How Analytics Teams Using SSAS Can Embrace Big Data and the Cloud
- 1. How Analytics Teams Using SSAS Can Embrace Big Data and the Cloud Li Kang, Sr. Director, Strategic Business
- 2. © Kyligence Inc. 2019, Confidential. Challenges facing SSAS Kyligence Data Analysis Platform Case studies Q&A Agenda
- 3. © Kyligence Inc. 2019, Confidential. Big Data Era: Critical Capabilities for BI Platforms Scalability & Model Complexity " The degree to which the in-memory engine or in-database architecture handles high volumes of data, complex data models and large user deployments." Critical Capabilities for Analytics and Business Intelligence Platforms - Gartner
- 4. © Kyligence Inc. 2019, Confidential. Multi-dimensional OLAP Integration with Excel and BI Tools Advanced Analysis Capabilities MDX/DAX Interfaces SSAS: A Complete OLAP Solution
- 5. © Kyligence Inc. 2019, Confidential. Cloud native OLAP analysis Volume Dimensions Concurrency Cloud From gigabytes to petabytes 100s of concurrent users 10s or 100s of dimensions SSAS: Challenges for Today’s Enterprises
- 6. © Kyligence Inc. 2019, Confidential. Title Query Performance Cost to Support Concurrency Query Performance Concurrency Support Hadoop SQL Engines Cloud Data Warehouse Data Virtualization Challenges of Alternative Solutions Query Performance
- 7. © Kyligence Inc. 2019, Confidential. Agenda Challenges facing SSAS Kyligence Data Analysis Platform Case studies Q&A
- 8. © Kyligence Inc. 2019, Confidential. Top level Apache project • The only open-source OLAP on Big Data platform Sub-second Interactive queries • Large scale +high concurrency + sub-second query latency + multi-dimension Best open source big data tool • InfoWorld's Bossies (Best of Open Source Software awards) in 2015 and 2016 1000+ organizations • Adopted by thousands of organizations globally Apache Kylin: Open Source OLAP for Big Data
- 9. © Kyligence Inc. 2019, Confidential. Powered by Kylin Founded in 2016 by the creators of Apache Kylin Global Presence Headquarters: San Jose, US; Shanghai, China Well Funded Investors: Redpoint, Cisco, Coatue, etc. Kyligence = Kylin + Intelligence Open source community Enterprise-level Product Professional services Cloud Computing Industry Solutions Award Winning CRN Top-10 big data startups
- 10. © Kyligence Inc. 2019, Confidential. 1000+ Enterprise Customers Worldwide
- 11. © Kyligence Inc. 2019, Confidential. Kyligence: Big Data Analysis Platform • Sub-second query latency • 1,000s to 100,000s concurrent users • Rich BI Semantic layer • On-premise or in the cloud • Support ANSI SQL and MDX • Seamless integration with: • Excel, Power BI, Tableau, Qlik, MSTR... Data Application Data Source Semantic layer Query PushdownData Flow Intelligent Modeling OLAP Cube Enterprise-level Security MDX ServiceODBC/JDBC CloudOn-premise RDBMS Cloud Hive MPP Kafka MDXSQL Tableau PowerBI Excel
- 12. © Kyligence Inc. 2019, Confidential. Intelligent Routing Detailed Query Kyligence Enterprise Aggregate Query Query Pushdown Ad-hoc Query SSAS Cube CubeCube Aggregate Inquiry • Kyligence provides complete data analysis services, supporting aggregate queries, detailed queries, and ad-hoc queries to transparently accelerate data access. • Patented Query Pushdown technology ensures ad-hoc queries get returned when there are no pre-computations to return. Completed data analysis services
- 13. © Kyligence Inc. 2019, Confidential. Before SSAS • Cube-level and record level permissions only • No column level permission Kyligence • Granular security control, down to to cell-level • Integrates with enterprise directories • Supports big data security frameworks After Project Level Column LevelRecord LevelTable LevelCube Level Column LevelRecord LevelProject Level Enterprise Grade Security
- 14. © Kyligence Inc. 2019, Confidential. One-click sync with existing Kyligence Cubes Sync Cubes Define A Semantic Layer Semantic layer for measurements & many-to-many relationships Analyze Interactively Quick insights using Pivot Tables or other BI tools Integration with BI products in easy steps
- 15. © Kyligence Inc. 2019, Confidential. SSAS vs. Kyligence: Architecture Comparison • Support more data sources • Scale out architecture • Efficient data processing • Support ad-hoc analysis • Flexible queries • Enhanced security Advanced Analysis BI Connectors Complex Relationships Self-service Data Fetch Query Engine Multi-level Permissions Multi-tenancy Distributed Architecture Business semantic layer Data service layer InquireData flow Order CRM POS SSAS DW/DM Business Analysis Multi-dimensional analysis MDX Vertical Expansion Cache Business Analysis
