Case Studies

Enhancing System Performance for Analyst Teams

Customer Challenges

Our customer has teams of developers applied to creating new and faster ways of enabling analysts to derive insight from big data. Their analytical applications were built primarily on a monolithic code base within their own data centers, with no centralized scheduling for data science and other containerized applications. The performance and scalability of these applications were also a major challenge, since the existing architecture supported multi-tenant environments with unique software level dependencies for each group analyst. The customer needed a standardized platform to support the work of their developers, with the aim of delivering analytics solutions to their analyst teams faster.

Solution Features

Our team of specialist engineers at ATG redesigned the customer’s infrastructure into microservices, using Kubernetes for container orchestration and Docker for containerization. As part of this development, we were able to build on-demand scaling and new failover policies. We configured custom metrics for Kubernetes, enabling the customer’s IT team to define and apply finer-grained scaling rules, such as tasks in a queue, rather than having to rely on less precise CPU and memory metrics. This has enabled the customer’s IT team to focus its attention on managing the common orchestration layer across all containerized applications. The net result has been high confidence in the team’s ability to meet demanding SLAs in support of their analyst teams.

Benefits to the Customer Mission

This new environment has enabled substantial improvements in developer productivity and speed of delivery by:

 

  • Creating “self-service” environments for data science specialists and other users.
  • Enabling better resource utilization through centralized scheduling of data science and other analytical workloads.
  • Simplifying the deployment and management of complex infrastructures for the customer.

 

 

An Acacia Group Company