Title

AWS re:Invent 2022 - Syngenta’s Cloud Journey to a Decision Science–Based R&D (PRT056)

Summary

• Jim Mellon, head of architecture technology data operations at Syngenta, shared the company's two-year journey in enhancing decision science and decision-making capabilities.
• Syngenta, a large agrochemical company, operates in 100 countries and is heavily R&D-focused, generating billions of data points annually from field trials and lab research.
• The company migrated from Hadoop to Amazon S3, managing about five petabytes of data across a hybrid infrastructure of on-premises and AWS environments.
• Syngenta aimed to modernize its applications from monolithic designs to a layered architecture, emphasizing speed and data protection.
• The company transitioned from siloed data sets to data platforms, with a focus on data stewardship, rationalization, and access for applications and data scientists.
• Data virtualization was introduced using Denodo, allowing for the creation of logical data views and a unified data access layer without moving data physically.
• Redshift was initially used for data acceleration, later replaced by Snowflake for improved performance, facilitated by the virtualization layer.
• Future plans include creating a global data mesh to comply with data localization requirements due to policy and government regulations.
• Key learnings highlighted the importance of non-judgmental attitudes towards past practices, the expectation of bumps in the road, and the need for fast, minimal-impact transitions.
• The benefits realized include faster onboarding of new data sets, increased user base for data-driven decision-making, and the creation of over 500 business views.

Insights

• Syngenta's journey underscores the importance of data management and modernization in R&D-intensive industries, where decision-making is heavily reliant on large volumes of diverse data.
• The transition from monolithic applications to a layered architecture with a focus on data platforms reflects a broader industry trend towards microservices and cloud-native technologies.
• The use of data virtualization technology like Denodo demonstrates the value of logical data management in complex environments, enabling agility and reducing the need for data duplication.
• The switch from Redshift to Snowflake within six weeks showcases the flexibility that virtualization provides, allowing for rapid adaptation to new technologies and performance requirements.
• Syngenta's approach to handling data localization challenges by creating a global data mesh could serve as a case study for other multinational companies facing similar regulatory pressures.
• The increase in user base from a handful of scientists to over 350 within a year indicates successful adoption and the effectiveness of the new data infrastructure in supporting data-driven decision-making.
• The lessons learned about maintaining a non-judgmental attitude and the importance of learning from mistakes are valuable insights for any organization undergoing significant technological change.