Success Stories





Migration from bare-metal
Health Insurance Distribution Platform

The client runs a health insurance distribution and health plan management platform that helps to decrease administrative expenses, simplify the overall process of choosing the best-fit health insurance and integrates advanced health products. The client maintained their own servers and network since inception.

Reason for migration

The client’s product development organization of 300+ staff members was substantially limited by their existing bare-metal infrastructure. The IT team had limited resources and a high degree of difficulty maintaining their on-prem infrastructure and were heavily challenged with the business needs for  performance, scalability, security and disaster recovery. It was difficult to scale and coordinate the teams and develop new services with CI/CD in place from day one. AWS was determinte to be the best option to achieve their objectives of elasticity and administrative simplicity, as well as meeting HIPAA compliance requirements.


To ensure the deployment of additional cluster capacities, A2C Cloud created an automated deployment template to accelerate the setup of new instances, from the initial 40-80 man hours to just 2-4 hours of automated processes.

The infrastructure was moved from bare-metal servers to the Amazon province cluster (EKS), which is incorporated into two subnets of one VPC – public for the world and private for the team. Cuber's nodes are in autoscale groups and spread over public/private parts of the network. Workers nodes with the app located in the private zone.

Throughout the process of moving to the AWS cloud, the following tasks were accomplished:
  • Kubernetes cluster in it’s own VPC.
  • Elastic Load Balancer and Public Ingress Node as for the public part.
  • OpenVPN for access to the private part of the network.
  • For AppsServices subnet – a Kafka cluster consisting of 3 servers.
  • Implementation of RDS PSQL, ElasticSearch, ActiveMQ, S3, CloudFront, Route53.
  • Special protection measures - App+DB_s were hidden by a private network and closed for access from the Internet
  • Visible part - only LoadBalancers.
  • Access to the cluster restricted by IAM+VPC endpoints.
  • MongoDB Atlas as a service for DB
  • Auto-Scalable function and Userdata deploy function
  • Support for three CloudFormation templates - one for VPC, one for RedHat SSO, and one for everything else
  • Use of peering connections, security groups, IAM roles/policies


Migration Summary

Project Duration - 5 months (initial system review, new environment configuration, data migration, deployment, DNS switching to work on new IPs)



Dedicated bare-metal servers with Kubernetes and virtual machines

Separate server per service

10 simultaneous users

About 100 simultaneous users in a minimal state and up to 10,000 in a boosted state

40-80 hours for environment deployment

4 hours max for environment deployment

Manual deployment

Automated deployment recipes