WavebreakmediaMicro - Fotolia
VMware Explore in Barcelona this week saw the virtualisation pioneer continue its push to go all-in on generative AI (GenAI) as a core capability for enterprises. This saw VMware extend its vision of a hybrid cloud and multicloud future built around generative AI and applications that can run in locations that include private cloud, public cloud and the “cloud edge”.
Core to the announcements were new partnerships with Intel and IBM watsonx to add capabilities for the VMware Private AI framework, revealed at its US Explore event in August. These will comprise reference architectures and hardware compatibility to deliver GenAI infrastructure. VMware Private AI allows customers to deploy GenAI infrastructure on-premise and in the cloud via its VMware Cloud Foundation suite.
VMware Private AI with Intel will allow customers to make use of the latest Intel Xeon CPUs with built-in AI processing, as well as Intel GPUs. The move will also see IBM watsonx capability added for customers that want to run it on-premise via VMware Cloud Foundation.
These moves come in addition to existing Nvidia and open source reference architectures launched at VMware Explore in the US.
VMware has in mind that generative AI is a growth area for software development, customer support and content generation, with an estimated annual economic value of $44tn, according to figures it quoted from McKinsey research.
VMware CEO Raghu Raghuram said generative AI and multicloud were “the two key things” in mind for VMware.
“Multicloud is the norm. Three out of four of you use more than one cloud, and some are using the edge cloud too,” said Raghuram. “We talk about being ‘cloud smart’, to let application requirements dictate where to deploy, whether in the private cloud or public.”
VMware Private AI hopes to reap some of that $44tn by providing secure and compliant AI capabilities to enterprises via its VMware Cloud Foundation for cloud and on-premise deployments.
Read more about AI and cloud
- How generative AI and cloud complement each other: McKinsey partner explains the symbiotic relationship between generative AI and cloud, enabling organisations to speed up cloud migration and harness the benefits of AI.
- Putting artificial intelligence and machine learning workloads in the cloud: We look at the pros and cons of putting artificial intelligence and machine learning applications in the cloud.
VMware Cloud Foundation unifies the management of compute, storage, networking and security across private, public and edge clouds, while supporting modern application development and multicloud operations.
VMware president Sumit Dhawan said the company had experienced 30% year-on-year growth in VMware Cloud Foundation and that “no other cloud provider has seen that”.
In Barcelona, VMware announced additions to VMware Cloud Foundation, which reached version 5.0 at its Explore event in the US. It is now on version 5.1, with additions that are said to bring two times CPU capacity, four times storage performance and enhanced self-service capabilities.
Key among the storage upgrades is the addition of VMware cloud-based vSAN Max connectivity to hardware from Dell EMC, HPE, Lenovo, Cisco, SuperMicro and XFusion.
vSAN Max offers up to 24 storage nodes and 8.6PB (petabytes) of capacity with 3.6 million IOPS.
Also added is Data Services Manager 2.0, aimed at rapid provisioning, instant replication and data protection in VMware Cloud Foundation. Key among the additions here are the PostGres-compatible Google Alloy DB Omni and MinIO Object Store as options to deploy for developers. These come in addition to existing PostGreSQL and MySQL capabilities.
Meanwhile, VMware Ransomware Recovery brings threat detection in snapshots – so presumably isolated from production data – and rapid recovery points based on those snapshots.
A new capability announced is VMware Live Recovery, which puts ransomware recovery and disaster recovery into a single management plane.
Further enhancements were announced to VMware Sovereign Cloud, which include the addition of data services such as MongoDB, Kafka and Greenplum, as well as in VMware’s Tanzu application development and Kubernetes platforms, which saw the addition of VMware Tanzu Application Service (for application development) and VMware Tanzu Intelligence Services (to manage and monitor costs, etc, in cloud application deployment).
Audi to use VMware in shift to AI-powered production lines
The shift from internal combustion engines to battery electric vehicles will provide the opportunity for Audi to automate car production and bring in “software-defined automation”, said Audi senior production manager Henning Loeser at VMware Explore in Barcelona this week.
According to VMware’s senior vice-president and general manager for service provider and edge, Sanjay Uppal, machine learning at the edge will see 10 times growth by 2026. That figure backs up VMware’s focus on the “cloud edge” and the use of generative AI at the edge, with Audi’s Loeser talking of the car maker’s plans in this area.
“The move to battery electric vehicles will provide the opportunity to put in new production lines with sensors on the production line and allow us to build in compute to help understand that automation in real time,” said Loeser.
Loeser likened the idea of building in vastly increased amounts of sensors, with compute and the ability to react in near real time on vehicle production lines as being like hyper-converged infrastructure.
Core to the plans for Audi will be the deployment of virtualised PLCs to enable software-defined automation.
PLCs – or programmable logic controllers – are key to manufacturing processes. They replaced mechanical means of production process control with programmable units that run strictly defined functionality. Virtual PLCs will form part of the edge cloud, with generative AI processing capability and, potentially, the ability to process production data from monitoring points across manufacturing environments.
Audi has already started to automate the quality of spot welds on its lines. The deployment of virtual PLCs in one part of a manufacturing process, such as this, could reference and learn from monitoring elsewhere in production using generative AI. In this way, the use of the edge cloud and AI would enable fully automated and self-learning continuous improvement as used in lean manufacturing schemas.