KG2 to grade 11. The research-based, valid, and reliable Arabic assessments are widely adopted throughout the region. 

These assessments play a crucial role in identifying student learning gaps, and monitoring progress in early literacy skills, including phonemic and phonological awareness, alphabetic understanding, oral reading fluency, and proficiency in reading, writing, and vocabulary at both elementary and secondary levels.

Children Taking Diglossia Exams

The Challenge

Diglossia’s platform was initially built to cater to a small, closed, cohort  of students. The design decisions taken initially was based on the limited testing that was planned. As a result, with the platform becoming widely popular with educational authorities across the region, scalability and resilience was lacking. 

As testing grew, incidence of platform unavailability and slow platform responsiveness grew, forcing Diglossia to do a re-think on the software and infrastructure. It was decided that a massive modernization effort, both for the application and the infrastructure was necessary, and reliability, resilience and scalability would need to be the key for any future design considerations. 

The Solution

Initially, we identified the criticality of each applications, acceptable downtime, and data loss. Since the customers were mostly public sector education regulators and providers, a detailed regulatory and security requirement study was undertaken. The expected load and performance requirements were analyzed and documented.

High Availability and Fault Tolerance was ensured with a Multi-AZ Deployment model with istribution of EKS nodes across multiple Availability Zones (AZs) to ensure high availability. Cluster Autoscaler and Horizontal Pod Autoscaler were configured to handle varying loads.

Regular backups of etcd data, EBS volumes was scheduled and AWS Backup was also configured. Cross-Region Replication to another region using services like Amazon RDS Cross-Region Read Replicas and S3 Cross-Region Replication was configured to ensure that a single region failure would not affect the platform for a prolonged time.

Built and stored container images in Amazon ECR (Elastic Container Registry) and Helm Chartsare used for managing Kubernetes applications. CI/CD pipeline using AWS CodePipeline and CodeBuild for continuous integration and deployment was setup for automated deployments.

CloudWatch was setup along with Prometheus and Grafana Cloud for monitoring EKS clusters and applications. Fluentd is used to send logs to CloudWatch Logs.

Adequate security was ensured with the use of security groups and network ACLs to control traffic to the EKS nodes. For kubernetes cluster access, Role-Based Access Control (RBAC) was implemented for controlling access. AWS Secrets Manager is used for managing sensitive information.

Implemented chaos engineering practices to simulate failures and test system behavior, as well as regular load testing using external tools were also implemented.

We conduct regular disaster recovery drills to ensure our teams are always prepared for an actual disaster.

The Result

Diglossia was operating a platform that was quite limited in our ability to scale due to the constraints of the local data centers and the varied usage of our testing apllications – heavy during school terms, but hardly any usage during the holidays. The move to AWS and the modernization that was done changed the business entirely. The platform is now extremely scalable, reslient and able to cater to multiple customers, with new customer platform onboarding taking just minutes. Improved business revenue and increased customer satisfaction was the immediate benefits of the modernization saga.