- 16. © Kyligence Inc. 2019, Confidential. Category Subcategory SSAS（Multidimensional mode） Kyligence Technical Architecture Scalability • Storage mode: cache • Scale-up architecture • Expansion requires data migration after hardware upgrades • Storage mode: HDFS / Cloud Object Storage • Scale-out architecture • Requires only additional nodes for additional capacity • No downtime upgrades required Operational costs • Limits to dimension quantity and Cube size • Single Cube is up to 10TB • No limitations for the quantity of dimensions or the cube size • Reduced operational costs by replacing multiple traditional cubes with just one single cube BI tools integration Excel, Power BI, Tableau, MicroStrategy Excel, Power BI, Tableau, MicroStrategy, Qlik, Cognos, BO, OBIEE Max Data Volume Up to 200 million rows per partition Unlimited Query interfaces MDX MDX / SQL / REST API Support various OLAP mode MOLAP / ROLAP / HOLAP Refresh mode Incremental / Full User Analysis Multi fact table model Supported Hierarchy definition Supported measure definition Supported Many-to-many relationship definitions Supported Measure group definition Supported SSAS vs. Kyligence: Key Feature Comparison
- 17. © Kyligence Inc. 2019, Confidential. Challenges facing SSAS Kyligence Data Analysis Platform Case studies Q&A Agenda
- 18. © Kyligence Inc. 2019, Confidential. • Timely and accurate monitoring of all operations across the country through a “one-stop” decision support information service. • Greatly improve the efficiency of data analysis, provide complete data analysis services and more analytical explorations • Multi-level permission management to effectively ensure data assets security • Familiar cube technology means team can reuse their technical skills Value Big Data Exploration and Practice in Food Industry Case Study: MOLAP at A Global Fast Food Chain • Analytics for restaurant operations was running on GreenPlum + SSAS + Excel/Tableau. • SSAS struggles with the growing data volume. • Store expansion, product diversification, and orders via web and mobile led to explosive growth in data. • Looking for a unified big data analytics platform to improve the efficiency of analysis and support growth. Background • The fast food industry has complex data analysis models, multiple dimension indicators, complicated business logic and flexible analysis method. • Business requirements change often, but development and operations are must still be simple and efficient. • Big data platforms need to meet current business users expectations . Challenges • Kyligence as an intelligent data warehouse on the CDH platform, including BI semantic layer. • Integration with Excel and Tableau • Kyligence query service cluster provides high-availability and high-concurrency. Solution
- 19. © Kyligence Inc. 2019, Confidential. Restaurant order analysis, 30 billion records, 50 million records/day Sub-second Latency SSAS • Slow response for aggregate queries: over 10 mins • COUNT DISTINCT does not work at this level • Limited business analysis Kyligence • 80% of aggregate queries finish in 1 second • COUNT DISTINCT functions finish in seconds Case Study: MOLAP at A Global Fast Food Chain
- 20. © Kyligence Inc. 2019, Confidential. • Timely risk analysis and decision making based on large amount of historical data • Dramatically reduced total cost of ownership • Increased stability and high availability due to the distributed architecture • Familiar cube technology means team can reuse their technical skills Value Big Data OLAP in the Cloud Case Study: Global Investment Bank • Billions of rows of historical data to support risk analysis • Existing large SSAS deployment with many cubes • Fast response time is critical to make timely decisions • Enterprise is moving to the cloud – MS Azure Background • There is no native SSAS MOLAP service on Azure • SSAS scale up architecture becomes cost prohibitive to support the data volume • Too many cubes to manage due to the size limit of each cube • Cube loading and query takes too much time Challenges • Kyligence Cloud multi-dimensional analysis on Azure • Integration with Excel and other BI tools • Leverages cloud elasticity to support high performance query and keep the cost down Solution
- 21. © Kyligence Inc. 2019, Confidential. Case Study: Global Investment Bank Lower TCO • 350 billion rows • 100 concurrent users • $300/day cloud cost Easy to maintain • 1200+ dimensions • 1 cube Faster Cube-Building • 8 billion new rows daily • < 1 hour build time Cloud Test Results 90% SQL queries finish within 5 seconds.
- 22. © Kyligence Inc. 2019, Confidential. Agenda Challenges facing SSAS Kyligence Data Analysis Platform Case studies Q&A
- 23. Questions?
- 24. Contact Us 99 Almaden Boulevard Suite #320 San Jose, CA 95113 Address +1 (669) 256-3378 Phone [email protected] E-mail
- 25. THANK YOU