IBM Cloud Pak for Data System
By 2024, with proactive, hyperspeed operational changes and market reactions, artificial intelligence (AI)-powered enterprises will respond to customers, competitors, regulators, and partners 50% faster than their peers. These digital transformation (DX) initiatives will be supported by artificial intelligence (AI) capabilities, providing timely critical insights, richer and immersive user experiences, and improved business outcomes.
IDC forecasts that global AI spending will reach $97.9 billion by 2023, driven mostly by deployments
in banking, retail, and manufacturing. However, AI adoption has been slow. Automated customer
service agents, IT automation, sales process recommendation, and automation are the current top
uses, but we expect to see automated human resources, digital assistants for enterprise knowledge
workers, regulatory intelligence, and advanced digital simulation emerge as the fastest growing use
cases over the next five years.
1. ML training. ML training consists of the steps required to build the ML model and can
include model generation, model build, and model fit.
2. ML inference. ML inference (i.e. prediction, scoring, or model serve) generates the
insights that need to be integrated into a business use case, creating an ML business
application that ultimately generates customer value.
Figure 1. Worldwide Global DataSphere (102.6ZB by 2023)
Figure 2. Current Reality of AI Deployments
Figure 3. Current Reality of AI Deployments Top Factors Holding Back AI Deployments
Figure 4. Data and Deployment Tasks are Time Consumings
Core Capabilities of IBM Cloud Pak for Data System
The IBM Cloud Pak for Data System provides a modular approach to compute, network, and storage
on standard hardware. Its core capabilities include:
• Red Hat OpenShift support which is certified across IBM Cloud Pak for Data services
• Open source governance capabilities for managing risks and accelerating open source-based AI
projects delivery
• Automated development of AI models supported by AutoAI
• Built-in data science and machine learning
• High performance analytics (powered by IBM Performance Server for PostgreSQL, which is 100%
compatible with Netezza) in “cloud-in-a-box” setup to take advantage of hyper-converged
modularity
• Visual application building, near real-time visual debugging, and support for Red Hat AMQ
Streams
• New industry-specific accelerators
• New offerings called packages which includes IBM Cloud Pak for Data entitlements that are
required to run the service
• New licensing model for IBM Cloud Pak for Data that enables businesses to purchase licenses to
IBM software in a cloud-centric model aligned with a Red Hat subscription model to deliver a
consistent buying experience across both product portfolios
• Extensible third-party services like Figure Eight to help annotate training data and fuel machine
learning initiatives.
Continue Reading
Get this whitepaper right in your inbox for free. We never spam!
