Modal Title
Data Science / Kubernetes / Machine Learning

Moving Fast and Smart with Data Using Kubernetes and AI

AI and machine learning are amazing technologies, but getting the most out of them means focusing on access to data.
Dec 15th, 2021 6:07am by
Featued image for: Moving Fast and Smart with Data Using Kubernetes and AI
Photo by Kelly L from Pexels.

Murli Thirumale
Murli is general manager of Pure Storage's Cloud Native Business Unit - Portworx, where he is responsible for strategy, operations and solutions that deliver multicloud data services for Kubernetes. He holds an MBA from Northwestern’s Kellogg Graduate School of Management, where he was an F.C. Austin Distinguished Scholar.

As we head into 2022, moving fast and moving smart with data should be the IT mantra for business success.

The new technology stack based on Kubernetes and containers is foundational to the path to fast. The path to smart? Machine learning and artificial intelligence (AI), of course, and the new data as-a-service tools that help unlock them.

In a recent survey commissioned by Pure Storage, 68% of respondents said they had increased use of Kubernetes as a result of the pandemic, to speed deployment of new applications and increase their use of automation. That makes sense. Across industries, companies found themselves accelerating digital projects and rolling out new apps and services to address the shift to remote work, supply chain disruptions and other market trends — or simply to tighten their fiscal belts.

The benefits of that increased usage will persist as Kubernetes expands its role within the enterprise to manage more demanding workloads in addition to the container orchestration that is at its core. We’ve found a surprisingly high rate of Kubernetes environments used to test or develop AI models and applications — 84% of respondents in our survey, in fact. That’s a very real sign of its strength as a platform for the most demanding workloads.

Driving Innovation with AI

The adoption and maturation of AI is leading to dramatic changes and successful applications in many industries. In financial services, for example, AI is driving innovation in key areas such as market insights, customer experience, fraud detection and algorithmic trading.

In retail and grocery over the next few years, we’ll see accelerated investments in technologies to support development of things like digital shelves, real-time inventory visibility, robotic fulfillment and automated checkout. Companies like Kroger are using AI to reduce self-checkout errors and machine algorithms to power smart search capabilities that can identify in-stock items, or point customers to other nearby locations where inventory is better stocked.

In quick-serve restaurants, industry leaders like McDonald’s are famously employing AI to fine-tune advance ordering, speed customers through the drive-thru lanes and enhance personalized menu recommendations.

Data science teams working with AI need an infrastructure that allows them to quickly experiment with different algorithms and models in a range of different computing environments. Consider the strengths of Kubernetes — easy scalability, portability and fast iteration cycle — and it’s easy to understand why it is an effective path to rapid development.

The Data Services Difference

Looking ahead, data services are becoming an important part of the mix for AI and other modern apps, as enterprises reinvent their businesses along their path to “smart” by building these modern applications based on an architecture that includes microservices, containers and Kubernetes.

Database-as-a-service, for example, is among the key as-a-service pipelines that represent a new frontier in IT innovation.

Microservices use many different data services rather than a single monolithic database, including SQL options like Postgres and MySQL, or NoSQL options like MongoDB and others. Yet each of these has its own way of being deployed, its own way of monitoring, and its own way of managing backup and recovery. Database-as-a-service can provide scalable data repositories that are available across teams and applications so developers can focus on using data services, not managing them.

The Road Ahead: A Dual Role for Kubernetes

Beyond AI and other modern apps, Kubernetes has blossomed into a dual role: orchestrating containers and orchestrating infrastructure. Already extensions like the Container Network Interface (CNI), Container Runtime Interface (CRI) and Container Storage Interface (CSI) are enabling organizations to manage and automate their storage, network and compute infrastructure.

That role will continue to grow as Kubernetes becomes a key middleware component for managing data and hardware. Leading Fortune 2000 companies are well on their way to maximizing their use of Kubernetes as a service control plane for the future.

After all, AI and machine learning are amazing technologies, but getting the most out of them means focusing on access to data — making it more readily available, making it more ubiquitous and making it super fast.

Group Created with Sketch.
TNS owner Insight Partners is an investor in: Unit.
THE NEW STACK UPDATE A newsletter digest of the week’s most important stories & analyses.