In the days preceding the kick-off AWS Re:Invent 2020, the annual conference that gathers AWS enthusiasts worldwide, there was a very special announcement about the release of a new service: AWS ROSA. The name, an acronym for Red Hat OpenShift Service on AWS, came to ignite the discussion about hybrid cloud management strategies and brought a new alternative to customers looking into migrating or extending their on-premises workloads to the public cloud.
But whileOpenShift Container Platform is new to AWS, it has been integrated with Azure in the Azure Red Hat OpenShift service for several years now. What do these services do and how can they help you? In this article we’ll take a closer look at these OpenShift options on AWS and Azure and what they offer hybrid cloud deployment at the enterprise level.
In hybrid cloud architectures, computing capabilities are split between a public cloud provider and an on-premises private cloud, with the exchange of applications and data between them fully enabled. This is a strategy adopted by many enterprise organizations worldwide. Red Hat OpenShift has been a prime choice for users with on-premises container-based workloads.
At its core, OpenShift is based on Kubernetes. However, OpenShift is much more than an enterprise-supported Kubernetes platform.
With OpenShift, Red Hat built an entire ecosystem of services to enable developers to create container-based systems. Services that you would typically have to set up elsewhere in addition to a Kubernetes cluster are available in your OpenShift ecosystem, such as container image repositories, logging services, monitoring tools, middleware applications, among others. OpenShift gives developers and system operators a complete platform that enables application development throughout its entire lifecycle.
By using Kubernetes at its core, OpenShift enables the same level of portability that makes containers and Kubernetes so popular among organizations which value that increased flexibility and vendor neutrality. While on-premises users have been the primary target audience for OpenShift, with the ubiquitous nature of Kubernetes at its core, these container workloads can run in multiple different target environments, be they on-premises, in the public cloud, or in the edge.
In the past few years, Red Hat bridged the gap between on-premises and cloud by teaming up with Microsoft, and more recently Amazon Web Services, to create OpenShift on AWS and OpenShift on Azure as fully managed services. It is however worth noting that Google Cloud Platform does not have a similar offering available yet.
Microsoft was the first major public cloud to provide a fully managed OpenShift on Azure service. The Azure Red Hat OpenShift, or ARO, enables customers to extend their Kubernetes capabilities with enterprise-grade tooling and resources. For existing OpenShift on-premises users, this is a minimal effort to extend their environment and create an Azure hybrid cloud, enabling them to take advantage of Microsoft infrastructure and services.
Red Hat and Microsoft teamed up to provide a completely integrated experience across both Azure and OpenShift. In fact, the value proposition of Azure Openshift is that you get all the capabilities of OpenShift but don’t need to operate any underlying infrastructure, such as virtual machines or hardware. Those aspects are being taken care of jointly by Microsoft and Red Hat.
Azure OpenShift deployment and billing is, as usual, included as part of your Azure subscription. The two biggest advantages of using ARO come from the simplicity in terms of low operational overhead, since Microsoft and Red Hat include enterprise support and ensure the service is up and running using their site reliability engineering teams, and the inherent integration and access to other Azure services such as an Azure Cosmos DB, Azure SQL DB and Azure Machine Learning, helping organizations to leverage other cloud native managed services.
Since its announcement back in 2018, Microsoft ARO has been developing steadily and growing in terms of availability and maturity. One interesting characteristic is that the ARO’s roadmap is publicly available, making it possible to have a glimpse of where its heading and to some extent reach out to product teams and influence it.
When it comes to down to cost, the customer has the possibility to choose a pay-as-you-go hourly pricing model or do a yearly upfront investment in exchange for a price reduction.
There are many variables to consider: Azure region, OpenShift version (3 or 4.x), and node instance type based on the needed vCPU and memory. On top of the hourly price of each node there is a small fee for OpenShift, while still keeping the total price per node under a few dollars per hour. When customers commit to 1- or 3-years upfront reservation on the node, they can enjoy between 30 to 60% savings.
With the enterprise support starting at $29 (though this varies according to the chosen plan) and a guaranteed 99.95% connectivity SLA for multiple instances, this is a very interesting option for existing OpenShift customers looking into setting up an Azure hybrid cloud architecture.
The newly announced Red Hat OpenShift Service on AWS (ROSA) takes a very similar approach to its competitor ARO, to offer an AWS hybrid cloud experience to existing OpenShift on-premises customers, or new customers looking to benefit from OpenShift on AWS.
The new AWS OpenShift service is currently in preview. To use it, customers need to register their interest and wait until they get accepted. While the entire range of capabilities are not publicly disclosed, there’s a good idea of what customers can expect.
From the get-go, customers will be able to experience a fully featured OpenShift cluster that is jointly managed and supported by both Red Hat and AWS. By using OpenShift on AWS, customers are expected to also gain the ability to leverage existing cloud managed services such as AWS Relational Database Service (RDS), AWS DynamoDB, Amazon S3, and AWS Simple Queue Service (SQS). One interesting aspect is that this can be done directly using the OpenShift Service Catalog in conjunction with an AWS Service Broker, providing an intermediate layer that enables users to deploy selected AWS services using native manifests and the OpenShift Console.
According to the AWS OpenShift product page, the AWS ROSA service will be available from the start in all AWS Regions. Moreover, it gives a glimpse to what the upcoming roadmap might look like, such as an improved IAM permission experience, AWS EC2 Networking integration, additional storage plugins and Amazon Controllers for Kubernetes.
AWS ROSA is a fully managed service, in which, like it’s Azure competitor, all the operational aspects are taken care of and the customer can simply enjoy a production-grade OpenShift cluster environment without having to spend time maintaining it.
The pricing is simple to understand and essentially calculated using two components, a fixed hourly fee for each cluster ($0.03/cluster) and the cost per worker node ($0.171 per 4vCPU /hour). Customers can decide on a flexible pay-as-you-go hourly pricing model or enjoy a 30 to 55% discount when reserving instances upfront for 1 or 3 years. The AWS OpenShift pricing is the same regardless of the chosen AWS Region.
The value proposition of peace of mind with near-zero operational overhead and simplified hybrid cloud experience is incredibly appealing to on-premises customers. For existing OpenShift customers, the ability to expand or migrate their workloads to the public cloud without having to re-engineer existing applications is quite tempting. The lack of underlying access to the nodes and possibilities to fully customize the infrastructure might be quite limiting for customers used to having that freedom in their on-premises environments.
For experienced cloud architects, it is inevitable to draw comparisons with Azure Kubernetes Service and AWS Elastic Kubernetes Service, both of which are existing managed Kubernetes services that also have the possibility to expand to non-cloud environments and are far less limiting. While OpenShift offers more out-of-the-box features than a managed Kubernetes platform, this is definitely something to take into consideration.
It is clear that these OpenShift on AWS and Azure are specifically targeted and recommended for enterprise users that have already made significant investments in on-premises OpenShift environments and might not have yet much cloud experience. It is however, an option worth considering for any customer looking for a cloud managed Kubernetes platform with zero infrastructure management and enterprise support.
A highly complex challenge to consider and solve when using a hybrid cloud architecture is data storage. While OpenShift does make it easier to shift workloads across environments, the same smooth experience does not apply to data.
To enable effortless data management, OpenShift users can take advantage of NetApp Cloud Volumes ONTAP for OpenShift. NetApp Cloud Volumes ONTAP, the premier cloud storage and management solution for enterprise customers, integrates natively with OpenShift to provide persistent volumes, data protection, cloning capabilities, built-in Ansible automation support, among other features.