Leveraging Kubernetes to Fuel Intelligence Everywhere
Way back in 1992 when I was leading development for a shipping company, the Head of Sales forever changed the way I approached using data in our business environment. The request from Sales was simple enough. They needed to provide a breakdown of revenues, costs, largest customers, and some other attributes that could not be answered from any of our existing thousand plus reports. Worse, it would take 2-3 weeks to create a new report. The response from our Sales leader turned my way of thinking upside down: “That’s not good enough. I need this for a board meeting tomorrow. We have new competitors that are disrupting our pricing.”
What became very clear to me at that moment was that business users needed insights from across our entire business and they needed it at the speed of the business. They needed to be in control and not subject to lengthy IT-imposed wait times. IT must be the enabler of this process rather than the governor.
Fast forward to 2020. Those needs are still with us, but many business users will begin to see the infusion of intelligence everywhere combined with the adoption of Kubernetes as the architectural foundation to deliver the technologies and platforms that enable information on demand.
Intelligence Everywhere: A Journey From Self-Service to Self-Driving
Businesses have always been using data to make decisions in an effort to grow or improve operational efficiencies. Self-service BI tools turned the traditional push model into a pull-and-push approach allowing users to formulate new queries on predefined data sets. The success of companies such as Business Objects, Cognos, and (more recently) Tableau are a testament to the value they bring to the market. Research from firms such as Gartner indicates that about a third of today’s worker community have access to these self-service type BI tools. However, this approach largely was limited to some predefined data schemas that represent the sphere of questions that could get answered.
The maturing and emergence of Artificial Intelligence (AI) and Machine Learning (ML) technologies are adding a new dimension to this. We now can discover entirely new answers and insights based on machines discovering patterns in data that had never been imagined. One of the biggest benefits of the application of AI is that we can begin to automate thousands of small decisions that are taking place every day across our businesses. We now can automatically approve or reject a loan application that is made on the web without human intervention; intelligent chatbots can learn from previous interactions with customers and provide new answers to questions without the intervention of a support agent.
In the next 12 months, I expect to see more intelligence embedded in nearly every facet of our lives, whether personal or business. The earlier pull model of self-service insights remains in place, but it will be augmented by a push and self-driving model where decisions are autonomously made to inform and shape machine-to-machine, machine-to-human, and even human-to-human interactions.
Kubernetes: The Architectural Framework to Power Today’s Insights-Driven World
The incredible complexity of managing data processing, data transformation, and AI/ML technologies has become the biggest challenge to capitalizing on this insight potential. This is because few organizations truly have the skills to run huge Hadoop or Spark platforms.
Thankfully the creators of these successful data processing platforms also brought us the next wave of innovation: Kubernetes. Kubernetes essentially is a container orchestration framework that brings a higher level of manageability to extremely complex technology. VMWare allowed us to virtualize a server into several servers, each running its own operating system, enabling us to then install our application stack on top of this virtualized server image. Containers build on this by encapsulating every element of the application stack that is then managed, scaled up or down on demand, and secured.
Where Kubernetes truly is going to set itself apart as the big winner is that our world is not becoming less homogeneous but more diverse. Best-of-breed applications are coming from dozens of providers, and no enterprise is adopting a single Cloud provider like Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform (GCP), and others. Software providers are adopting Kubernetes as the common framework standard (even though there are slight differences to Kubernetes on AWS versus Azure versus GCP) to manage Multi-Cloud deployments. Likewise, many business technology environments are adopting Kubernetes to provide a framework for managing hybrid On-premises and Cloud environments.
How to Prepare for the New World of Kubernetes and Intelligence
Tip 1: Get Started Today!
If you have not started on your journey to explore AI/ML in your business, it is time to get moving. One of the main challenges is finding top tier data scientists in your local markets. Fortunately, there are a number of software vendors that are making AI/ML less of a dark art for the few and are pushing to democratize this with predefined models and marketplaces to help get them started.
Tip 2: Scale Your Data Preparation to Match the Demand
The biggest oversight in many of these initiatives is that data processing and data/insight consumption technologies are deployed, but we forget that shaping, cleaning and consolidating the raw data into ready information often takes 80 percent of the effort. Leaving this to IT staff or taking a leap of faith in the hopes that a software platform can magically correct the data will not solve the problem. Often, the end result is that the data prep effort gets moved into the hands of business personnel.
Tip 3: Start the Kubernetes Conversation With Key Vendors to Understand Their Plans.
Although Kubernetes as a framework has achieved broad adoption, there is no single standard yet. It is time to have your voice heard and help motivate your vendors to come together around this topic and help drive standards because they will likely not come from the likes of AWS, Azure, or GCP.