In a previous article we introduced the open source container management orchestration tool Kubernetes. We looked at where Kubernetes can be used, how it works and the benefits of using such a tool to manage your containerised applications. As more organisations start using serverless architectures to deploy their applications, it has become increasingly popular. Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform (GCP) have all introduced Kubernetes into their service offerings to ensure that their users are able to manage, upgrade and deploy their Kubernetes-based architectures straight from the cloud. We are going to deep dive into the strengths and drawbacks of each cloud platform’s Kubernetes offering and how each can be beneficial to your organisation’s containerised applications.
First we have AWS’s Elastic Kubernetes Service (EKS) which is a fully managed service that allows users to integrate their pre-existing, on-premises Kubernetes solutions with AWS. EKS is compatible with applications that are already running on Kubernetes. However, a downside of EKS is that it can be difficult for users to set up and requires its users to have experience with containerised applications. EKS is a fully scalable and customisable solution that allows users to be in complete control of their Kubernetes architecture and deployments. On the other hand, GCP’s Google Kubernetes Engine (GKE) is well-known for its quick and easy setup, allowing users to easily deploy production ready, containerised applications using Kubernetes without having to install and manage Kubernetes clusters. The Azure Kubernetes Service (AKS) allows users to easily deploy their Kubernetes based applications from Azure and provides them with detailed analytics on their deployments, to ensure that they are using Kubernetes effectively and efficiently. Since AKS was introduced after EKS and GKE, updates to the service take significantly longer.
An important aspect of Kubernetes is ensuring that your clusters are upgraded. GKE allows for automated updates to clusters which can make it easier for users to manage their clusters. AKS requires users to run a single command to update their clusters, however, EKS is the least automated in that users are required to interact directly with their command line using commands to update their clusters which can make EKS significantly more difficult to work with than AKS and GKE. Out of the three cloud platforms, it is evident that GKE is the easiest to work with and manage in terms of upgrading your Kubernetes clusters.
In addition to cluster upgrades, another important aspect of Kubernetes is resource monitoring. GKE makes use of a GCP service known as Stackdriver to monitor your Kubernetes resources provisioned on the platform, requiring only a few manual configuration steps. AKS uses two services, Azure Monitor and Application Insights to monitor and evaluate the performance and health of your Kubernetes clusters running on the platform. AWS, however, does not readily provide services that can be integrated with EKS for monitoring clusters, therefore users are required to use third party tools for monitoring.
Autoscaling is an important concept in Kubernetes. It is vital that Kubernetes clusters scale up or down seamlessly according to the demand experienced by the applications, to ensure high availability as well as low latency. GKE is the most mature in its autoscaling capabilities as it is easy and simple to use, EKS requires only a few manual configurations for autoscaling while AKS has introduced auto scaling but there is still great complexity in configuring it, however, this will improve with time.
Determining the best Kubernetes cloud platform service to use, solely depends on the cloud platform that you are currently running within your organization. Each platform’s Kubernetes offering has its benefits and drawbacks, where some organizations may prefer to manually configure their Kubernetes clusters, others may prefer a fully managed solution. Regardless of the Kubernetes service you use, deploying your Kubernetes architectures from the cloud can ensure scalability, efficiency and easy maintenance of your containerized applications.
Contact us today on web@piidigital.co.za for us to help you make an informed decision that best suits your company!
References:
https://jaychapel.medium.com/managed-kubernetes-pricing-comparison-eks-vs-aks-vs-gke-dbf3f2c6290c
https://www.toptal.com/kubernetes/k8s-aws-vs-gcp-vs-azure-aks-eks-gke
https://www.sumologic.com/blog/kubernetes-aws-azure-gcp/
https://cloudacademy.com/blog/kubernetes-services-aws-vs-azure-vs-google-cloud/it
