In November 2022, I coauthored and published a book: ‘Data Fabric – An Intelligent Data Architecture for AI’. It discusses the value of the data fabric approach and architecture, along with some of the most common use cases organizations might encounter on their Data and AI journeys.
Getting started on that journey might seem a daunting task for some. To reduce any trepidation or anxiety, IBM recently announced a series of offerings to help organizations kickstart their journeys and build their data fabric over time while leveraging their existing infrastructure and investments.
I think of these products as ways to help organizations tackle their most pressing pain points first – and later expand their data fabric capabilities as and when needed. Indeed, these are in fact a subset of the same services offered with IBM Cloud Pak for Data – packaged up as a set of pre-built, preconfigured services and capabilities for easier consumption to address a particular challenge.
Let’s look at the three Express offerings first and the services they provide.
ELT Pushdown Express
Designed to help organizations quickly and easily build data pipelines through a visual interface, prebuilt connectors, and packaged transformations. Data flows are translated into native SQL. The intelligent pushdown technology automatically selects the best computational resource for each ELT flow, maximizing performance and efficiency with minimal effort. Organizations can deploy ELT Pushdown Express on-premises or on cloud. It offers:
- Support for more than 65 sources, making it easy to integrate data from any system
- Hundreds of prebuilt transformation functions, enabling users to build complex ELT flows and transformations in minutes
- Intelligent SQL Pushdown that selects the most optimal computational resource for each ELT flow
- Built-in visualizations to quickly plot and understand data
- Developer-centered interface with an open-source JavaScript Object Notation (JSON) format pipeline definition with APIs and command-line utilities for continuous integration and continuous delivery (CI/CD)
- No restriction on number of rows processed or number of users
Data Science and MLOPs Express
Designed to operationalize AI with an end-to-end data science & MLOps platform. Provides user self-service access to data with automated, integrated tools and processes designed to quickly and efficiently build, deploy and monitor models at scale. It offers:
- Support for team collaboration with built in project support
- Code, low-code, and no-code interfaces to enable analysis, data wrangling, model development, and reporting
- Best-in-class decision optimization
- Built-in visualizations to quickly plot and understand data
- One-click model deployment
- Visual orchestration capability to automate the Data Science and MLOps Express lifecycle
Data Governance Express
Designed to help organizations securely discover, list, and share data assets across the enterprise. As data is discovered, AI helps to more accurately and automatically classify it. An intuitive graphical user interface helps simplify data cataloging while offering transparency, efficiency, and trust. It offers:
- Automated discovery of data assets
- Classification of data as it is being catalogued, without user action
- Data sharing features to help users find and access data sets that are relevant to their jobs
- Security controls to protect sensitive data from unauthorized use
IBM AI Governance
In addition to Express, IBM also offers IBM AI Governance in a pre-configured software offering for common AI governance scenarios. It is designed to helps data science and enterprise risk teams quickly initiate and implement robust AI governance while simplifying the startup of model risk management and AI governance within an organization.
More information on these offerings is available from IBM here.
In closing, if these offerings have sparked your interest, or you want to find out more information on implementing certain scenarios within Cloud Pak for Data, I invite you to download ‘IBM Cloud Pak for Data Version 4.5: A practical, hands-on guide with best practices, examples, use cases, and walk-throughs’- a Redbooks publication which discusses implementing a data fabric using Cloud Pak for Data.