Modernizing the Operational Data Store with Hadoop
TDWI Speaker: Philip Russom, TDWI Research Director
Date: Thursday, January 29, 2015 (moved from 12/4/14)
Time: 9:00 a.m. PT, 12:00 p.m. ET
Operational data stores (ODSs) are currently experiencing a dramatic evolution, as are many data platforms and practices within data warehousing and enterprise data management. The evolution of the ODS is driven mostly by users’ increased usage of big data and advanced analytics, but also by changing practices in data archiving, data staging, and data integration. The result is that ODSs today manage greater data volumes, handle more diverse data, and serve more practical uses than ever before.
As users expand and modernize their operational data stores, many are migrating ODSs to Hadoop. Most ODSs (regardless of specific use) involve simple data models (e.g., transaction records in tables) but in massive quantity; these are easily ingested by Hadoop. Processing ODS data (e.g., rescoring performance metrics) is likewise straightforward but at scale, which Hadoop handles ably. Many ODSs are designed to be row stores and live archives, which again are a good fit for Hadoop. Furthermore, ODSs are prone to proliferation (akin to data marts), and Hadoop can be the point of consolidation for redundant and rogue stores, but strong governance capability will be required. Finally, new practices like data lakes and enterprise data hubs can operate as one giant ODS or as a collection of ODSs.
In this Webinar, you’ll learn:
- Why operational data stores need to be modernized and how some users are doing it
- How ODSs still support data warehouse architectures, but also support new approaches to advanced analytics, data archiving, data lakes, and enterprise data hubs
- How Hadoop makes a highly scalable and cost-effective platform for ODSs, especially when compared to traditional relational databases
- Learn how to industrialize the data integration process for a data hub
Philip Russom, Ph.D.