Big Performance Gains for Big Data

Categories: Analytic Database Enterprise Data Hub Partners

Intel, Oracle, and Cloudera collaboration pays off for easier, faster big data analytics

By Prabha Ganapathy

Imagine two teams of data analysts working on the same goal: to extract usable business intelligence (BI) from massive, growing data sets.

One decides to build a do-it-yourself (DIY) Apache Hadoop cluster. They have in-house expertise and think the DIY approach will help control costs. Unfortunately, it takes them weeks to nail down their hardware and software needs and manually deploy and configure everything. And that doesn’t even cover the additional days and dollars spent fine-tuning the cluster for performance.

The other team deploys a preconfigured Oracle Big Data Appliance, which means they’re running queries long before the other team has finished building and configuring their DIY cluster. Plus, their queries are running at nearly double the speed of the team with the DIY cluster.

So what accounts for the big difference between the two solutions? It comes down to collaboration. Joint engineering between Intel, Oracle, and Cloudera is bringing us optimized technologies, like the Oracle Big Data Appliance, that can deliver amazing performance and value.

Joint Expertise, Built-In

I recently attended the JavaOne conference where Mike Olson, co-founder and Chief Strategy Officer of Cloudera, joined Intel VP Michael Greene on stage to talk about that collaboration. Mike pointed out how closely Cloudera has been working with Intel and Oracle to optimize throughput and speed for their software running on the Oracle Big Data Appliance. The joint efforts have really paid off, with significant performance increases over a do-it-yourself cluster.

Intel-Oracle_BDA_Blog_Post_Final_-_Google_Docs

In fact, when Intel engineers recently tested the Oracle Big Data Appliance against a DIY cluster, they observed nearly two times faster performance for the Oracle appliance after general tuning, and nearly three times faster performance after workload-specific tuning. Those gains are due to the tight integration between software and hardware within the appliance. Powered by Intel Xeon processors, the Oracle appliance makes full use of integrated Intel technologies for fast, efficient processing, data compression, and data encryption.

I’m looking forward to seeing what comes next from this partnership, as the three companies continue to provide increasing performance gains, along with affordability and stronger security.

Take a look at this white paper Intel Performance Study on Big Data Appliance vs. DIY Hadoop for details on the Oracle appliance test results and performance differences. And check out this blog post from Jean-Pierre Dijcks at Oracle on this effort and on new products coming next year.

Follow me on Twitter at @PrabhaGana

facebooktwittergoogle_pluslinkedinmail

2 responses on “Big Performance Gains for Big Data

Leave a Reply