This is precisely why LONTEX, a provider of e-toll systems,
faced the challenge of deploying an architecture capable of processing huge data streams from 100,000 lorries in near real time. A team of engineers from LONTEX designed a modern, advanced cloud environment from scratch and selected the optimal technology stack.
The Hostersi team was brought in to validate, implement and optimise this demanding project, supporting the process with their expertise in AWS, automation and security.
The Challenge
The main task was to efficiently and securely deploy a system in the cloud to handle data from on-board devices in lorries, in accordance with the architecture developed by the LONTEX team. Key requirements included:
- High throughput: Handling 100k devices sending data frames every 5 seconds.
- Low latency: The need to decode and optimise traffic before sending it to the Kafka cluster.
- Data criticality: Ensuring full synchronisation with the e-toll system, including an acknowledgement (ACK) mechanism.
- Migration and automation: Transition from the old test environment to modern managed services, whilst fully restructuring the infrastructure.
The solution and Hostersi expert role
The infrastructure designed by the client’s team was divided into four environments (Prod, Demo, Dev and Tools). At the heart of the solution was an Amazon MSK (Kafka) cluster, recognised as the most critical element of the system, storing data from the entire previous week.
Our role focused on the implementation, audit and optimisation of the individual layers of the system:
- Code-based infrastructure provisioning (IaC): we set up the complete production and pre-production infrastructure, describing it entirely using code (Terraform) and Helm charts. This enabled automation, rapid cloning of environments and the elimination of human error.
- EKS optimisation and configuration: The application environments host microservices (Frontend and Middleware) on Amazon EKS clusters. Our team was responsible for the advanced configuration and optimisation of the clusters, supported by the Karpenter tool for dynamic node management. Furthermore, to minimise latency, the ingress workers (AVL Ingester) were deployed on dedicated Auto Scaling Groups (ASGs) outside the EKS cluster.
- Databases and Cache: The architecture utilised RDS PostgreSQL for relational data and ElastiCache for Redis for high-speed in-memory operations.
- Security audit and optimisation: we carried out a comprehensive infrastructure audit. We configured and optimised the security stack, utilising, amongst other tools, AWS Inspector, which ensured the highest standard of cloud data protection. Access to the private layer was secured via a VPN, whilst the dashboard for management and clients was based on Cloudflare Zero Trust (OTP authentication), which eliminated the need for non-technical users to use a VPN.
- Observability and monitoring: Within the Tools environment, we configured and optimised a central monitoring system based on the Grafana Stack, including Mimir (metrics) and Loki (logs), whilst also integrating them with AWS CloudWatch to ensure full visibility of processes.
The MAP Programme and Funding
Implementing such a complex project also required a strategic approach to funding. As an experienced AWS partner, we enrolled LONTEX in the Migration Acceleration Programme (MAP). As a result, the client received significant financial support for the implementation work, as well as credits to cover the costs of AWS infrastructure, which greatly lowered the barrier to entry into the new, advanced architecture.
Results and next steps
The implemented and optimised architecture provides a stable foundation for the e-toll system, enabling the smooth management of traffic whilst maintaining full control over costs, performance and security. Our partnership is constantly evolving. We are currently finalising the next stages of our collaboration:
- Implementation of administrative support
- AWS reselling
- Implementation of the DRP plan
