Cloud Pak for Data
Published in

Cloud Pak for Data

Getting Started with IBM Cloud Private for Data with IBM Power Systems

As artificial intelligence (AI) capabilities mature, enterprises are continuously evaluating use cases that can transform their business. A key challenge that slows down AI adoption is the abundant but untamed data that is not ready for AI. There is a high correlation between companies that are outperforming in AI adoption and the ones that have a robust data infrastructure aligned with their business architecture. According to the 2018 IBM Business Value survey, shifting towards enterprise grade AI, 65% of outperformers surveyed capture, manage and access business, technology and operational information on key corporate data with a high degree of consistency across the organization versus 52% of all others surveyed.

To that end, IBM recently introduced IBM Cloud Private for Data (ICP4D), a data and analytics platform, to help make your data estate ready for AI. It simplifies how you collect, organize and analyze your data in a cloud native platform and feed your AI models with the data they need. It’s comprised by a set of well-integrated microservices that employ a cloud-native architecture and provides a robust end-to-end platform for all of your enterprise data and analytic needs. Now you can collect, organize and analyze your data all in a single platform, as shown below:

IBM Cloud Private for Data Platform

And now IBM Cloud Private for Data can be deployed on IBM Power®Systems™ — the purpose-built server infrastructure for data intensive and AI workloads! Power Systems provide an excellent compute fabric for IBM Cloud Private for Data with their industry-leading ability to move information between high-bandwidth interconnects as well as their ability to exploit industry-leading NVLink 2.0 interconnects between CPUs and GPUs — a critical factor for today’s AI, machine learning and deep learning applications.

Who are the actual users of IBM Cloud Private for Data? Let’s break it down quickly:

1. Data Engineer

a. Create new Db2 (and more) relational databases and warehouses

b. Facilitate cross-cloud hybrid data access and movements

c. Build data movement and transformation flows to prepare data for other users to consume

2. Data Steward

a. Build an enterprise catalog through auto-discovery and classify existing data sources

b. Tag and annotate data sets and other assets, index for search — making assets easy to find

3. Data Scientist

a. Find data of interest across all sources within the enterprise

b. Explore, visualize and understand data

c. Share analytic assets with others and publish into the enterprise catalog

d. Train machine learning models and deploy scoring services

4. System Administrator

a. Quickly provision IBM Cloud Private for Data with the complete set of capabilities

b. Allocate and manage resource usage and scale environment as needed

c. Monitor the ongoing health of the system

While there are an endless number of scenarios in which IBM Cloud Private for Data can be used, there are four principle scenarios to highlight:

1. Manage Data Anywhere
Now you can manage all of your data regardless of where it lives via data virtualization. You can also gain control and leverage your data from connected devices (imagine streaming analytics).

2. Operationalize Data Science and AI
Build, deploy, manage and govern data models and data at scale to improve your business outcomes.

3. Shift to Next-Gen Workloads
Shift development methodologies to cloud-native practices — provision and scale in minutes, build once and deploy anywhere in the multi-cloud world and inherent automation and collaboration to increase your productivity.

4. Smarter Governance
Auto-discover metadata, manage governance rules and policies and enforce privacy to mitigate risk and ensure compliance (e.g., GDPR).

The common theme of these use cases… is your data! And given that your IBM Power Systems house all of your enterprise’s critical data, there is simply no better pairing for your data than IBM Cloud Private for Data. Whether you’re running Oracle, Db2 or open source databases — IBM Cloud Private for Data has you covered. You can immediately start operationalizing this data with transformations, deep analytics, artificial intelligence model creation and more (check out the IBM Cloud Private for Data content hub for the latest and greatest information)!

IBM Cloud Private for Data is a private cloud analytics platform that leverages IBM Power Systems for their superior processor, memory and accelerator technologies. Shown below is the heterogeneous platform reference configuration to run IBM Cloud Private for Data with IBM Power Systems (along with the necessary x86 worker nodes to run ancillary services that require an x86 compute platform).

Logical Architecture

The figure below depicts a reference configuration on scale-out systems to realize the above logical architecture. This reference configuration runs with a highly available master node topology to tolerate a physical server failure within the overall IBM Cloud Private for Data deployment. While the configuration below depicts a scale-out model, it’s important to highlight that IBM Cloud Private for Data can also run atop any enterprise Power System as well (i.e., systems running PowerVM).

Reference System Configuration

The configuration above is adequate for up to approximately 15 users who are working on analytic projects and can be used for production deployments. For more complex workloads (e.g., ETL), the environment should be adjusted accordingly to maintain effective concurrency levels. For example, if you are planning to deploy an integrated Db2 Warehouse instance, you should add appropriately sized worker nodes to accommodate the additional workload and scale requirements. For more detailed system configuration information, refer to the IBM Cloud Private for Data system requirements (if you are exploring a proof-of-concept deployment, you can consolidate the management and worker nodes down into five total nodes — i.e., three ppc64le nodes that are dual-purpose master and worker nodes and two x86–64 worker nodes).

IBM Cloud Private for Data provides a single, enterprise-wide fabric for aggregating information. IBM Cloud Private for Data operates across a heterogeneous compute fabric and can interface with data residing on existing enterprise platforms (e.g., Oracle atop AIX on Enterprise Power Systems). It also exploits POWER’s architectural strengths in AI — unmatched by any other computing platform. Last, but certainly not least, it is the only solution in the market that provides an end-to-end analytics workflow for Enterprise AI (as shown below) that gives you all the tools you need to unlock your data to fuel a new wave of innovations.

End-to-end Analytics Workflow

Try IBM Cloud Private for Data today by leveraging our hosted trial environment! For more information about planning for, installing and operating IBM Cloud Private for Data, you can visit the online Resource Center. You can also check out the IBM Cloud Private for Data online community, which contains a wealth of information and access to a thriving ecosystem of other IBM Cloud Private for Data users. For additional questions or comments on how to start your journey into AI with IBM Cloud Private for Data, please contact your local IBM sales representative or business partner today!



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Joe Cropper

Joe Cropper

Passionate about the intersection of technology and business to make the world a better place. Oh, and cars too! Opinions are my own.