How Daiichi Sankyo modernized drug discovery using AWS Parallel Computing Service

This article describes how Daiichi Sankyo modernized their drug discovery HPC infrastructure by migrating from AWS ParallelCluster to AWS Parallel Computing Service (PCS).

• AWS PCS is a managed HPC service using Slurm job scheduler on EC2
• Daiichi Sankyo reduced operational burden and eliminated person-dependent management
• Custom AMIs created via EC2 Image Builder streamlined cluster deployment
• AWS Step Functions and SSM Documents automated Linux user/group provisioning
• Console-based operations enabled staff with limited cloud experience to manage clusters
• Flexible resource scaling supports diverse research department needs
• PCS improved failure recovery and reduced Slurm troubleshooting time
• Cost justified by reduced operational management burden and increased efficiency

PCS enables Daiichi Sankyo to maintain sustainable, flexible HPC environments for drug discovery research with significantly reduced administrative overhead.