One cloud isn’t enough
For some, that’s a provocative idea. For most, though, it’s obvious. Gartner tells us that over 75% of midsize and large organizations have adopted an approach to IT that’s built on hybrid multicloud: on-premises cloud combined with more than one public cloud.
And cloud isn’t enough without the edge
Now that’s more provocative. But let’s evaluate Gartner’s perspective that the edge completes the cloud. It’s true enough that many organizations face challenges — including needs for ultralow latency compute and data transfer — that can’t be solved by a centralized cloud. As a result, for many use cases, edge will start to complement cloud and core. We’ll continue to see millions of edge infrastructure devices that combine compute, storage, networking, and management software. Edge devices are also aggregating Internet of Things (IoT) data, accelerating video performance over 5G, and crunching data from smart city sensors and airplanes. If you don’t believe us, look at the AWS Snow Family for a hint of the future.
So, what we’re actually seeing is a new strategy emerging for IT: one that spans core to cloud to edge, with one or more private clouds, one or more public clouds, and more than one edge. Does that sound complicated? It is, and to extract full value from this complexity, enterprises want to manage everything as a seamless aggregated collection of services that can be used to support any application anywhere. They want edge to core to cloud hybrid cloud management.
Hybrid cloud management isn’t a baked, matured, standardized approach with a single set of tools; it’s more of a discipline. That’s because so many critical capabilities are still emerging, and new technologies, like microservices orchestration, are still being integrated and improved upon. Yes, there are dashboards and tools that can help. Yes, there are thousands of APIs, tools, and integrations that can help. But there isn’t a one-size-fits-all approach that’s simple and easy.
Having said all that, there is an emerging set of understandings, best practices, and proven tactics that will help developers and leaders cope with change in this rapidly evolving world of hybrid cloud. In this blog post, we explore the pitfalls and best practices around hybrid cloud management. We hope our tips and tricks will help you discern new opportunities and feel confident about extracting more value from your hybrid clouds. You’ll get familiar with the direction of hybrid cloud so that you can position your organization to face the future.
The keys to managing your hybrid cloud environment — and the pitfalls to avoid
Let’s begin with a hypothetical hybrid cloud. Your IT teams have turned their on-premises infrastructure into two private clouds, using Azure Stack because you’re mostly a Windows shop and can harness existing expertise. They’ve chosen AWS for some workloads and Google Cloud for others because of distinctive best-in-class services. They’re extending out to the edge with edge gateways, a few micro data centers, and a handful of AWS Snowballs, all to support IoT aggregation. And everything is massively efficient, stable, and cost effective.
But let’s find the pitfalls and deliver improvements, shall we?
Pitfall #1: Hybrid cloud can be inefficient
When we discuss hybrid cloud with IT leaders, one of their first concerns is inefficiency. Even in a single cloud, everyone knows about the hidden costs of shadow IT and understands that it’s easy to overlook poor utilization, service duplication, and ineffective cost optimization. Compound these problems across your hypothetical Azure+AWS+Google Cloud+edge environment, and you could have a recipe for ineffective and costly cloud infrastructure. Resources aren’t properly utilized, data is stranded on the most costly cloud service, and users struggle to get the resources they need when they’re needed.
You can begin to counter some of that inefficiency with hybrid cloud management that includes service aggregation. The goal is to show services and applications running on multiple cloud environments in one place. By managing resources as a single unit, it’s easier to find overprovisioning, underprovisioning, and stranded resources, and also to compare costs across multiple cloud providers.
Of course, you can’t compare costs without workload and cost analytics. Each cloud provider offers rich, actionable data about what is running and what the costs are. Organizations must work out ways to combine disparate cost information (that’s in different formats) and conduct ongoing analysis to improve efficiency.
And finally, you must automate your hybrid cloud. The overall goals of automation are to reduce the amount of routine work, accelerate movement, and create breathing room for innovation. Automation includes orchestration and self-service capability within clouds. But smart organizations start building self-service provisioning, orchestration, and data or workload migration across clouds as well. Some also use AI-driven decisions to make their hybrid cloud better. Spot by NetApp is a good example of this approach. Spot automates the optimization of cloud resources to minimize manual changes, helping development teams work faster without the constraint of insufficient infrastructure.
Pitfall #2: There’s always a skill gap
In the world of hybrid cloud, new technologies emerge every week, and it’s a full-time job to keep up. Sometimes it seems as though there’s a small niche group of people who know what they’re doing. Most people know only half of what they need to know, and enterprise IT professionals struggle to catch up and keep up. For your hybrid cloud, counting third-party capabilities, open-source tools, and vendor software, you have thousands of services and a lot of inconsistent nomenclature. Just keeping up with the names of things is a hard task. To truly handle hybrid cloud management, you’ll to have to prioritize skill acquisition through training and hiring.
It’s also hard, as new capabilities emerge, to have people skilled enough to discern shifting strategic implications. For example, 5 years ago, only a handful of developers saw the power of containers; now thousands of organizations are redesigning their cloud strategies to take advantage of containerized applications. Staying on the forefront of emerging technologies helps you chart a course into the future.
