Background
The Challenge
- Limited interactivity and collaboration due to lack of real-time communication and live user sessions
- Infrastructure and scalability constraints during compute-heavy and memory-intensive genetic optimisation tasks
- Inflexible architecture lacking tenant isolation and support for dynamic, AI-powered SaaS delivery
- Rigid deployment pipelines that restricted adaptability for multi-tenant environments and seasonal demand spikes
Our Solution
The architecture leverages Amazon EKS for secure, scalable genetic algorithm processing. Microservices and autoscaling ensure reliable performance and smooth collaboration across fluctuating workloads.
- Scalable compute environment using Amazon EKS with optimised node groups, autoscaling, and cluster-level elasticity
- Microservices design separated WebSocket handling, genetic algorithms, and backend logic for modular, responsive performance
- Real-time collaboration supported by sticky sessions, session pooling, and health checks across multi-user environments
- Robust traffic management with an Application Load Balancer (ALB) handling SSL, WebSocket upgrades, and path-based routing for computation workloads
The Result
The architecture delivered a cutting-edge foundation for MateSel’s future product roadmap, enabling expansion into full-scale SaaS delivery with real-time collaboration and genetic modelling capabilities.
- Enhanced user experience with 100% improved interactivity, live updates, and seamless multi-user collaboration
- Faster, more reliable performance including 50% quicker genetic computations and 99.9% platform availability
- Cost and resource efficiency achieved with 65% lower compute costs via autoscaling and optimised resource usage
- Scalable, secure SaaS delivery supporting multi-tenant isolation and faster innovation through CI/CD pipelines
