Think IBM's latest attempt at relevance in the cloud world continued at its Think conference by giving its Cloud Pak for Data another beating with the AI stick and unleashing Watson on IT pros.
AutoSQL with Cloud Pak for Data
The IBM Cloud Pak for Data turned up three years ago and was geared towards collecting and organising data for AI analysis. Aimed at pretty much any hybrid cloud environment, Big Blue targeted enterprises with a need to bring diverse data together and make sense of it all.
Dealing with disparate data sources has been taxing many in the community, and behemoths such as Microsoft have had their own crack at it with the likes of Azure Synapse. For IBM, it's an opportunity slap on the "Watson Anywhere" branding and what it describes as an "intelligent data fabric."
To that end, the biz is rolling out AutoSQL – a technology aimed at automating access, integration and management of data for AI without having to move it. The same query engine is used over multiple data lakes, warehouses, and streaming data. Unsurprisingly, IBM also has a bunch of benchmarks to show just how whizzy its tech is when compared to the competition.
IBM reckons its tech handily spanks the behinds of AWS Snowflake and Azure Synapse on both performance and price/performance measures although, as with all such metrics, we'd recommend a good hard look at the type of loads you're likely to put through it.
As well as AutoSQL, other new capabilities heading to Cloud Pak for Data include AutoCatalog, to maintain a real-time catalog of data assets and their relationships, and AutoPrivacy, to automate the identification of and enforce policies on sensitive data.
Perhaps more interesting for end users is the arrival of Watson Orchestrate, aimed at allowing business folk to use natural language for interaction and automation "without needing any IT skills."
Built on Red Hat OpenShift (and so able to run on more places than the competition), IBM's vision has a user, for example, asking Watson Orchestrate to keep an eye on business opportunities and alert (and set up meetings) when a sale progresses, while also understanding and learning context based on a user's prior interactions.
- IBM says it's built the world's first 2nm semiconductor chips
- IBM has another crack at HCI, analysts say container-centric software-defined storage approach could shake things up
- AWS on track to be bigger than IBM by Christmas, once Kyndryl is spun out
- IBM teases new AIX release – the first since 2015
Applications, tools, data and history are connected by the system, which relies on a set of pre-packaged skills needed to perform a given task. It all sounds a little creepy to us, and not unlike some of the technologies available in the Microsoft 365 world.
Watson Orchestrate is currently in preview in the IBM Cloud Paks for Automation.
Gartner analyst Chirag Dekate told us: "An AI-infused Cloud stack enables intelligent automation of technology and business process functions for enterprises seeking to accelerate productivity and value gains from their cloud investments.
"IBM differentiates its cloud strategy by infusing Watson AI across its automation framework used across multiple Cloud Paks. IBM's AI and Cloud strategy potentially enables value capture and maximization from secure, enterprise-grade integrated applications on top of the foundational hybrid and multicloud IT Estate."
Big Blue has quite the hill to climb. According to Gartner's figures, its share of the Worldwide IaaS Public Cloud Services Market was a mere 1.7 per cent in 2019, while its showing in the Cloud Infrastructure and Platform Services Market stood at 2.3 per cent. Both percentages were down on 2018, although revenues had risen.
A September poll run by Gartner had 75 per cent of respondents planning to continue or start new AI initiatives in the next six months, although security and privacy concerns, complexity (particularly with existing architectures), and data volume were highlighted as barriers in a July 2020 survey. ®