24 March 2020
There is so much jargon in the cloud world, sometimes its hard to follow the conversation! Different vendors use the same words to describe different things, and different words to describe similar things. And the meanings are evolving over time. One of the most-used and abused phrases today is some variation of ‘Hybrid Cloud’ or ‘Hybrid Multi-cloud’. Is this just hype or some vendor fud (fear, uncertainty and doubt)? Or is it a real thing?
My goal in this blog is to explain the term, and highlight the criticality of Hybrid Cloud to your enterprise computing strategy. Not only is Hybrid Cloud a critical interim step in the goal of getting to the cloud, I believe it is going to part of the long-term strategy for most large enterprises.
What is Hybrid Cloud Anyways?
NIST – the US National Institute of Standards and Technology defines the Hybrid Cloud as “a composition of two or more clouds (private, community or public) that remain unique entities, but are bound together, offering the benefits of multiple deployment models.”
While NIST has the benefit of being a standards organization, I think more commonly, people use the term ‘Hybrid Cloud’ most frequently to describe systems spanning public cloud and private or dedicated cloud, or even public cloud and traditional on-premise solutions, rather than any two clouds as NIST defines. That’s how I’ll use it here, but I will come back the NIST definition at the end. To highlight the point that this term is still evolving, I’ve collected some other example definitions at the end of the blog.
So for the purpose of this article, “Hybrid Cloud” is where different parts of an application system work together and are spread across your on-premise environment and one or more public cloud. (See diagram) That’s how most people are using the term today. “Separate clouds become hybrid when those environments are connected as seamlessly as possible.”Some Hybrid Cloud Examples
To understand where you might want to use Hybrid Cloud, think about the difference between your on-premise environment and Public Cloud. You have more control over the on-prem environment since it runs in your data center, you can run transactional workloads fast, but it is not elastic like Public Cloud is. If you need to scale on premise, then you need to buy new hardware, where as on Public Cloud (and dedicated), if you need to scale you just add more resources and that provisioning is handled separately. Access to the on-premise environment is at the speed of your local network, access to Public Cloud is at the speed of your internet connection.
One of the simplest and most common examples of hybrid cloud is where your production environment remains on premise in your data centre but you’ve moved your development and test environments out to the cloud. That gives you the flexibility to create and destroy test environments as needed. As someone who’s led a lot of large complex systems integration projects, we used to beg for ‘one more test environment’, and as soon as we got it, we’d want another one!
This ‘prod on-premise, test in the cloud’ is a Hybrid Cloud example because to work effectively, the environments need to be connected with a DevOps toolchain that spans the on-prem and public cloud environment. And the big benefit – besides keeping your test team happy – is that you only pay for those added environments while they are in use, and can eliminate the cost when you are in a cycle of low or no testing.
A second Hybrid Cloud example and one step up on the complexity curve would be an on-prem environment that is using a cloud-based API. Perhaps your application is calling out to Watson APIs, or needs access to Google Maps. In that scenario, Watson or the Map APIs are a small Public Cloud-based component of your overall solution. Data is mostly flowing out to the cloud, with a small answer set coming back.
Getting to the highest end of a complex Hybrid Cloud application, you have applications that share large amounts of data. Perhaps an ERP system that is on prem, is updating a data lake in the cloud. And that ERP is also communicating with a CRM solution like Salesforce on the cloud for quote or order management so that manufacturing and distribution understands backlog requirements.
This third example starts to show why hybrid will be around for a long time. As more applications move to the cloud; more and more of them will need to communicate with those ‘left behind’.
For lots of reasons it will take years to get all large enterprise applications to the cloud, if ever. Some applications will be considered ‘crown jewels’ or contain data too sensitive to move out, some applications that have evolved over years may be too complex to move all at once, and for some applications that have been stable for years, there may be no value in migration.
So what’s not Hybrid? You’re not on Hybrid Cloud, for example, “if a company is using a SaaS application for a project but there is no movement of data from that application into the company’s data center.” (Dummies.com) That’s just a single cloud, working standalone from the rest of your application portfolio.
