Tenaris’s data science team has introduced a new technology in its IT infrastructure called Kubernetes to increase the deployment speed of applications, reduce errors in all operational processes and standardize the data science tools.
The platform was rolled out during the lockdown due to COVID-19 by Tenaris’s data science team, based at the Dalmine, Italy, mill. The team’s main responsibility is to apply data science and machine learning to increase safety, quality and efficiency across industrial processes.
The first step was done in 2018, when the data science team started to leverage containerization for its workloads. Kubernetes allows having reproducible environments, where software and its dependencies can run in isolation, with a positive impact on security. However, they used to rely on time-consuming manual release cycles for deployments and updates.
“By adopting this cutting-edge technology, any data scientist on the team can now directly put into production an application,” explained Vincenzo Manzoni, Data Science Director Tenaris. “This means we can answer to operational needs in less time and with more precision, granting a high level of reliability to the whole system. Thanks to a highly skilled team, we were able to deploy both the Data Science Lab and the model service infrastructure 100 percent remotely.”
Deployment time of a new release went from hours to minutes. With 14 nodes and 200+ pods, Tenaris’s data science team is now running 90% of its most critical workload through this platform, with no downtime since the infrastructure commissioning.
Tenaris is one of the only companies in the steel industry that has an in-house Industrial Innovation department, which the Data Science team is part of. Formed by 40 employees across Italy, Argentina, Mexico and the Netherlands, this team seeks to find innovative solutions to improve Tenaris’s industrial performance.