New BI and IMC
Download as PPTX, PDF3 likes432 views
The document discusses the evolution of business analytics and intelligence (BI), contrasting traditional methods with emerging trends emphasizing real-time, cloud-based solutions. Key topics include data visualization, self-service BI, and the impact of technologies such as IoT and AI on analytics. It highlights the importance of immediate data access and simplicity in database design to drive contemporary BI applications.
![Architecture
• Database is memory-resident and not disk-resident
• Database design is simpler
• Fetching columns is simpler than querying from complex tables
• High Availability and Fail-over mechanisms
• RAM/ Battery RAM/ NVRAM [ferroelectric RAM (FeRAM), magnetoresistive RAM
(MRAM) & phase change RAM (PRAM)] ensure almost zero-loss availability
Top 3 IMDB: SAP HANA, Couchbase Server, MemSQL](https://image.slidesharecdn.com/newbiandimc-170227085519/85/New-BI-and-IMC-9-320.jpg)
New BI and IMC
- 1. New BI: IMC FUTURE OF BUSINESS ANALYTICS AND BUSINESS INTELLIGENCE
- 3. Business Analytics & Intelligence Life Cycle Business Requirement Data Engineering Analytics Modelling Data Visualization &DSS
- 4. New BI versus Old BI • Old is not yet old! • Traditional ETL + EDW + Dashboards • Tableau / Qlik / Microsoft / SAP BO / SAS / Pyramid / Oracle Hyperion… • Pre-requisites: On-Premise Infrastructure, Specialized Data Transformation Tools, Visualization • New is still evolving… • Ubiquity + Cloud + Mobility + Visualization • Tableau / Qlik / Microsoft / SAP BI & Fiori / Hadoop & SAS… • New requisites: Real Time, On Time, Big Data, AI, IOT, IMC
- 5. Evolution in Computing Technology for BI Real TimeDynamicStaticReporting Technology Data Information DSS Intelligence Business Orientation User Environment
- 6. Top 7 BI Trends – 2017 (n-2772; Scores >6/10) • Data Visualization • Self Service BI • MD / DQ Management • Analytical Databases • Data Governance • DM & Predictive Analytics • Agile BI Development Source Lowest Five: Hadoop, Data Science, BIaaS, Open Data, Visual design Standards Score < 5/10 Traditional Unstructured New Sources Data Sources Reporting
- 7. In-Memory Computing Stores information in the main “Quick-access RAM” of dedicated servers instead of complicated RDBMS structures operating on comparatively “Slow Disk Drives” Drivers of IMC • Cloud and Ubiquity • Business Systems + IOT + Unstructured Data Big Data • Information @ speed of light !!! IMC
- 8. IMC at work Carpe Diem Do it now!
- 9. Architecture • Database is memory-resident and not disk-resident • Database design is simpler • Fetching columns is simpler than querying from complex tables • High Availability and Fail-over mechanisms • RAM/ Battery RAM/ NVRAM [ferroelectric RAM (FeRAM), magnetoresistive RAM (MRAM) & phase change RAM (PRAM)] ensure almost zero-loss availability Top 3 IMDB: SAP HANA, Couchbase Server, MemSQL
- 10. IMC Use Cases • Internet of Things – using real time data from sensors and connected systems • Banking and Insurance industry – insights, claims and modelling • Artificial Intelligence – based on real-time machine learning and cognitive computing • Real time Advertising – based on consumer behaviour and availability / hyper-local • Outreach and product targeting – sentiment analysis through unstructured data • Image and Location Analytics – GIS, media and medical imaging Tests with Intel’s Nehalem architecture shows that latency of storage is reduced from 1 million CPU cycles to 250 cycles
- 11. © Sandeep Kumar kumarsahab@gmail.com https://www.linkedin.com/in/kumarsahab f https://www.facebook.com/kumarsaahab t @kumarsahab
Editor's Notes
- #7: Self-service BI – Casual users (70%), Power users (25%), Business Analysts (5%). Agile methodologies – iterative development, where requirements and solutions evolve through collaboration between self-organizing cross-functional teams. Data governance - overall management of the availability, usability, integrity, and security of the data employed in an enterprise.
- #10: Because their design is typically simpler than that of on-disk databases, IMDSs can also impose significantly lower memory and CPU requirements. Caching only speeds up retrieval of information, or “database reads.” Any database write – that is, an update to a record or creation of a new record – must still be written through the cache, to disk.