Red Hat Previews Generative AI Capabilities for OpenShift
Red Hat today revealed it has made available a developer preview of generative artificial intelligence (AI) capabilities that promise to make it simpler to manage instances of the Red Hat OpenShift platform based on Kubernetes clusters.
Designed to work with multiple large language models, Red Hat OpenShift Lightspeed provides IT teams with a chat interface through which they can launch prompts that provide, for example, guidance to help invoke specific capabilities or a YAML file that could be used to complete a task. Additionally, IT teams can fine-tune prompts by exposing additional data using retrieval-augmented generation (RAG) techniques.
Each IT team can decide for itself which type of large language model (LLM) they prefer to invoke via the Red Hat OpenShift Lightspeed interface, so long as they are running version 4.15 or higher of the platform. Alternatively, IT teams can opt to bring their own LLM if they so choose.
Stu Miniman, senior director of market insights for hybrid platforms at Red Hat, said ultimately Red Hat is working toward leveraging generative AI to manage the underlying Kubernetes clusters themselves. In addition, Red Hat via a project with IBM is embedding generative AI capabilities to Red Hat Ansible that can be applied to instances of Red Hat OpenShift.
Ultimately, Red Hat OpenShift Lightspeed should reduce the current level of cognitive load required to use the platform. Most instances of Kubernetes platforms are typically managed by DevOps teams, but with the rise of generative AI tools, it is becoming more feasible for IT administrators to take on more of those tasks. Those capabilities should make it easier for more organizations to build and deploy cloud-native applications without necessarily having to centralize the management of all these platforms under the auspices of a platform engineering team that is trying to enforce best DevOps practices at scale.
Longer term, it also remains to be seen what impact generative AI will have on managed services that many organizations currently rely on to consume Kubernetes clusters, noted Miniman. As it becomes simpler to deploy and manage Kubernetes environments, the demand for those services might decrease or, conversely, as the cost of delivering those services theoretically declines demand might increase.
Regardless of approach, the number of Kubernetes clusters distributed from the edge to the cloud is only going to steadily increase. The challenge now is finding a way to manage those clusters in a way that doesn’t require nearly as much specialized expertise in an era where finding and retaining IT professionals that have those skills is still a challenge. Generative AI, in effect, is providing a higher level of abstraction for managing IT platforms.
As generative AI capabilities become pervasively available across every type of IT platform the silos that exist in enterprise IT environments should start to fade away. In the case of cloud-native computing, however, that’s especially critical for cloud-native applications because arguably the single biggest issue holding back the adoption of Kubernetes platforms has been complexity.
