Introducing Trillium DQ for Big Data: Powerful Profiling and Data Quality for the Data Lake
1 like756 views
The document introduces Trillium DQ for Big Data, a solution aimed at addressing various data quality challenges faced by organizations in managing extensive datasets. It emphasizes the importance of data profiling, cleansing, and matching to ensure accurate, usable data for decision-making, while also highlighting features that enable scalable processing and integration with big data frameworks. The presentation is led by Harald Smith, director of product marketing, and aims to enhance trust and collaboration in organizations' data analytics efforts.
Introducing Trillium DQ for Big Data: Powerful Profiling and Data Quality for the Data Lake
- 1. Introducing: Trillium DQ for Big Data Harald Smith, Director Product Marketing
- 2. Housekeeping Webcast Audio • Today’s webcast audio is streamed through your computer speakers. • If you need technical assistance with the web interface or audio, please reach out to us using the chat window. Questions Welcome • Submit your questions at any time during the presentation using the chat window. • We will answer them during our Q&A session following the presentation. Recording and slides • This webcast is being recorded. You will receive an email following the webcast with a link to download both the recording and the slides.
- 3. Speaker Harald Smith • Director of Product Marketing, Syncsort • 20+ years in Information Management with a focus on data quality, integration, and governance • Co-author of Patterns of Information Management • Author of two Redbooks on Information Governance and Data Integration • Blog on InfoWorld: “Data Democratized” 3
- 4. Data challenges across the business Business Leaders Lack trust in data needed to make rapid, accurate decisions that grow business Business Analysts Can’t access or understand data and spend excessive time on investigating Information Leaders Must facilitate business collaboration and data transparency and governance Chief Data Officers Make data a strategic business asset utilizing scientific skills from basic spreadsheet knowledge 4
- 5. Only 35% of senior executives have a high level of trust in the accuracy of their Big Data Analytics 92% of executives are concerned about the negative impact of data and analytics on corporate reputation New survey indicates nearly 80% of AI/ML projects stalling due to poor data quality 84% of CEOs are concerned about the quality of the data they’re basing decisions on Big Data Needs Data Quality
- 6. 6 Data Quality Challenges of Big Data Profiling Data • Organizations are storing vast amounts of data in data lakes and the Cloud – from many different sources – but that data isn’t usable unless it is understood and to understand it, the business users who work with the data must be able to access and profile it without constant IT help Matching Entities Accurately • Distinguishing matches that indicate a single specific entity across so much data requires sophisticated multi-field matching algorithms – that need to be understandable by business users to be meaningful Scalability • Distinguishing matches across massive datasets requires a lot of compute power - compare everything has to be compared to everything else, multiple times in multiple ways • Taking advantage of Big Data processing for scalability requires specialized skills and takes a long time – and requires tuning, re-writing as technology changes • Traditional data quality tools are not designed to work on that scale of data
- 7. Trillium DQ for Big Data Understand, Evaluate, and Resolve Big Data Quality Problems Trillium Discovery for Big Data Data Profiling Gain a complete picture of your data before use • Understand the data • Analyze the data • Find data quality problems • Build and evaluate data quality rules 7 Trillium DQ for Big Data On Premises or via Trillium Cloud Deploy any or all products to the cloud - Completely managed SaaS in AWS or Azure Trillium Quality for Big Data Data Cleansing and Matching Cleanse, standardize, and connect data in accordance with your predefined standards • Entity matching and resolution • Data cleansing and correction • Data record enrichment
- 8. Feature-rich data profiling and data quality processing engines • Leveraging over two decades of data quality expertise An efficient orchestration of this engine in Big Data distributed frameworks • Powered by an architecture that has been in production with very large (2000+ node) environments running natively across the cluster • Partnered with Cloudera and Hortonworks closely, native integration with the stack • Syncsort has been a major contributor to Apache Hadoop open source project • With efficient orchestration, we can process any number of attributes with a handful of MapReduce jobs • Same architecture is used for Apache Spark “Design once, deploy anywhere” architecture • Native connectivity providing breadth and performance • “Intelligent Execution” to optimize process execution at run-time (MapReduce, Spark 1.x, Spark 2.x) • On-premise and in the cloud (e.g. Amazon EMR) 8 Data Quality for Your Big Data Needs
- 9. Key Outcomes • Reduce the time for business analysts to discover and understand data on Big Data platforms • Allow business analysts who understand the data but have little technical expertise to quickly find data and run data profiling in three steps • Let analysts explore results and drilldown to details within 2-5 seconds per view to review and then report on data issues to business leaders • Scale to large volumes of data sources & attributes so that business analysts can understand the contents of any data source needed for business decisions • Data is always secured in process and at rest and only available to authorized users to comply with regulations and avoid fines 9 Trillium Discovery for Big Data
