Enterprises are looking for fresher data – from daily, to hourly, to real-time – as well as access to data from more sources and for longer periods of time. And they need it faster and cheaper. Meanwhile, traditional approaches for transforming data in the data warehouse (ELT) can’t keep pace; retention windows are shrinking to as little as eight weeks; and, as data volumes grow, IT is forced to make cost and performance trade-offs.
Many organizations have discovered that shifting ELT to Hadoop dramatically reduces costs, allows them to incorporate new data faster, and frees up data warehouse capacity for faster analytics and end-user response times.
This presentation will outline a three-step framework for optimizing the data warehouse with Hadoop and demonstrate an end-to-end approach that takes you from data integration to data discovery and visualization with the click of a button. No trade-offs, no compromises.
1.Identify, analyze and document ELT workloads. SQL is still widely used for data integration. In most cases, 20% of data transformations consume up to 80% of resources. Learn how to identify and understand these workloads – including SQL ELT – best suited to be offloaded into Hadoop.
2. Access and shift heavy ELT workloads to Hadoop. Learn how to access virtually any data, from any source; and optimize the data warehouse by quickly and securely translating expensive workloads to efficient MapReduce processes, without coding.
3. Optimize & secure the new environment. Learn how to meet these key requirements with file-based metadata, support for Kerberos and LDAP, automated cluster management, and advanced monitoring tools. Now, you’re ready for new insights with data discovery and visualization.
Jorge A. Lopez, Director, Product Marketing, Syncsort
David Tishgart, Director, Product Marketing, Cloudera
Nikhil Kumar, Technical Product Manager, Syncsort