Infusing AI into IT operations

An automation strategy is the foundation for AI adoption and is critical for success. Let’s explore some examples of this.

In my previous blog post, we delved into how uptime and availability are not just technical issues; they are essential to the success of AI-driven business operations. For businesses leveraging AI to improve efficiency, foster innovation, and boost customer satisfaction, ensuring their infrastructure is reliable and always accessible is vital. Automation is essential for AI, as it enables the reliability and efficiency necessary to handle AI workloads.

Red Hat Ansible Automation Platform plays an important role in establishing a strong foundation for AI implementations by facilitating standardized deployment, scalability, configuration management, high availability, monitoring integration, disaster recovery, version control and documentation. By automating routine tasks and thus enforcing best practices, Ansible Automation Platform enables the reliability, performance and resilience of AI infrastructure within IT operations.

3 pillars of automation use cases

Let’s take a closer look at three categories of automation use cases involving AI: orchestration with AIOps, enabling AIOps and infrastructure optimization.

Orchestration with AIOps

It is worth exploring how automation facilitates the shift from reactive to proactive IT strategies. AIOps proactively identifies and addresses issues before they impact important business metrics, enabling better informed decision-making and faster problem resolution. So, how can you harness these capabilities, and what is the role of automation?

Consider the fact that many of the existing technology vendors in your infrastructure today already have integrated predictive AI. Are you leveraging these capabilities to their full extent? Automation use cases, involving orchestration, can help organizations put AI capabilities to use immediately. In many instances, the built-in AI features are not enabled by default. Configuration management and automation must be used in order to take advantage of these components and integrate them with your IT infrastructure.

One such example involves network devices with built-in threat intelligence. These network switches can autonomously pull network data and contextualize potential threats uncovered in the environment. To put this feature to use, it must be enabled on every single switch. Imagine manually configuring and authenticating 500 switches! This is a perfect example of how you can leverage automation for configuring the switches in order to realize the value of the AI capabilities. Using Ansible Automation Platform, you can leverage the built-in event-driven automation feature, allowing you to trigger automation jobs that perform actions based on predefined event conditions. With a simple set of rules, you can initiate specific actions automatically, in response to those events, shifting to a more proactive IT strategy. In this particular example, an automatic notification is generated and an immediate blocking of the user involved.

Enabling AIOps

AI can empower IT infrastructure engineers with enhanced capabilities when integrated into the automation workflow. Consider an example involving a self-healing infrastructure scenario. With Ansible Automation Platform, event-driven automation can be leveraged when the event is recognized and rulebooks are in place. But what happens when we don’t know what the event means? Red Hat AI solutions can be used to help determine the event/error. Then, you can create the automation to fix the error. We can now put this into a full cycle loop to fully realize self-healing infrastructure.

Let’s explore this a bit further. An event occurs and is picked up by the observability tool. An Ansible Rulebook then detects the event and runs a corresponding playbook, if it is a known event (example: an application has failed and won’t restart). An automatic service ticket would then be created. If it’s not a known event, the event can be funneled into a platform such as Red Hat Enterprise Linux AI  or Red Hat OpenShift AI to identify the issue. You could then use a generative AI tool such as Red Hat Ansible Lightspeed to create a playbook that will fix the issue. Then, automated policy as code can help keep your automation guard rails on. Policy as code entails embedding operational policies and best practices into automation code, so that internal requirements, security standards and specific mandates are integrated into every process and will be enforced across all operations and domains. This will also reduce downtime and increase confidence in your operations. For example, when using Ansible Lightspeed to create an automation playbook, the AI may not inherently be aware of the organization's existing policies, so this capability would assist. Other examples that automated policy as code can assess include: whether or not the modules being used are allowed by the company; if the modules are signed and not modified; and if a new instance being spun up is allowed by the organization. In summary, with the combination of AI and generative AI, this particular automation use case allows an organization to realize a fully self-healing (and compliant!) infrastructure, solving potential issues before they cause an outage.

Infrastructure optimization

How can Ansible Automation Platform help Red Hat Enterprise Linux AI and OpenShift AI solutions? Ansible Automation Platform can be used to set up and configure AI services, install and download different models, configure and verify GPU acceleration and coordinate different parts of your infrastructure. AI frameworks can involve many different disparate parts even on a single compute node, and Ansible Automation Platform can help standardize and provide effective delivery of your AI services to your customers. Where Ansible Automation Platform can be especially critical is with edge deployments. In many customer environments you need data to get back to your AI compute layer so that you can train the models to continually enhance them. This is where Ansible Automation Platform can be used for installations, retrieving data, manipulating data into the appropriate syntax if needed, and help everything function properly.

A use case in the infrastructure optimization realm involves streamlining the onboarding of new edge deployments, such as IoT devices, to collect and coordinate their data with AI solutions. Wherever you are orchestrating data, models, or Git repository code, you can use Ansible Automation Platform to avoid manually "bootstrapping" or scripting your way to deployment. As model updates are registered, the CI pipeline downloads the model and creates a container. Ansible Automation Platform helps with: registering the model at the near edge; devices at the far edge can pull the latest version; and Ansible Automation Platform provisions and deploys the edge configuration.

Summary

As you can see, automation is critical for taking full advantage of AI capabilities in your organization. Whether automation is helping improve your IT infrastructure by improving resiliency and uptime, coordinating your existing IT vendors and their solutions to take advantage of new AI solutions, or even helping your IT infrastructure engineers directly by enabling AIOps, automation is the foundational layer that helps realize AI in the modern enterprise.

In the remaining blogs in this series, we’ll continue to explore additional use case examples within orchestration, operationalization and infrastructure optimization.

Learn more

• Why choose Red Hat Ansible Automation Platform as your AI foundation?

• Watch this webinar: IT Automation: a key enabler for enterprise AI adoption