Powering drug discovery and innovation by dynamically creating and scaling high-performance computing clusters

About the customer

Johnson & Johnson, the world’s largest healthcare business and a top 50 Fortune 500 company, produces pharmaceuticals, medical devices and consumer goods that benefit a billion people worldwide. Johnson & Johnson owns Janssen Pharmaceuticals, which created the one-dose COVID-19 vaccine that’s preventing infection and saving lives in over 100 countries. This research-driven organization relies on high-performance computing (HPC) to power the discovery and production of effective, broadly available pharmaceuticals.

Their challenge

The nature of Johnson & Johnson’s research requires on-demand compute capacity at scale and the ability to easily reduce that capacity once it’s no longer needed, which can only be accomplished with cloud infrastructure. Janssen runs over 10 production HPC clusters on Amazon Web Services (AWS) – used by scientists and developers around the world, including those working on the Janssen COVID-19 vaccine – and was looking for an off-the-shelf solution to replace open-source Grid Engine and a cloud management tool that no longer supported their vendor of choice. Challenges included accommodating existing infrastructure and systems that had evolved over a decade, a complex networking setup, and the necessity of integration into configuration and change management systems. Additionally, every cluster was configured differently.

Johnson & Johnson, owner of Janssen Pharmaceuticals, needed the right HPC management software for its cloud-based infrastructure.

Our solution

Janssen upgraded workload management software to HPCWorks™ Grid Engine™ software and deployed HPCWorks Navops™ software to manage the company’s complex cloud deployments. Both are part of the Siemens Xcelerator business platform of software, hardware and services.

With a modern, commercially supported version of HPCWorks Grid Engine and a single HPCWorks Navops server managing multiple clusters, the solution seamlessly integrated with AWS cloud services. The package fulfilled Janssen’s requirements to work with automated creation and scaling of clusters, integration and compliance with internal security policies and networking, both commercial and homegrown applications, and Good Practice (GxP) compliance. It included groups for resource allocation, docker integration for containerized workloads, and Representational State Transfer Application Programming Interface (REST API). The result was a simplified, automated and extensible Janssen HPC infrastructure.

With the Siemens orchestration layer, Johnson & Johnson can scale on demand in a complex multi-tenant environment in which each group of customers has access to purpose-built clusters. The team can use current configuration components to create additional clusters quickly and easily as needed. Siemens infrastructure-as-code solution included complete user interface (UI)/command line interface (CLI)/API configurability, flexible configuration and integration options, and the ability to create a template for necessary components and easily edit configuration differences.

“The HPCWorks Grid Engine and HPCWorks Navops solutions enabled us to scale in the cloud and grow our HPC infrastructure to support critical science and research, including COVID-19 vaccine development,” says Martin Dellwo, manager of advanced computing, Janssen R&D.

Results

Janssen uses HPCWorks Navops to create and scale HPCWorks Grid Engine clusters dynamically, allowing the firm to continue innovating. With the Siemens solution, scalability limits from Janssen’s previous configuration were removed and their clusters are now running at three times the size of the previous implementation. With increased scalability and improved agility, operations, deployment and management of HPC infrastructure, the team can perform science on demand with the flexibility to tackle events like COVID-19 at an enterprise level.