Hybrid Hadoop: Using Cloud Compute with On-Premises Data - Seguridad Mania.com - España y América Latina

With datasets continuing to grow​ larger and cloud computing getting easier to set up and use, one of the biggest challenges facing IT departments responsible for managing Apache Hadoop environments is how to efficiently support and enable data storage. The main choice for moving data to cloud-based Hadoop deployments is to use HDFS, but efficiencies and processes have not gotten to a level that is both sustainable and repeatable while being affordable for IT organizations. In this 30-minute webinar, speaker Scott Jeschonek will look at the current best practices of using HDFS and Hadoop completely in the cloud. Next, he will highlight the limitations of this approach by contrasting classic on-premises Hadoop with Hadoop in the cloud, suggesting that the two are not the same. Finally, a new approach will be proposed that enables the use of Hadoop compute clusters while data is kept in its on-premises source instead of loading into the cloud. Participants will be able to consider a "Hybrid Hadoop" workflow that provides the economics needed for replication with more flexibility in the data source location, and the efficiency of eliminating the load process before cloud compute jobs.