Advanced Computing in the Age of AI | Monday, July 15, 2024

IBM Expanding AI, ML Capabilities in IBM Cloud Satellite Through the Cloud Pak for Data Add-On 

Image credit: Shutterstock

With the recent release of its IBM Cloud Pak for Data, IBM is quickly expanding the tools it offers to enterprise customers that want to put their workloads onto IBM’s Cloud Satellite.

The IBM Cloud Pak for Data, which became generally available to customers on March 26, is the first major add-on for Cloud Satellite since the nascent service was launched on March 1. IBM Cloud Satellite lets enterprise users build, deploy and run workloads wherever their data resides, using any public cloud, any on-premises servers or at the edge, without having to manage the resulting complex configurations on their own.

Cloud Pak for Data is basically our data and AI capabilities, including our Watson capabilities,” Jason McGee, an IBM Fellow and vice president and CTO for IBM Cloud Platform, told EnterpriseAI. It’s built to help customers that are performing data analytics and machine learning on data, but who want to quickly switch it to run on another cloud such as Amazon S3, on a bare metal server or anywhere else whenever they want to make the change, he said.

Jason McGee, IBM

“Before, if you want to do analytics and AI [elsewhere] on that data, you would have to send a copy of that data to IBM Cloud to do that processing on it,” said McGee. “With this new capability you can run your analytics notebook on the data in Amazon S3 by literally moving the runtime to Amazon S3” with a few clicks. “You get better latency, you don't have to [manually] shift the data around, and you don't have costs for doing egress with data from Amazon to some other place. Cloud Satellite is what powers that capability.”

Behind the scenes, Cloud Satellite makes this happen by using as-a-service IBM Watson capabilities to run the data’s notebook runtime in an Amazon Web Services (AWS) data center while being managed from the IBM cloud, according to McGee. “That runtime can physically run that data in the same data center, even though it is sitting in something like S3. And now you basically have local access to the data.”

Ed Anderson, a cloud service market analyst with Gartner, told EnterpriseAI that IBM’s Cloud Satellite approach addresses a growing customer need to run cloud services in locations where the demands of specific application scenarios can be met.

Ed Anderson, Gartner

“There are many applications that must run in specific locations due to latency, compliance, or data sovereignty issues,” said Anderson. “As organizations seek the benefits of cloud for location-sensitive applications, distributed cloud offerings such as IBM Cloud Satellite will be required.”

Another analyst, Bob O'Donnell, the founder and chief analyst at TECHnalysis Research, agrees, saying that services like these can be helpful especially to longtime IBM customers that may not be using private cloud options from AWS or Microsoft Azure on their own.

Bob O'Donnell, TECHnalysis

“It fits a particular need,” said O’Donnell. “It's not for everybody, but it also opens up IBM Cloud to a wider group of people who simply wouldn't go with any cloud solution that didn't let them run something remotely. It fills a critical gap they needed to fill,” while helping in some industries which IBM traditionally serves, such as financial services and healthcare. “This allows customers to do whatever they want to do and not have to deal with a zillion other places, but do it all through IBM Cloud Satellite, which might be more comfortable for them.”

IBM is delivering its Cloud Satellite through a partnership with enterprise technology platform vendor Lumen Technologies, which is using its global Edge Compute platform that features more than 180,000 connected enterprise locations on the Lumen network. In addition, IBM Cloud Satellite also includes collaborations with more than 65 other ecosystem partners, including Cisco, Dell Technologies and Intel to assist customers.

Those Cloud Satellite options could be helpful for smaller customers who may not have had relationships with these partner vendors or even with IBM in the past, said O’Donnell. “They may have been interested by some of the services that IBM offered, but they didn't feel like they had a way to get there,” he said. “This gives them an on-ramp to get there. And because they're partnering with a bunch of other companies to provide services and other capabilities that they may not have had before give it more of a solution-like feel.”