Uniphore is an AI innovation company that faced challenges with its legacy Windows Server infrastructure, which was costly and raised security concerns due to its end-of-support status. To modernize, Uniphore utilized Amazon Web Services (AWS) for containerization and cloud-native solutions, achieving a 30% reduction in costs and improved operational capabilities. The migration involved transitioning from 50 bare-metal Windows Server 2008 R2 instances to a Linux environment on AWS, optimizing call-center analytics workloads and enabling better training of large language models.
The modernization strategy included three key areas: application modernization through refactoring and containerization, a custom data migration solution using Type 2 hypervisor technology, and a cloud infrastructure setup with Amazon EC2, Amazon EKS, Amazon EFS, and Amazon S3. The migration was executed in three phases: deploying the hypervisor and DataSync agent, secure data migration and validation, and production cutover with application deployment.
Best practices adopted during the process included Infrastructure as Code (IaC) with Terraform, CI/CD pipelines using GitLab, comprehensive monitoring with DATADOG and CloudWatch, and enforcement of AWS Backup policies. The modernization resulted in benefits such as improved scalability and performance, flexibility and portability of applications, enhanced security and compliance, efficient data management, reduced operational overhead, and significant cost savings.