- 10. 10 Trillium Discovery for Big Data • Delivers enterprise trusted Trillium Discovery on distributed big data platforms (e.g. Hadoop, Spark) for high-volume, scalable data profiling • Provides complete Trillium Discovery data profiling for analysis & review • Attribute metadata, value & pattern frequencies, key & dependency analysis, cross-source join analysis, drill down to any outlier or issue, and more… • Provides easily configured native connectivity for Big Data sources • Provides managing and monitoring for task execution • Integrates with the security frameworks (Kerberos, AD, LDAP) of Big Data platforms
- 11. Run Profiling 1 n . . . . . . . . . . . . . . . . . . . . . . 11 Trillium Discovery for Big Data – Data Profiling at Scale Select Source Explore ProfilesRun Profiling Stored Profiling Results ▪ Metadata & Statistics ▪ Frequency Distributions ▪ Drilldown Indices Share & Govern Results Integration (APIs) Notification Collaboration Native Connectors ▪ HDFS source directories ▪ … Drilldown to IssuesEvaluate Business Rules
- 12. Key Outcomes • Match and link any data entity – customers, suppliers, products, etc. – into a trusted single view to support a broad array of business-critical use cases (e.g. Customer 360, fraud, AML) • Parse and standardize complex multi-domain data, extended with enrichment and verification of critical address and geolocation data – all leveraging out-of-the-box templates • Utilize “design once, deploy anywhere” approach to speed time-to- value and focus on building data quality business logic while letting the product handle the technical aspects of framework execution with no coding or tuning required • Leverage the high-performance compute power of distributed Big Data frameworks including Hadoop MapReduce and Spark to process high volumes within targeted time windows to meet critical Service Level Agreements (SLA’s) 12 Trillium Quality for Big Data
- 13. 13 Trillium Quality for Big Data • Integrate, parse, standardize, and match new and legacy customer data from multiple disparate sources. • Provide high-quality entity resolution through multi-domain deduplication and matching with the most comprehensive set of match comparisons available, including fuzzy matching, distance comparisons, and more. • Standardize, enhance, and match international data sets with postal and country-code validation. • Deploy data quality workflows as native, parallel MapReduce or Spark processes for optimal efficiency. • Process hundreds of millions of records of data. • Increase processing efficiency. • Support failover through Hadoop’s fault-tolerant design; during a node failure, processing is redirected to another node.
- 14. Trillium Quality for Big Data – Data Cleansing at Scale Boost effectiveness of machine learning, AI with complete, standardized, matched data. 1. Visually create and test data quality processes locally 2. Execute in MapReduce or Spark On premise or in the Cloud Big Data Platform 14
- 15. Syncsort Trillium Delivers Data You can Trust Data Profiling Business Rules & Data Quality Assessment Data Validation, Standardization, Enrichment & more Matching, Entity Resolution & Verification •Customer 360 •AI/ML Operational Integrations •Analytics & Reporting Data Governance Trillium Discovery for Big Data Trillium Quality for Big Data + Global Address Verification Trillium DQ for Big Data 15
- 16. Trillium DQ for Big Data Use Cases 16
- 17. Turn your Big Data into a trusted view of your customers, products and more Power machine learning and advanced analytics with reliable, fit-for- purpose data Gain actionable business insights from high-volume disparate data sets from across the enterprise Deploy industry- leading data quality processes at massive scale, with no coding or Big Data skills required Trillium DQ for Big Data evaluates & transforms your Big Data for trusted business insights
- 18. Anti-Money Laundering on Hadoop at Global Bank S O LU T I O N CHAL L ENGE • Must provide highly accurate entity resolution • Must be secure – Kerberos, LDAP • Must have lineage – data origin to end point • Massive data volumes • Scattered data – Mainframe, RDBMS, Cloud, … • Must archive unaltered mainframe data Full Anti-Money Laundering regulatory compliance with financial crimes data lake – high performance results at massive scale. • Full end-to-end data lineage supplied to Apache Atlas and ASG Data Intelligence • Cluster-native data verification, enrichment, and demanding multi-field entity resolution on Spark • Unmodified mainframe “Golden Records” stored on Hadoop Bank must monitor transactions to detect Money Laundering for FCA compliance. Machine learning can detect patterns, but … Requires large amounts of current, clean data. • Trillium DQ for Big Data • Connect CDC • Connect for Big Data 18
- 19. Trillium DQ for Big Data Cleanses Credit Data for Creditsafe C H A L L E N G E Ensure ALL DATA on each company is analyzed – and NO DATA from another company is accidentally included – to get accurate corporate credit ratings. • Need to profile, cleanse and enhance data to evaluate credit ratings for 80 million companies in U.S. alone • Existing solution lacked flexible de-dupe matching rules, scalability • Millions of records to analyze per company, in multiple inconsistent data sources, about 800 million/day total and growing • Solution must scale! S O LU T I O N • Amazon EMR Cloud • Trillium DQ for Big Data cleansed, standardized and matched over 130 million recs/hour on basic 10-node test cluster– met the business SLA with room to grow 96% Address Matching Accuracy after Trillium cleansing, standardization Saved software costs – Replaced multiple solutions and tools Saved Amazon cluster costs and left room for company growth “We can’t afford to miss information, or mix up information about businesses with similar names. Companies count on our highly accurate predictive scoring to provide fast, accurate ratings for their potential customers and vendors.” 19
- 20. Next Steps For more information on Trillium DQ for Big Data and our other Syncsort Trillium data quality solutions, please visit: https://www.syncsort.com/en/products/trillium-dq-for-big-data And: https://www.syncsort.com/en/integrate
- 21. Q & A 21