Review hybrid cloud offerings that bring the cloud on premises

Many organizations are drawn to the benefits of public cloud services, but not all of them can ditch their on-premises environments, infrastructure or legacy applications. Connectivity, regulatory requirements and latency concerns are just a few of the factors that prevent certain companies from moving fully to the cloud.

To cater to enterprises unable to host all their workloads in the cloud, public cloud providers developed hybrid cloud offerings that push their services into on-premises environments. These efforts to bring the public cloud on premises have enabled organizations to choose which workloads to run in the cloud and which to keep in the data center. Those organizations can then link the two environments with dedicated networking.

Get to know the hybrid cloud and on-premises infrastructure services from AWS, Microsoft and Google that aim to bring the cloud directly to customer locations.

Google Anthos. Google Anthos is a hybrid cloud container environment built for consistency across the cloud and on premises. Launched in April 2019, IT teams use this software to manage cloud technologies -- particularly Kubernetes containers -- on internal hardware and legacy applications.

Anthos is based on container clusters that run on the managed container service Google Kubernetes Engine (GKE). Users can access Anthos in the cloud with GKE and in data centers with GKE on-prem. Anthos runs on any server that supports Kubernetes. This Google Cloud service offers multi-cloud functionality, with management capabilities for workloads on AWS and Microsoft Azure, as well as other third-party options. This enables users to get the benefits of a public cloud deployment without being locked into one cloud platform.

A core component of this hybrid cloud offering is Anthos Migrate, which automates the migration of legacy virtualized workloads in the cloud or on premises into Kubernetes containers. Anthos also comes with a set of management, monitoring, security and authentication tools bundled for on-premises and cloud environments. Anthos' key features include Anthos Config Management for cluster administration and management, Istio on GKE service mesh, Strackdriver monitoring, Cloud Run for event-driven workloads and more.

Google has partnered with hardware providers -- including Cisco, Dell EMC and NetApp -- to bring Anthos to more customers. These partners provide Anthos systems that will be available on hyper-converged infrastructure.

AWS Outposts. Outposts is the on-premises version of AWS cloud infrastructure. Outposts are specialized racks that house AWS-designed compute and storage servers, as well as some AWS services and tools. AWS delivers, installs, maintains and updates these racks.

Each Outpost is connected to a local AWS Availability Zone for access to all of the tools and services available in that Region. Outposts are built for organizations that want to utilize the AWS platform but have low latency or regulatory needs that require workloads to remain in their data centers.

Outposts are designed to provide many of the AWS tools, services, APIs and security controls in on-premises environments. Supported services currently include a range of EC2 instances, Elastic Block Store, Amazon Relational Database Service and AWS CloudFormation. AWS plans to expand the supported offerings over time.

There are two versions of AWS Outposts: AWS native Outposts and VMware Cloud on AWS Outposts. The native version, which launched at AWS re:Invent 2019, is managed entirely by AWS and accessed through the Management Console. The VMware Cloud variation, which is still in beta, runs VMware's software-defined data center software and enables customers to use VMware control planes and APIs.

Unlike some other public cloud providers' on-premises services, AWS Outposts is not a multi-cloud offering. The AWS hybrid cloud platform is a single-vendor service that does not support workloads in other clouds.

Azure Stack. Microsoft launched Azure Stack in 2017 as a way for users to host Azure cloud services on premises. The initial Azure Stack service has since expanded into a portfolio of three services. Each one is designed to tailor Azure capabilities to different infrastructure needs and environments.

• Azure Stack Hub. This hybrid cloud offering enables users to host applications in the cloud and on premises, or entirely in their data center. Hub offers tools, services and management options for on-premises computing that's consistent with Azure's public cloud environments.
• Azure Stack HCI. This version of Azure Stack runs local VMs on hyper-converged infrastructure (HCI). Azure Stack HCI is based on Windows Server 2019, which provides the compute, storage and networking technologies in a software-defined architecture. Azure Stack HCI uses hardware provided by OEM vendors, such as Dell or Lenovo. Azure Stack HCI also enables users to access Azure cloud services on premises.
• Azure Stack Edge. Azure Stack Edge is an appliance for users that need on-site computing, typically outside a traditional data center. Microsoft manages, installs and updates the devices. The appliance runs managed VMs, containers and machine learning models at edge locations. It brings Azure compute, storage and other cloud capabilities directly to users. It also provides low-latency data processing and high connectivity.

Azure Arc. Microsoft's latest hybrid cloud service is in the preview stage of development. IT teams use Arc to deploy Azure services and management capabilities in the cloud or on premises. Azure Arc can control Windows and Linux server farms on premises. It also supports Kubernetes container deployment and management across cloud and on-premises environments. The Azure Arc preview has built-in support for Microsoft's public cloud data analytics services, such as Azure SQL Database and Azure Database for PostgreSQL Hyperscale.

Arc is based on Azure Resource Manager, the framework that manages deployment, administration and governance of Azure resources. However, Azure Arc expands the capabilities of Resource Manager into multi-cloud and hybrid cloud environments.