Blindspot is a fast-growing company based in Prague that helps other companies to master the power of artificial intelligence by designing tailored AI solutions. One of their flagship products is the Anomaly Detection Framework (ADF). ADF enterprise platform tools dive in the streams of data from various systems in computer networks to detect cybersecurity threats or unusual malware types across critical devices, such as servers or databases, or applications that are vital for smoothly running operations.
Deployment challenges
With the latest decision to push towards the cloud the team was given a task to migrate and deploy their single-tenant/on-premise platform as a multitenant containerized solution for one of the Blindspot partners with the tight deadline of a less than 3 months. The project included not only the provisioning of several AWS environments connected with on-premise datacenter but it also required the deployment of CI/CD pipeline and integration with remaining systems of that particular partner.
Flexible and scalable environment in AWS
One of the main goals was to set up the enterprise grade ecosystem that would allow seamless migration from former tools and simple integration of new end users while taking into account specific requirements for resources driven by customer data rates and volumes.
The foundation part uses AWS Landing Zone with multiple accounts deployed by AWS Control Tower and Security Hub for ongoing security compliance assessment.The infrastructure consists of multiple Amazon Elastic Kubernetes Service clusters (AWS EKS) and Amazon Elastic Container Service for container orchestration (AWS ECS). Connectivity between AWS accounts and on-premise data center is done via AWS Transit Gateway.
While user workloads are hosted in different EKS clusters, the use of shared managed services like AWS Aurora or AWS EFS, apart from other monitoring tools, helps to increase the efficiency of the communication between multiple ADF components, reduces the cost of the overall solution and provides additional insight into it.
Benefits of AWS deployment
One of the key success metrics is the speed of the service provisioning for the new end user. Compared to former on-premise systems, now when the service is configured and automated in AWS the user onboarding time is reduced by roughly 41 %. Also, AWS based deployment removes the on-premise HW limitations and can easily adapt to the ever changing system requirements coming from specific end users.
With a container based approach, the scalability and elasticity to achieve desired capacity according to the needs of the workload is almost instant.
Finally, different environment stages and tight integration with CI/CD pipelines allow Blindspot and their partner to constantly develop new features while reducing the operation risk for the end users which makes the innovation more efficient, stable and faster.
| BEFORE | NOW |
| on-premise data center |  AWS |
| slow onboarding time for new users | onboarding time of new users reduced by 41 % |
| HW limitations | dynamic scaling |
| 0 possibility to scaling and adapt to user’s requirements | no limitations to innovation |
| high cost of HW upgrades | cost reduction |
| AWS and on-premise data center connected |
