By Ian Lumb | January 13, 2015 | CMGUI, Bright Cluster Manager, Hadoop, Hadoop Cluster Management, Hadoop Summit, Bare Metal Provisioning, Bright Computing, Application Performance, Apache Hadoop, Apache Spark, Hadoop Analytics Stack, Big Data Analytics, CDH
The Summit features 6 tracks. By tapping the expertise of our Hadoop braintrust, we’ve submitted 4 ideas to 3 different tracks. Because we’d like to earn your votes, please allow us elaborate.
Hadoop Governance, Security & Operations Track
Our Hadoop engineers Denis Shestakov and Michele Lamarca made two submissions to this track.
Wrangling choice is the emphasis of Distro-Independent Hadoop Cluster Management. This is one of the primary areas where Bright Cluster Manager adds value in the Hadoop arena - by allowing customers to choose a Hadoop distribution suitable to their needs. Using the GUI provided by Bright, we’ll illustrate how a Hadoop distribution can be configured and managed with ease on an ongoing basis.
Hadoop performance is the focus of our Automatic Deployment of Cost-effective Hadoop Clusters submission. Because Bright Cluster Manager for Apache Hadoop starts at bare metal, and then progresses up the stack to the Hadoop distro and analytics apps, we’re able to assert a holistic approach for managing performance. To demonstrate, we’ll share an approach that is mindful of all resources in automating configuration and tuning for optimized Hadoop clusters.
Data Science & Hadoop Track
Our solutions architect, Robert Stober, submitted Next-Generation Management Solutions for Big Data Analytics: An Illustration Using a Real-World Use Case to this track. Deployment of CDH on Lustre filesystem, not HDFS(!), makes this use case unique and intriguing. Working alongside Dell and Intel at Tulane University, the TOP500-ranked Cypress hybrid supercomputer is what makes this use case very, very real.
Applications of Hadoop and the Data Driven Business Track
By tapping into my own background in geophysics, I recontextualize seismic processing via Reverse-Time Migration (RTM) in the submission RTM Using Hadoop: Is There a Case for Migration? Computationally intensive RTM aims to remove unwanted artifacts from seismic data so that targets of economic interest (e.g., petroleum reservoirs) can literally be brought into focus. Although Big Data Analytics has been applied successfully in other areas of seismology, using it in the context of RTM appears to be novel.
That’s it. 4 submissions spanning 3 tracks. As you can see, we’re really embracing the elephant. And if you feel it’s warranted, we’d appreciate your vote of support. Please vote here.