With the emergence of Big Data and AI analytic use cases, Enterprise IT is striving to transform and modernize existing infrastructure to support new business models. In this presentation, we will discuss how an Open Infrastructure architecture can be utilized for this type of next generation workloads. With the upcoming releases, the community is adding many new features including support of persistent memory. We will discuss our latest contribution to OpenStack and how we integrate these new features to build an open infrastructure for Big Data.
In this talk, we will first discuss the requirements of Big Data applications. Then we will share our experience on how to build and optimize an OpenStack Private Cloud for such workload. We will illustrate the OpenStack software architecture as well as the underlying hardware reference design for Big Data workload. We will describe the benchmark methodology and share the performance improvement on such implementation.
The audience will learn:
1. General requirements for Big Data application
2. OpenStack software and hardware architecture for Big Data
3. Big Data benchmark methodology
4. Performance improvement over persistent memory architecture
