The ideal cloud-based AI and data analytics service allows data scientists to configure their cloud compute, networking and storage resources simply and quickly without becoming Kubernetes experts.
Through our collaboration with Intel, Rakuten Cloud empowers enterprises and communications service providers to build and manage the lifecycle of these services with just a few clicks. To demonstrate just how easy it is to set up and manage a service, we’ve created a recap video of our joint presentation at KubeCon North America.
In the video, Jerry Huang, Partner Scale Account Manager, SMG Intel, and Abir Bhattacharyya, Principal Engineer, SQA, Rakuten Cloud, demonstrated how the private cloud for data analytics services is built on the Rakuten Cloud-Native Platform (CNP).
Rakuten CNP is an industry-leading Kubernetes platform that is optimized for running storage and network intensive AI and analytics applications such as data lakehouses. Other components in the platform include Rakuten Cloud-Native Storage (CNS), which provides app-aware storage for the variety of data types that come in a data lake environment. Rakuten Cloud-Native Orchestrator (CNO) provides automation and management features.
The software runs on servers based on 5th Gen Intel® Xeon Processors. This processor family has CPUs that offer between eight and 64 cores per processor, catering to diverse workload requirements. One CPU feature called out in the video is Intel® Advanced Matrix Extensions (Intel® AMX) that improves performance of deep learning training and inference on the CPU.
Rakuten Cloud leveraged these technologies to establish a private cloud for data analytics as a service blueprint, capable of transitioning from bare metal infrastructure to full AIOps in minutes.
To learn more about this solution, check out the related solution brief detailing how Intel and Rakuten Cloud developed an analytics service for a Tier 1 service provider.