Many large enterprises still aren’t ready to move into the public cloud space due to security concerns, questions about intellectual property rights, and current investments in hardware and data centers. In response, public cloud providers are changing the way they enable organizations to take advantage of the power of cloud technologies. AWS, Azure, Google Cloud, IBM, and VMware have responded to this available market share with enterprise tools and platforms that organizations are starting to use in order to accelerate their delivery of value to customers.
Hybrid cloud solutions in particular leverage both public cloud resources and an organization’s existing physical data centers. Due to the rise of on-premise cloud platforms and the popularity of Kubernetes, today hybrid cloud solutions can be integrated into an existing enterprise infrastructure that runs the pre-built services and tools that public cloud providers supply. As a result, enterprises can begin to adopt new tools and technologies within the confines of their own data centers while also protecting their data and IP. In addition, they can start cloud-native development inside their organization and then migrate workloads to the public cloud sector when they’re ready. Basically, hybrid solutions enable them to optimize existing hardware investments and contracts and decide whether to move to a public cloud infrastructure or continue with their own hardware. (It’s estimated that the hybrid cloud market will be worth nearly $97.6 billion by 2023.)
Below, I’ll discuss how five major technology providers are responding to enterprise hesitation about moving into the public cloud with forward-thinking hybrid cloud solutions. Understanding the solutions these providers offer (or will soon offer) can help enterprises better understand the current cloud technology landscape and accelerate their secure, cost-effective migration to the public cloud.
Microsoft was the first public cloud provider in the big 3 to provide on-premise hybrid cloud solutions to the enterprise. Azure Stack is composed of hardware and software purchased from a preferred partner like Dell EMC or HPE and software services billed via EA or CSP from Microsoft. They offer several options: (1) the Microsoft Azure Stack Development Kit, a single-node, non-production, proof-of-concept deployment; (2) Azure, an appliance purchased from an approved vendor that includes support and enables you to consume Azure Cloud services (Microsoft manages software and services on a usage basis, and the vendor manages the hardware); and (3) Azure Stack HCI, vendor-certified hardware with the optimal Hyper-V and Storage Spaces. The HCI stack runs on top of your Azure Stack and allows full access to Azure Cloud services. Both Azure Stack and Azure Stack HCI are production-ready, stack-integrated systems that have Azure software pre-installed and configured to run in production. Azure Stacks also require connectivity back out to the Azure Portal for management and use.
Split support (Microsoft for the software and services and the vendor for hardware) means you may have some difficulty getting support and upgrades. The Azure Stack model also keeps the current vendors in place instead of empowering you to outsource hardware and software maintenance. Note, too, that as of today Azure Stacks do not support Kubernetes services.
Google is the second public cloud provider in the big 3 to provide on-premise hybrid cloud solutions to the enterprise. Google has joined the enterprise hybrid cloud space, announcing Anthos at their latest conference. Google has taken a different approach to on-premise cloud services than AWS and Microsoft by providing only software and services and focusing on providing Google Kubernetes Engine (GKE) on premise. You will need to provide any hardware you wish to use, along with VMware vCenter 6.5 to create the VMs that will support the GKE on-prem cluster and F5 Networks BIG-IP load balancers to handle Layer 4 networking requirements. You will also need enough VMware licenses to support the VMs for Anthos and the necessary storage and compute capacity.
Anthos software and services are licensed by vCPU and can be purchased in 100 vCPU blocks with an annual contract at $10,000 a month. An enterprise-level support contract from Google runs an additional $15,000 per month or a percentage of your total spend, whichever is greater.
Amazon is the last of the big 3 cloud providers to offer an on-premise solution for cloud services. AWS Outposts is a hardware-based hybrid solution that comes with the AWS services that you choose from your AWS console (in other words, Outposts won’t allow you to leverage existing hardware capacity). AWS Outposts will be shipped directly to your data center and need to be plugged into power and networking in racks in your data center or co-location. It has two variations: (1) VMware Cloud on AWS Outposts, which allows you to use the same VMware control plane used in your data center today, and (2) the same AWS-native hardware and services that AWS provides in their public cloud offering.
