As enterprises begin migrating to the public cloud, it’s important to realize this movement doesn’t happen in one fell swoop. Typically,
cloud migration workloads move in groups related to applications or business functions—for a large organization, this process can take years. In some cases, they may run in this hybrid cloud pattern forever, leading to management challenges. The major cloud providers—Azure
®, Google
® Cloud Platform (GCP), and AWS
®—have offerings to manage existing on-premises VMs through the cloud platform, taking advantage of automated cloud provisioning tools and managing on-premises workloads as if they were cloud resources.
In my experience working with many customers, even smaller clients end up with resources not just in the public cloud but spread across various data centers and co-location facilities. There has been a rapid evolution of services available for hybrid cloud models in recent years. Keeping the hybrid service offerings in sync with cloud models can be challenging for the cloud providers, as the hybrid offerings are a combination of managed hardware and software, which may introduce variables in deployment patterns. However, in some cases, vendors are taking advantage of
cloud-native platforms like Kubernetes—for example, Azure Arc Data Services allows customers to run their version of Microsoft’s Platform as a Service (PaaS) database managed service, Azure SQL Managed Instance. The Managed Instance service can run anywhere you have Kubernetes, opening the service up to multi-cloud deployments.
While many of these hybrid cloud services are dependent on relationships between hardware vendors like Cisco and HPE, VMware
® has taken a slightly different approach. Since virtualization is the basis of cloud solutions, VMware’s solution is simply placing
their ESXi virtualization stack into public clouds while providing a single control plane in vCenter. The advantage of this solution is customers can move their existing VMs in the VMware cloud to a public cloud provider (or hybrid solution) without making significant changes to their network or storage configuration.
What is Hybrid Cloud?
Google loosely defines hybrid cloud as “…one in which applications are running in a combination of different environments.” There are many different implementations, but the way it typically works is some IT services are fulfilled by the internal IT department and data center, while others are fulfilled by the cloud provider. A key part of this is having a common service catalog built to abstract some technical details away from the business users, so the user can request a set of resources to support a given application type. This application could be a machine learning project, Internet of Things (IoT) solution, etc. Behind the service catalog, there is an orchestration of resources created by automation, either with cloud resources or in the company’s data center.
Beyond having a service catalog, another key aspect of
choosing a hybrid cloud solution is hybrid networking. Hybrid networking takes several forms, from small VPN connections to a dedicated high-bandwidth, low-latency connection up to 100 GB/sec. A robust network connection facilitates options like using the cloud as a disaster recovery solution. In some extreme cases, I have seen customers have cross-connections between their data centers and two different clouds—it’s worth noting, though, the costs of those solutions are high because data center space between two public cloud data centers is at a premium.
Comparing Dedicated Hybrid Cloud Solutions
When
selecting a cloud platform, you know know each of the major public cloud providers has its own hybrid cloud solution. These solutions have much in common—typically, cloud resources are deployed through the cloud control plane, and resources are monitored using the same control plane as in the public cloud. The costs can be based on the amount of hardware purchased and the number of resources used.
These cloud solutions include:
- Azure Stack - Azure Stack is built upon partner hardware and uses hyper-converged infrastructure (HCI). It supports building your own VMs, using the Azure Kubernetes Services and software-defined networking (SDN). It connects to Azure and maintains updates and monitoring. It also fully supports the Azure ARC feature set. As for costs, Azure Stack is priced per your usage.
- Google Anthos - Google’s service offering is built around the Google Kubernetes Engine and allows you to run virtual machines. It can be installed on your bare-metal servers or on top of existing virtualized infrastructure. Anthos also allows you to manage your workloads using policy to meet security and compliance requirements. The pricing for Anthos is based on the number of vCPUs you have allocated in your clusters, whether you run on another cloud or in your own data center.
- Amazon® Outposts - AWS hybrid solution is more similar to Azure Stack than Anthos. It allows you to purchase different sizes of hardware with other associated services. If you purchase an Amazon Outposts rack, Amazon supports many of its popular services, including EC2 VMs, S3 storage, EBS storage, Relational Database Service (RDS), and Elastic Kubernetes Service (EKS). Pricing is based on the hardware costs and the resources consumed.
Each of these solutions has similar features and use cases, and each works by emulating the public cloud within your own data center. They involve using the cloud management interface offered by the service provider and using evergreen services always kept up to date by the cloud vendor. While they do cost more than running your hardware stack, you can reduce the effort level to move into a cloud environment by emulating cloud infrastructure.
Advantages of Using a Hybrid Cloud Solution
Over time, more and more organizations are moving their workloads into the public cloud. In working with many clients, I’ve learned that means they want to have single management and deployment experience for their workloads, including those that may not be running directly in the cloud. Whether it’s part of their digital transformation strategy or looking for additional scalability for large-scale applications, these companies are quickly moving forward with their cloud migrations.
In most cases, organizations will benefit from hybrid cloud architecture. They can start moving processes to public cloud APIs and deployment models while still maintaining some systems in their own data center. While some legacy apps may permanently remain in the local data centers, organizations should build their cloud strategy around what is available in the public cloud providers, whether it be Infrastructure as a Service (IaaS), Platform as a Service (PaaS), or Software as a Service (SaaS) solutions.
For a deeper, fuller picture across on-premises and cloud environments, learn more about the
SolarWinds Observability Self-Hosted (formerly known as Hybrid Cloud Observability) built on the SolarWinds Platform, a next-generation, integrated IT infrastructure, application, and database performance monitoring solution offering organizations of any size a comprehensive and cost-effective way to optimize performance, ensure availability, and
reduce remediation time.