Smart organizations are always working on filling the skill gap, but they also keep a list of experienced professionals: vendor consultants who can augment and extend in-house capabilities. For example, we have a team of consultants who talk technology, business value, and strategy for organizations of all sizes, every single day. For your hypothetical Azure+AWS+Google+edge hybrid cloud infrastructure, you need all the skills and insights you can muster.
Pitfall #3: Making infrastructure application-led for any app
Banks still have 50-year-old critical applications running on mainframes, so if any organization thinks that their enterprise applications are going away, they should rethink. The trends are to keep enterprise applications alive and well, but power them and manage them with cloud-native applications. Many organizations are working toward managing cloud-native containerized applications and microservices alongside applications in virtual machines and on bare-metal servers. It’s all hybrid cloud, and as such, it’s all encompassed within hybrid cloud management.
But leading organizations are taking a step beyond that approach. In the past, organizations built their infrastructure and then added applications. Increasingly, organizations seek better ways to make their infrastructure application-led. Many of the capabilities we’ve mentioned, including orchestration, automation, and AI, are really there to facilitate the shift toward application-led infrastructure for both enterprise and cloud-native applications. That’s why NetApp acquired Spot. We saw the value of an approach that combines analytics, guidance, and automation to continuously optimize cloud compute and storage for any and every application requirement.
However, tools for service aggregation, monitoring, automation, and control across both enterprise and cloud-native applications are still works in progress. It’s common for an organization to have one set of tools for cloud-native and another set of tools for enterprise. But organizations have tactics that get them closer to an ideal of application-led infrastructure. These tactics include refactoring enterprise applications to become cloud-native, testing emerging toolkits, and choosing underlying technologies that span both enterprise and cloud-native requirements.
Speaking of which…
Pitfall #4: Choosing the right building blocks
Google Cloud has storage services. Azure has storage services. AWS has storage services. Capabilities vary, costs vary, complexities vary. But consistency is good. It’s smart to standardize on cross-cloud services — services that can function across multiple cloud stacks. Or, to put it another way, hybrid cloud management needs hybrid-cloud-specific building blocks.
Your hypothetical hybrid cloud — Azure, Google Cloud, AWS, and more — has dozens of services and hundreds of integrations. Every additional service and integration adds complexity, risk, and cost. There’s a lot to be said for finding services that can span multiple cloud infrastructures.
Organizations that are building a hybrid cloud need a storage service that can span multiple clouds and serve as a hybrid cloud integration touchpoint. Choosing a hybrid-cloud-optimized building block not only offers simplicity and consistency; it’s also a good way to reduce your skill gap and make costs more predictable while maintaining data durability. Can you imagine how much easier your hypothetical hybrid cloud would be with a consistent set of storage services instead of incompatible ones for all the different cloud elements?
And the more versatile the hybrid cloud building block, the better. It should be able to support a wide range of use cases, performance tiers, and SLAs. Your hybrid cloud building blocks also need to be able to support both enterprise and cloud-native applications. Toward these goals, NetApp® ONTAP® data management software has already extended itself into the world of containers. We’ll no doubt see tighter integration between on-premises ONTAP, multicloud ONTAP, and ONTAP for both stateless and stateful applications in Kubernetes.
Oh, and it’s helpful if your building blocks come with open APIs or even their own management capabilities to improve utilization, performance, and cost controls. That’s a place where NetApp excels.
Pitfall #5: If it’s not mature, be careful
Many cloud technologies are emerging, open-source, relatively untested, and certainly unproven for enterprise applications at scale. Smart organizations rely on tested, proven, mature technologies. By choosing mature technologies and then integrating them into highly innovative, rapidly evolving cloud and edge infrastructures, organizations can balance time, cost, and risk. Over the history of cloud, organizations have bumped again and again into unforeseen problems and limitations with cloud services that might be a year or two old. NetApp® Cloud Volumes ONTAP has a track record that other cloud storage capabilities can’t match. It’s proven for enterprise reliability, performance, and value.
And mature technologies come with an added advantage: a rich feature set. NetApp Snapshot™ copies and replication, data protection algorithms, and compatibility with both enterprise and cloud technologies are capabilities that other technologies struggle to meet. That rich feature set helps Cloud Volumes ONTAP support file services, databases, DevOps, or any other enterprise workload.
From a single cloud to independent clouds, to more or less coupled multiclouds, to tightly coupled multiple clouds — maybe all public, or a mix of public and private, along with edge — the world of IT is more complicated than ever before. But it’s clear that hybrid cloud and hybrid cloud management are the ways forward. There are pitfalls, to be sure, but we’re in a world of rapidly improving advantages and opportunities. That’s why NetApp prioritizes accelerating the innovative capabilities of Cloud Volumes ONTAP, which is available on premises in traditional storage and on AWS, Azure, and Google Cloud.
NetApp helps organizations of all sizes standardize and optimize their storage for enterprise and cloud-native applications with the most compelling mix of efficiency, reliability, and value. To learn more, visit NetApp Cloud Central. While you’re there, check out our fully managed cloud file services, NetApp Cloud Volumes Service for AWS, NetApp Cloud Volumes Service for Google Cloud, and Azure NetApp Files.
Use “performance tier” (for SSD aggregates) or “cloud tier” (for object storage). Use “capacity tier” instead of “cloud tier” if applicable.