Big Data Evolution: The Big Data Platform Grows Up
TDWI Speaker: David Loshin, President of Knowledge Integrity
Date: Tuesday, October 28, 2014
Time: 9:00 a.m. PT, 12:00 p.m. ET
The big data software ecosystem has evolved into a robust framework for developing analytics applications spanning a wide range of complexity. At the same time, big data deployments more commonly center on the platform as an expansion of the corporate file system. The concept of the data lake resonates with enterprises desiring to offload data assets into a common platform for analysis, yet the analyses often remain batch-oriented—summarizations, aggregations, and other ETL-like tasks.
As software ecosystems mature, greater demands are placed on the infrastructure platform. For example, enterprise requirements for interactivity and low latency have given rise to a bevy of SQL-on-Hadoop projects, in-memory engines, and stream processing capabilities. Business users are increasingly looking to big data platforms for high performance execution: massive parallelism, real-time response, and in particular, interactive applications instead of sequences of batch processes.
In this talk we discuss different ways that big data is growing up. Specifically, we look at the emerging trends in platform configurations to meet the need for “supercomputing-class” analytics. Attendees will learn about:
You will learn:
- Relevant innovations in Hadoop 2.0, Spark, and other emerging technologies
- The requirement for fast response times and minimal data movement for analytic productivity
- Exploiting the memory hierarchy for increased performance (caches, in-memory computing, fast interconnects, and SSD)
- Motivating balance between storage nodes and compute nodes
- The evolving use cases for integrated big data analytics