Spectro Cloud today extended the reach of its management framework for Kubernetes clusters to now include instances running on edge computing platforms.
Tenry Fu, Spectro Cloud CEO, says Palette Edge is an extension of the Real Metal architecture the company added to its Palette management framework last year. The goal is to provide a framework for managing multiple Kubernetes distributions running everywhere from the edge to the cloud regardless of whether they are deployed on bare metal servers or virtual machines, he said.
That framework makes it possible to deploy pre-validated and curated stacks of software on Kubernetes clusters, including operating systems and add-on application services. It leverages open source technologies, such as the Cluster application programming interface (API) defined by the Cloud Native Computing Foundation (CNCF), the Metal-as-a-Service (MaaS) framework created by Canonical and Cluster API MAAS project created by Spectro Cloud.
The Palette framework employs those technologies to enable both DevOps teams and IT administrators to centrally manage distributed Kubernetes environments. This can be accomplished either via APIs or a graphical user interface (GUI) that Spectro Cloud created to make Kubernetes more accessible. In effect, the management of Kubernetes clusters becomes plug-and-play, says Fu.
In addition to enabling IT teams to apply and enforce policies, the Palette framework also makes it possible to orchestrate rolling upgrades of fleets of Kubernetes clusters without any downtime.
It’s not clear to what degree IT organizations are still intimidated by the complexity of Kubernetes. As management frameworks that abstract away the complexity of Kubernetes become more widely available, it is becoming simpler for IT administrators to manage and govern Kubernetes environments. A recent survey published by Spectro Cloud finds more than three-quarters of respondents (77%) have been working with Kubernetes clusters in some form for the past two years. The survey also finds only 40% have deployed them in a production environment. Only 21% say they currently have more than half of their clusters in production.
A full 98% of respondents report myriad Kubernetes challenges ranging from implementing consistent management and controls for enterprise environments (49%) to managing multicluster (46%) and heterogeneous multi-cloud environments (42%). These are followed by challenges with integrating required services (41%), managing security (41%) and meeting compliance requirements due to configuration drift (34%), managing the needs of one-off use cases such as artificial intelligence / machine learning ( AI / ML) or GPU support (33%) and scaling to large environments (24%).
More than three-quarters (77%) of respondents say that being able to run various Kubernetes stacks across different environments is vital but that it also makes it harder to manage those clusters. Two-thirds (66%) say they do not believe they can have both flexibility and usability in a Kubernetes environment.
The management of Kubernetes clusters is clearly evolving. DevOps teams that initially deployed a cluster are looking for ways to enable IT administrators — that often lack programming skills — to assume more responsibility for managing routine tasks across a fleet of Kubernetes clusters. After all, every minute a DevOps team spends managing an existing cluster is one less minute they can spend on rolling out the next great application.