Envision Energy Case study

Project: DevOps for high performance computing

Company: Envision Energy

Industry: Energy

About Envision Energy

Envision Energy is committed to creating a sustainable future by revolutionizing the energy industry through advanced technologies. Since its foundation in 2007, Envision Energy has maintained rapid growth in its business operations, becoming a top tier global provider for smart energy technologies and services, including sales of smart turbines, smart energy management software and technological services. Envision Energy’s R&D efforts include building machine learning and data analysis technologies that turn data into valuable information.

Challenges

In a traditional architecture, gaining actionable insights for various computational intensive workloads can present a challenge in areas of cost, velocity, and scale. Due to the vast amount of data being computed, a new DevOps and HPC approach was desired in order to quickly scale the compute resources needed to keep R&D velocity. The economics of purchasing IT solutions for peak workloads that lasted only a few minutes each day did not make sense.

Challenge 1

Intensive workloads increased cost.

Challenge 2

Large amounts of data needed to scale to keep R&D velocity.

Challenge 3

Supported scaling around the clock.

Solutions

Envision depends on ML and data analysis to turn renewable energy data into actionable insights. With high-performance compute and large storage requirements, Eplexity designed and implemented a DevOps process to support Envision’s rapid, dynamic workload scaling. Automation included cluster sizing based on job metrics and compute node availability, spin-up and clean-up of HPC Clusters, large data set archival, and DR systems provisioning. Utilized EC2, ELB, S3, EFS, Glacier, CFNCluster, and SLURM.