Written by Vikram Murali , Director, Integrated Analytics Systems Development, IBM Data and AI
Originally posted May 22, 2019
Last year, IBM announced IBM Cloud Pak for Data, a robust, end-to-end solution for data and analytic needs within an enterprise. One of its design principles was to help organizations access a vast array of data sources on-premises and in the cloud—all while applying IBM’s deep data management and analytics within a private cloud setting.
As a refresher, IBM Cloud Pak for Data is an integrated collection of data and analytics microservices built on a cloud-native architecture. From a software perspective, it’s pre-configured and requires no assembly. It helps organizations collect, organize and analyze their data—ultimately preparing them for AI infused applications. In essence, IBM Cloud Pak for Data is a multicloud data and AI platform delivering an information architecture for AI.
What we’ve learned from AI engagements is that every step of the ladder is critical. AI is not magic. It requires a thoughtful and well-architected approach. Success with AI models is dependent on achieving success first with how you collect and organize data.
The approach of preconfigured software services helps remove the need for organizations to perform complex and time-consuming software integration themselves and allows them to install it on their own hardware. However, some companies may not have the time or resources to perform the typical hardware and software infrastructure tasks such as building network infrastructure, or installing and patching an operating system, to name a few.
Introducing IBM Cloud Pak for Data System version 1.0
Today, we’re announcing IBM Cloud Pak for Data System, an Intel x86-based, hyper-converged data and AI platform. It contains all the IBM Cloud Pak for Data capabilities mentioned above while removing the burden of businesses having to configure hardware and software infrastructure. The system combines storage, compute, networking and software into a single system to help further reduce complexity and increase horizontal scalability.
AI in IBM Cloud Pak for Data System
One of the key aims of IBM Cloud Pak for Data is to deliver AI capabilities out-of-the-box. For instance, Watson Studio is included in the configuration. Watson Machine Learning and Watson Machine Learning Accelerator are available as add-ons. It also comes with a comprehensive, end-to-end data science toolkit that helps data scientists of all skill levels to:
- Prepare data
- Build AI models, both machine learning (ML) and deep learning (DL)
- Train AI models, either through an interactive or batch paradigm
- Deploy and manage lifecycle of models
- Enable GPU acceleration to train models as well as leverage GPUs and field programmable gate arrays (FPGAs) for inferencing
- Scale to enterprise-wide deployments of data scientists and ML/DL models
Other components in the offering include ML-based serviceability and Db2-augmented data. The system will include Watson Assistant as an add-on in June.
Why you should explore IBM Cloud Pak for Data System
IBM Cloud Pak for Data System is an integrated, end-to-end platform that is cloud-native by design, architected as microservices and containerized workloads. It offers instant, pre-assembled provisioning and has capabilities to collect, organize and analyze data.
It takes the IBM Cloud Pak for Data experience further by providing a modular approach to compute, network and storage on standard hardware—a building-block approach under unified management.
Built on a philosophy of plug-and-play, it also provides “plug-and-grow” capabilities, allowing bare metal compute and storage nodes to be added, recognized and provisioned to help meet the need of the business. Tests in IBM labs have shown that the IBM Cloud Pak for Data System can be up and running in less than four hours. Built-in data virtualization enables access to sources without moving data. It can help businesses gain insights into their enterprise data within minutes with a single, unified console that helps to provide a seamless user experience of data and AI capabilities.