Enterprise Storage Forum content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Container usage continues to go up, up, up. The adoption curve of containers may even surpass that of virtual machines (VM) eventually, according to some analysts. But surging popularity tends to flush out other issues. And one of the big ones with containers is how to manage them more effectively.
Container orchestration tools such as Kubernetes have simplified the management of containers and boosted their efficiency. They add automation in areas such as deciding which server will host which container, launching and grouping containers, deploying updates, testing, and adding and managing security, storage, and networking services for applications.
Here are some of the top trends in the container management market:
Incorporation of Artificial Intelligence
Improved container management courtesy of Kubernetes is being harnessed in AI apps to drive growth. As a result, he global enterprise AI market is anticipated to grow at a Compound Annual Growth Rate (CAGR) of 39.7% to $309.6 billion by 2026.
“New workloads like AI require massive datasets, a high degree of parallelization, and high-performance compute and storage,” said Morgan Littlewood, Senior Vice President of Product Management, iXsystems.
“KubeVirt provides integration of VMs with Container Storage Interface (CSI) CSI, and traditional virtualization approaches like vSphere and OpenStack will migrate to classic status.”
KubeVirt is another way to help development teams adopt and benefit more from Kubernetes. As some existing VM workloads are difficult to containerize, it offers a unified development platform for containers and VMs. The end result is an enhanced ability to containerize VM workloads as well as improved development workflows and the easier management of VMs that may never be containerized.
MLOps
Beyond AI, container management is having an impact in the evolving fields of MLOps and AIOps (often used interchangeably). MLOps is a shortening of the words Machine Learning Operations. It is rapidly becoming a key aspect of ML engineering as it helps streamline how ML models are taken through to the production stage. But it doesn’t stop there. It also is involved in maintaining and monitoring machine learning models by bringing about greater collaboration between IT, data scientists, and DevOps.
Improved container management, then, has been of material assistance in aiding in the forward progress in the MLOps sector. But it works both ways. MLOps has added a new dimension to container management.
“Containerization being the key, Kubernetes will aid cloud-native MLOps in integrating with more mature technologies,” said Bin Fan, Vice President of Open Source and Founding Engineer at Alluxio.
“To keep up with this trend, organizations can find their AI workloads running on more flexible cloud environments in conjunction with Kubernetes.”
MLOps enables IT to address application performance challenges without them having to spend days figuring out what processor, memory, storage, networking, or IOPS resources are to blame. Its value comes from the fact that the scale of modern computing is such that problems can no longer be solved by relying solely on human interaction and manual labor. Automation is required and if it is backed by AI or ML, it is far easier and faster to add context to large volumes of data and refine conclusions in very short time frames. This adds value in event correlation, event analysis, anomaly detection, root cause analysis, natural language processing, automation, and diagnostics. How? It can rapidly discover patterns that can be used to predict incidents, spot emerging behavior, determine the reason for slows, and drive automation.
Improved Storage Management
A major container management trend is how it is augmenting storage management and causing changes in the storage management space. Cloud co-location is growing more into hybrid cloud designs as a way to help avoid cloud-wide outages by shifting to hybrid cloud and on-premise solutions.
“The use of containers with disaggregated resources will continue to grow, blurring the line between storage and physical hardware,” said John Scaramuzzo, CRO of Nyriad.
“Block, file, and object will all be residing on common hardware and Hybrid Storage as a Service will see continued growth utilizing a virtualized consumption model between cloud and on-premise storage.”
Better Scaling
Early container deployments didn’t always scale well. Kubernetes and other container management functions have addressed this challenge. Red Hat OpenShift, for example, provides automatic scaling without the need to configure the number of replicas, or idling. You can scale to zero when not in use, auto scale to thousands during peak, with built-in reliability and fault-tolerance, according to Naina Singh, principal product manager at Red Hat.
Other platforms, too, are rising to the scaling challenge via container management advancements. IBM Turbonomic, for example, adds management capabilities to heighten application performance, optimize cloud spend, and improve the scaling of applications in containerized environments.
