AI and Analytics at Scale

Chapter 5. Example Data Infrastructure: HPE Ezmeral Data Fabric

Just as the new technology of curtain walls and steel beams completely changed the way buildings could be constructed, giving rise (literally) to skyscrapers, so too does new technology for data infrastructure support new ways of working with AI and analytics at scale in production systems.

HPE Ezmeral Data Fabric is an example of a data technology designed to support these approaches, and the real-world use cases we have described throughout the report use this data fabric as the data infrastructure. Through these examples, we have described how certain capabilities are required of a data infrastructure to enable scale-efficient systems. In this chapter, we explain how HPE Ezmeral Data Fabric meets those needs.

Of course, data infrastructure is only one part of what you need to achieve scale efficiency. There are a number of other technological frameworks that help build systems like the ones we have described. We have previously mentioned Kubernetes, KubeDirector, and Kubeflow, which help you orchestrate and manage computation (particularly containerized applications), and help you build efficient workflows for analytics and AI. We have also mentioned SPIFFE, an open source framework that helps you manage identities across trust boundaries. But if your data infrastructure doesn’t provide key capabilities, you will find it difficult to follow the design recommendations we’ve made in this report. So let’s consider ...