Platform9 Arrives at Simplified Kubernetes With Fairwinds
Platform9 is a company that has styled itself on its ability to simplify cloud operations with a platform and toolset designed to combat cloud computing’s convolutions and contortions. As the market for enterprise-grade fully managed services rolls inevitably forward, the company is looking for ways to apply its cloud cost optimization technologies to new software topologies.
Key among those attack surfaces is Kubernetes.
By colluding with Fairwinds and centralizing on its managed Kubernetes-as-a-Service technology, the two companies hope they can create a Kubernetes cost control mechanism for public cloud deployment that is worthy of mission-critical applications.
Why do Kubernetes Costs Spiral?
It’s no secret among those system and software engineers deploying Kubernetes, deployments often cost more than initial project plans have projected. This is down to several reasons including skewed project scope, a lack of visibility into the active state of a Kubernetes cluster and the fact that – as previously explained on Cloud Native Now – Kubernetes is typically a “shared computing resource” so team usage responsibilities (and corresponding billing assignment) can be hard to pin down without “show back measures” and other chargeback mechanisms.
Then, after all those challenges, there is (again, no prizes here) Kubernetes over-provisioning. As we know, developers over-provision resources to ensure their applications have the CPU and memory needed during peak demand.
“Over-provisioning Kubernetes is a problem many organizations are facing. Instead of continuing to waste resources, Platform9’s Elastic Machine Pool seeks to tackle this problem by enabling users to run existing Amazon Elastic Kubernetes Service (EKS) workloads on fewer nodes,” said Madhura Maskasky, Platform9’s co-founder and VP of product.
Live Cloud Compute Rebalancing
Platform9’s Elastic Machine Pool is a compute engine for EKS that runs on AWS Bare Metal to target utilization in EKS VM instances at the hypervisor layer. Designed for DevOps and FinOps teams, EMP promises to help deliver automated EKS cluster efficiency. How does it work? It is capable of “live rebalancing” of cloud compute resources with no application disruption required.
Because IT teams typically have to rely on manual optimization to combat these Kubernetes cost management, Platform9 says its technology proposition and place in the market is therefore validated. The company insists that a more efficient and automated approach to cost optimization is essential for operational efficiency.
Maskasky has also pointed to the “growth of AIML and container workloads” as key to surging Kubernetes demand. For the acronym challenged, artificial intelligence markup language (AIML) is an XML dialectic subset that software application developers can use for a variety of purposes. Primary use cases for this technology are in the creation of natural language software agents typically used in chatbots.
Kubernetes-as-a-Service
Fairwinds provides a managed Kubernetes-as-a-Service that works with platform engineering teams to manage infrastructure, perform upgrades and support decision-making, allowing engineers to focus on business-specific developer needs.
“Over-provisioning Kubernetes is a problem many organizations are facing. Instead of continuing to waste resources, Platform9’s Elastic Machine Pool seeks to tackle this problem by enabling users to run existing EKS workloads on fewer nodes,” said Andy Suderman, CTO at Fairwinds. “Getting EMP up and running during a proof of concept showcased that with a managed Kubernetes-as-a-Service to help streamline implementation, enterprises can seamlessly add a layer of efficiency optimization to Kubernetes environments.”
As the application of automation hits every level of the IT stack, it is logical enough to see cloud compute convolutions being intelligently rebalanced and mapped out in this way. We might wistfully hope that cloud architects and software engineers at all levels might use these tools to create more efficient applications and subsequently extract the DNA from those projects to build their next project more efficiently. More likely is a reality where client requirements jolt the best-laid plans out of kilter and we still need an Elastic Machine Pool to cool off in.