Overview of IBM AI Optimizer for Z
IBM® AI Optimizer for Z (AI Optimizer for Z) is a purposefully built software application that empowers you to leverage generative artificial intelligence (gen AI) technologies and accelerate the adoption of gen AI solutions across your IBM Z® workloads. Designed to save time, effort, and cost, AI Optimizer for Z enables you to deploy, integrate, and monitor onboarded gen AI solutions in your enterprise with speed, simplicity, and ease.
Adopting gen AI solutions across an enterprise generally requires a significant number of skills and resources in hardware and infrastructure setup, software and application integration, data connection and processing, network and security configuration. These requirements make the deployment of a new gen AI product difficult, time-consuming, and costly. For some organizations, the challenges do not end with onboarding. They face the continuous challenges of monitoring system health, balancing resource usage, and managing system performance.
AI Optimizer for Z completely transforms the traditional product onboarding and use experiences. This enterprise-grade solution eliminates the complex manual installations and configurations. It automates the deployment, integration, and provisioning of an onboarded gen AI product, such as IBM watsonx Assistant™ for Z or IBM watsonx Code Assistant™ for Z, with minimal intervention and in a fraction of time.
AI Optimizer for Z is platform agnostic and can be deployed to a Red Hat® OpenShift® Container Platform (OCP) cluster that runs on an s390 or x86 system. It's perfectly suited for deploying gen AI applications on IBM Z and LinuxONE where the world’s largest AI workloads run. Seamlessly integrated with the z/OS® or Linux® on Z ecosystem, AI Optimizer for Z leverages the mainframe hardware, the entire software stack, and the secure data connections on Z. The tight integration enables you to exploit advanced AI processors on the mainframe, utilize readily available Retrieval Augmented Generation (RAG) capabilities, and run foundation models with Z data across your AI workloads at any scale. It also provides you a hybrid approach to optimize the capacity and infrastructure inferencing based on availability, cost, and energy consumption across your cloud and on-prem AI workloads.
As shown below, each AI Optimizer for Z instance consists of a collection of deployable services, an operator, a web user interface (UI), and a set of RESTful application programming interfaces (APIs). The deployable services are the onboarded IBM gen AI products. The products are built, configured, and packaged natively in the OpenShift environment as services for autonomous installation, rapid integration, and frequent update. Powered by Red Hat OCP, the AI Optimizer for Z operator automates the installation and integration of the deployable gen AI services.
The AI Optimizer for Z UI features a dashboard that collects, categorizes, and visualizes real-time data of the deployed gen AI services, system resource utilization, and LLM consumption. As an OpenShift or system administrator, you can use the UI and the APIs to manage deployed services, allocate system resources, and monitor the overall system health.