Hardware and services will be accessed via the AWS Management Console, so AWS Outposts will require network connectivity between your data center and AWS. This means you’ll need to either purchase AWS Direct Connect to the AWS availability zone nearest you or set up a VPN over the internet to AWS. (We recommend AWS Direct Connect to ensure consistent connectivity and low latency.) All patches and software upgrades will be applied to the Outposts by AWS, and you’ll be alerted when these updates will occur.
AWS Outposts supporting Amazon ECS(Elastic Container Service), EKS(Elastic Kubernetes), EMR(Elastic Map Reduce), and RDS(Relational Database Service) are expected to become available in late 2019, though pricing for these services has not yet been released.
IBM Cloud Paks
IBM Cloud Paks were announced in early August after IBM’s acquisition of Red Hat. This hybrid cloud packaged solution runs on Red Hat OpenShift on either IBM Cloud or on-prem Red Hat Enterprise Linux. It includes containerized IBM middleware and common software services. Each targets a specific use case and can be combined for multiple uses. The initial System Cloud Pak is required to install and set up the additional Cloud Paks that IBM offers. It is designed to work on Intel and IBM Power Architecture platforms and with workloads on REHL, IBM Power OS, and on IBM AIX. You will also need a subscription to IBM PureApplication System to implement the other subscription-based Cloud Paks: Applications, Data, Integration, Automation, and Multicloud Management. The System Cloud Pack includes VMware and supports Kubernetes.
At present, it’s not clear if you have to use OpenShift or if you can use the standard Kubernetes setup. There is also reference to using Helm charts within the OpenShift/Kubernetes Cloud Pak. To implement an IBM Cloud Pak on premise, you will need to provide your own hardware licenses for RedHat Enterprise Linux. Licensing ratio pricing will be based on the virtual processor cores that the Cloud Paks are running on. Detailed subscription and licensing costs aren’t yet available.
VMware Project Pacific
VMware recently announced the future release of Project Pacific, a complete re-architecture of VSphere with Kubernetes as the control plane. Designed to make everything look like a Kubernetes cluster, it has declarative syntax for managing cloud resources like VMs, disks, and networks. While everything will still look and feel like the older VSphere, it will include the capability to manage the entire application and not just the underlying VMs. Project Pacific allows an organization to manage both containers and VMs inside of VSphere and embed both containers and VMs in a K8s namespace. It’s also partnered with VMware Tanzu, which is a product to manage Kubernetes clusters across all cloud providers. It will also introduce Supervisor Clusters, a specialized kind of K8s cluster that uses ESXi as its worker nodes instead of Linux. A Kubelet (Spherelet) is directly integrated into ESXi. There is also a built-in, centralized container registry service.
Hybrid Cloud Solutions for the Enterprise
Of the five approaches described above, three stand out: AWS Outposts, Google Anthos, and VMware Project Pacific. These innovative solutions all represent a change in how enterprises are implementing cloud services. AWS Outposts are full-package solutions composed of both hardware and software services with support for Kubernetes. Google Anthos is software solution that you can install on your existing hardware to implement Kubernetes. VMware’s Project Pacific brings new services and offerings to customers already leveraging VMware’s products and builds on a Kubernetes-based foundation.
The biggest takeaway: an increasingly large segment of the enterprise market is starting to adopt cloud-based services. In response, vendors are now competing for both public and on-premise cloud offerings for the enterprise. We will continue to watch how these hybrid cloud solutions are implemented at enterprise organizations in order to make a deeper assessment going forward. In the meantime, if you have any questions, reach out!
Liatrio is a collaborative, end-to-end Enterprise Delivery Acceleration consulting firm that helps enterprises transform the way they work. We work as boots-on-the-ground change agents, helping our clients improve their development practices, react more quickly to market shifts, and get better at delivering value from conception to deployment.