Tamr helps enterprise companies solve large data unification challenges by breaking down multiple data silos and unifying data into pipelines through automation. Tamr was founded in 2013 and is based in Boston, Massachusetts.
To support its data unification products, Tamr Steward and Unify, the company wanted to improve IT performance and scalability while reducing hardware costs.
For Tamr, supporting a wide array of customers requires powerful IT resources. The company frequently works with clients such as Societe Generale, Carnival, GSK, and Toyota with established IT infrastructures that range from full on-premises implementations to hybrid on-prem and cloud back-ends. Given the variety of supported environments, Tamr needed to be able to adjust its services to work alongside differing customer setups.
Tamr solutions also have to support high-resource performance demands such as automatically sorting through multiple customer records and verifying identities via machine learning. All of these factors encouraged the company to prioritize improving its platform’s performance and versatility.
“We’ve always developed our products in direct response to our customers’ most pressing business requirements, including whether they aim to be cloud-based or operate exclusively on-prem,” says Daniel Bruckner, Co-founder and Engineer at Tamr. “We needed a platform that would allow us to grow and maintain this open approach.”
Migrating to Google Cloud Platform (GCP) helps ensure that Tamr can reliably deliver its data unification solutions to customers and sustain future growth. Tamr uses GCP to support its main products: Unify, which harnesses machine learning and automation to establish a centralized data pipeline, and Tamr Steward, an issue tracker that integrates into custom and common analytic applications such as Tableau and Microsoft Excel and is used by customers to collect feedback on data & analytics.
Unify is powered by Compute Engine, which helps support in-house Tamr machine learning resources that combine traditional models with human expertise that helps guide the models. Cloud Dataproc supports Apache Spark workloads, while Cloud Bigtable assists with database storage.
For Tamr Steward, Google Kubernetes Engine and Cloud SQL provide containerization and database resources. Cloud Pub/Sub and Cloud Run establish alternative development channels used for features such as in-app notifications and activity triggers, while Stackdriver offers additional system management support. For general platform needs, Tamr has seen cost savings by moving to Cloud Identity Access Management and Firebase Authentication. The company also uses Cloud Storage and G Suite for additional platform support and to enhance everyday employee work.
To optimize their Google Cloud usage and manage costs, Tamr engaged with Google Cloud MSP, SADA. SADA participated in consultation and guidance around networking and security, connected Tamr with the necessary big data resources, and helped them identify potential cost savings and ongoing cost optimization. With the cost savings identified by SADA and security architecture set in place, Tamr was able to focus on innovation and developing new products.
Tamr has seen immediate performance gains alongside cost savings since migrating to GCP. Routine tasks such as making host provision calls or running requests in Tamr Steward that used to take minutes can now be completed in seconds via GCP scalability and processing power. Improved processing times also helped spur a new developer service that immediately launches instances from pull requests. “The launch would not have been possible on its legacy back-end,” explains Daniel. Immediately following the migration, the Tamr team estimates that the company achieved nearly 50 percent cost savings.
The performance gains from GCP have had a significant impact not just on Tamr but also on its customers. Within Tamr, faster-performing tools promote developer innovation and improved speed to market. By having versions of Tamr Steward and Unify that run faster, Tamr’s customers ultimately benefit from a greater understanding of their data pipeline so they can realize higher-quality insights.
“By moving to Google Cloud Platform, we achieved a nearly 10 percent performance boost across our entire platform,” says John Imbrescia, DevOps Lead at Tamr.
Results and Benefits
- Improves product responsiveness by nearly 10% through Google Cloud Platform performance gains
- Reduces infrastructure costs after replacing legacy platform with cloud-native approach
- Creates up to 50% cost savings following migration to Google Cloud
Source: Google Cloud Customer Stories