Turbonomic has expanded the capabilities of its application resource management (ARM) platform to enable IT teams to optimize and scale infrastructure resources for Kubernetes clusters automatically in real-time by resizing the containers that make up a cloud-native application.
At the same time, Turbonomic added the Fast and Accurate Container Planning capability, which allows IT teams to simulate capacity requirements for applications and services that might be added to a Kubernetes cluster.
Asena Hertz, director of product marketing for Turbonomic, says these latest capabilities extend its ARM platform, which relies on machine learning algorithms to make it easier for IT teams to manage resources in Kubernetes environments, where applications tend to scale up and down dynamically. In the absence of such tools, many IT teams have hesitated to deploy Kubernetes cluster at scale, she notes.
Most of the organizations that have deployed Kubernetes at scale are farther along the DevOps maturity curve, Hertz says, so the goal now is to provide more visibility into these complex environments. Once that goal is achieved, it then becomes feasible to automate a wide range of IT infrastructure management processes based the service level optimization (SLO) an IT team is trying to achieve and maintain, she adds.
One the challenges organizations face is the difference among different IT teams within the same organization when it comes to their DevOps maturity, as organizations deploy Kubernetes on virtual machines and bare-metal servers in the cloud and in on-premises environments. Tools such as ARM should make it easier to bring teams just getting started with containers and Kubernetes along that maturity curve faster, notes Hertz.
It’s still early in terms of Kubernetes adoption, so not that many organizations are deploying multiple cloud-native applications on the same cluster. However, it’s only a matter of time before just about every organization looks to maximize the IT infrastructure resources being consumed by Kubernetes clusters. Otherwise, the cost of Kubernetes cluster sprawl would be unsustainable.
Less clear right now is to what degree the average IT administrator will be able to manage Kubernetes clusters effectively. Most of the teams managing Kubernetes clusters today tend to have extensive programming expertise that the average IT administrator typically lacks. However, as more graphical tools for managing Kubernetes become available, Kubernetes clusters should become more accessible to those IT administrators.
In the meantime, Turbonomic is betting Kubernetes will spur demand for IT management tools that take advantage of machine learning algorithms to automate highly dynamic IT environments. Many IT teams have been hesitant to adopt such tools because they hand have tended to limit visibility into the IT environment. However, many cloud-native applications environments now change faster than any human IT administrator can track, much less optimize manually within seconds.
Of course, many of those same IT administrators are wary that machine learning algorithms and other forms of artificial intelligence (AI) might one day replace them. However, given the complexity of the IT environments that must be managed at scale, it’s far more likely they will find themselves orchestrating a small army of digital assistants that have been trained to handle a very specific set of tasks.