Hardware Deep Dive for BigData Processing, BigData in the Cloud

Scale-up vs. Scale-out, how massive parallel multicore systems are

  • Brief Introduction to CouchDB incl. Live - Demo of Cloudant - App deployed to Bluemix

  • Intro to DB2 and DB2 BLU vectorized column processing

  • Intro to MySQL as a Service

  • Intro to PostgreSQL as a Service

  • Intro to Cloudera as a Service

  • Intro to Cassandra as a Serviceshaping the BigData ecosystem

Hadoop workloads today rely on commodity x86 infrastructure, which although having cheaper costs, also provide lower reliability than UNIX Servers. IBM, through its proprietary POWER Architecture, wants to provide an alternative, aimed at achieving a higher performance at comparable prices. We will discuss IBM's Hadoop Reference Architecture, its performance, and the technical features of the POWER Chip, that enable Hadoop workloads to run faster, and achieve higher throughput per core.

On-Premise vs. Cloud: Free or highly competitive (BigData) data management solutions in the cloud: Column Stores, Document Stores, Hadoop

Open in Google Maps