Designing Shared Storage for Hadoop, Elastic, Kafka, TensorFlow

Designing Shared Storage for Hadoop, Elastic, Kafka, TensorFlow. As analytics environments like Hadoop, Elastic, Kafka and TensorFlow continue to scale, organizations need to find a way to create a shared infrastructure that can deliver the bandwidth, flexibility, and efficiency that these environments need. In a recent Storage Intensity podcast, Tom Lyon, founder and chief scientist of DriveScale and George Crump, Lead Analyst of Storage Switzerland, sat down to discuss a wide range of subjects including Non-Volatile Memory express, (NVMe), Non-Volatile Memory express over Fabric (NVMe-oF), and Composable Infrastructures.


Is NVMe Enough for Efficient Hyperscale Data Centers?

Hyperscale architectures typically sacrifice resource efficiency for performance by using direct attached storage instead of a shared storage solution. That lost efficiency though, means the organization is spending money on excess compute, graphics processing units (GPUs) and storage capacity that it doesn’t need.

Load More Posts