Enabling Real-Time Analytics with CDC at Philips

Divya van Mahajan Introduction Marco Blanca Chief Product Owner SRC Philips Lead Big Data Consultant Devoteam Frank Knot Sr. Enterprise Solutions Architect Fivetran

Agenda 3 ● The Company ● Fivetran technology ● The Challenge ● The Solution ● Use Cases ● The Journey ● Key Learnings

© Koninklijke Philips N.V. At Philips, we’re empowering people to take care of their health and well-being and helping doctors and nurses to provide better, more sustainable and more convenient care – in hospitals, clinics and the home.” Roy Jakobs CEO Philips “

Strong positions across our portfolio of businesses Last 12 months’ sales Connected Care 29% of sales Personal Health 20% of sales Diagnosis & Treatment 47% of sales Diagnostic Imaging Monitoring Sleep & Respiratory Care Personal Health Ultrasound Image Guided Therapy Enterprise Informatics Systems, smart devices, software and services, powered by AI-enabled informatics Supporting precision diagnosis and minimally invasive treatment in a growing number of therapeutic areas such as cardiology, peripheral vascular, neurology, surgery, and oncology Ambulatory, home-based and in-hospital monitoring and workflow solutions fueled by advanced interoperability and patient data insights Connecting patients and caregivers across care settings, delivering clinical, operational and therapeutic solutions Broad range of consumer solutions to support people in proactively managing their health and well-being Market-leading capabilities integrating platforms, informatics, and services

6 Data Lake and Analytics at Philips Data Lake foundation in Azure supporting DaaS; Real-time data; data application initiatives

7 Marco Blanca Introduction Tribe Lead Data & AI Strongest capabilities Cloud | DevOps | Data-Driven | Integration & API Local partnerships Some known clients Key facts & figures 300 tech experts | 100+ satisfied clients Strategic partnership focus

Reference Architecture Events Databases (Cloud or on prem) Application Files Data Sources Destinations Data Lakes Cloud Data Warehouses Data Tools Orchestration Governance Observation etc. Data Lakes Cloud Data Warehouses Fivetran Event Platforms End to End Automation, Reliability and Scalability

9 The Challenge Architecture before Fivetran/HVR adoption ● No Real-Time ● Source deletions cannot be replicated ● Limited extraction time window ● Synapse not suited for very large Audiences

The Solution: Fivetran & Databricks to deliver great value

● Increase the digitalization of factories and enable Data Driven decision to improve KPIs like productivity ● Standardize the landscape, increase data quality and reduce manual work Digital Factory

● Provide Supply Chain information in a timely manner, safely and cost efficiently. Supply Chain Optimization

13 2020 Start Replication with ADF, Synapse Dedicated as a target Paradigm Shift Replication via Fivetran for ERP, First tables landed on Delta Lake as a target 2021 Consolidation 5 main ERP Systems and manufactories on-boarded 2022 Optimization Introduction of Databricks Streaming Jobs to reduce costs and increase flexibility Summer 2022 Scale Introduction of Unity Catalog and Salesforce Fivetran connectors Today # Data Sources: > 50 # Tables: > 9000 Volume: > 1 PiB # Changes Daily (average): > 180M (with spikes > 3B) #Streaming Jobs: > 80 Our Journey to a Mature Data Lake

Close collaboration with business to deliver data at the expected velocity 14 Fivetran and Lakehouse have proven to be an efficient solution to extract data from SAP 60% Cost saving achieved using Spark Streaming for Silver Automate-First approach to properly handle the amount of data sources Key Learnings Close collaboration with Fivetran is a key success factor for technical challenges

Questions?