IBM Brings Watson AI to Any Public, Private Cloud

IBM THINK — IBM is making its Watson artificial-intelligence technology available on public clouds from such rivals as Amazon Web Services, Google Cloud and Microsoft Azure, a move that expands Big Blue’s efforts in the areas of hybrid cloud and multicloud.

The giant tech vendor is making Watson services like Watson Assistant and Watson OpenScale, which before had only been available on IBM’s own public cloud, more widely accessible by integrating them with IBM Cloud Private for Data (ICP for Data). Enterprises will be able to run the technology not only in public clouds but also in their own private clouds or on data-center servers.

IBM’s Ken Parmelee

“Through the integration of key Watson services with IBM Cloud Private for Data … we’re making it easier for organizations to bring AI to wherever their data resides,” Ken Parmelee, executive with IBM’s cloud integration business unit, told Channel Futures. “As a result, more businesses will be able to begin unleashing the power of AI and its ability to predict outcomes, automate and optimize business processes for strategic, data-driven decision-making.”

IBM announced the Watson strategy this week at its Think 2019 user conference, one of several moves the company made regarding the cloud, most focusing on hybrid and multiclouds. The company is pursuing an aggressive strategy of helping customers move their data and applications between their on-premises environments and multiple public clouds by leveraging such open technologies as Kubernetes. IBM’s proposed $34 billion acquisition of Red Hat also is a key part of the company’s larger hybrid and multicloud efforts.

“IBM knows firsthand that a one-cloud-fits-all approach doesn’t work for most companies,” Parmelee said. “Rather, most are using a hybrid-cloud strategy comprised of a combination of public, private and on-prem capabilities. Most of our enterprise clients are now entering chapter two of cloud, which is focused on modernizing and transforming the critical, more complex workloads and applying high-value cloud services such as AI and analytics.”

A key is being able to easily and quickly move applications and data between clouds and businesses that “need choice and flexibility so that they don’t find themselves locked-in in the midst of their cloud journey. They need to do all of this efficiently and securely,” Parmelee said.

One hurdle for enterprises that want to scale AI has been moving the data to where the AI service is. Now customers can leave the data in whichever cloud environment they’re most comfortable with and bring Watson to the data. ICP for Data is an open and cloud-native architecture for AI that includes integrated data science and data-engineering features, as well as application-building capabilities. The openness of the architecture is important; IBM estimates that 85 percent of Watson customers use open-source languages and frameworks, including Python, R and TensorFlow.

The Watson microservices available on ICP for Data are based on Kubernetes. Watson Studio and Watson Machine Learning are among those already available on the architecture. The new ones are Watson OpenScale for managing multiple instances of AI regardless of where they were developed, and Watson Assistant for building conversational interfaces into applications and devices. Later this year, IBM will add more Watson services to ICP for Data, such as …