When is Hybrid Cloud a Problem?
Hybrid Cloud is a computing pattern and like other patterns, it isn’t right for every situation.
Of course, if you’re organization started as ‘born on the cloud’, there is probably little likelihood that you will ever implement this solution. There isn’t likely a use case that will force you to build your own data centre if you don’t already have one.
Particularly as the solution gets more complex, you need to be really concerned about the amount and timeliness of data movement between the various environments. If there is a requirement for real-time or near real-time, bandwidth and latency can become significant issues. “Managing hybrid cloud is a complex task because each cloud solution has its own API, storage management protocols, networking capabilities, etc.” (Citrix)
How to Manage Hybrid Cloud
When you have a simple hybrid environment, you probably don’t need any new tools or techniques to manage your Hybrid Cloud. Your existing IT management processes will probably be fine. As more and more on-premise solutions are working in tandem with on public cloud solutions, it gets more complicated and you are more likely to need tools to manage hybrid cloud and/or multi-cloud environment that results.
The tooling to manage Hybrid Cloud and multi-cloud is still new and evolving. Clients are investing in their own custom solutions and vendors are building out capabilities. As your hybrid environment grows more complex you will need tools that provide
- Visibility – With workloads running in multiple clouds and on premise, you need a “single pane of glass” that tells you exactly what is running, and where it is running.
- Management – Tooling that gives you the ability to set security policies and track spend; and manage workloads based on those needs. Orchestrate how applications start, connect to each other, and scale.
- Automation – Automation will allow you to deploy applications across environments, help manage backup and disaster recovery, and provide the ability to move workloads from on-premise to cloud or vice-versa.
Focusing on these will help manage systems that have already been built. As you’re building new applications or modernizing existing ones, you will need ways to rapidly develop integration methods and data movement across cloud environments rapidly and consistently.
To make it easier to manage workloads and have consistent tooling across private cloud and public cloud, common tooling such as OpenShift from Red Hat / IBM or Anthos from Microsoft can help. The promise of these tools is that they will allow you to orchestrate all of your cloud environments using a single interface. Built on open source tools such as Kubernetes and Linux, they offer built-in governance and help manage orchestration across the Hybrid environment. Increasingly, Kubernetes and containers are the de facto organizational process for new application and microservices deployments, so these tools become increasingly critical.
In addition to OpenShift, other companies such as Microsoft (Anthos), Citrix and VMware are all offering competing management frameworks and one of the questions an IT department will need to answer is choosing between these platforms.
Future of Hybrid Cloud
I think it is clear from the examples above that hybrid cloud will continue to evolve over time. Eventually it is likely to be indistinguishable from ‘multi-cloud’ – two or more different clouds collaborating in an application or system. Which is basically what the NIST definition states.
For now, Hybrid Cloud offers benefits that can’t be gained with just on-premise or just ‘on-cloud’ solutions:
- It gives business comfort around security and compliance issues – with the right level of security for production by maintaining on-premise for high-security applications, and allowing development and test environments and lower-security needs systems to be on public cloud
- Integration with the remaining legacy on premise environments is easier (and higher-performance) from private cloud, which facilitates migration. Until a critical mass of the applications that communicate together are ready for cloud, you can keep the portfolio together using private cloud, then gradually move the ones to public when its practical).
- Hybrid Cloud offers scalability / workload management (ie, test environments, development environment, temp performance testing or training environments) that your on-premise data centre can’t match.
- Hybrid Cloud gives the potential for cost-optimization between existing assets and public cloud environments
- Embracing Hybrid cloud gives your on-premise environments access to public APIs and microservices such as AI technologies
- Hybrid Cloud gives your IT department the opportunity to balance among the competing goals of control, portability, scalability and cost
- Hybrid Cloud allows applications to move across boundaries over time and to operate across boundaries
- For more on benefits of Hybrid Cloud see here (RCR Wireless) and here (Gigamon)
New use cases are coming along that will bring new forms of Hybrid Cloud. One example is IoT (the Internet of Things). As IoT becomes more mainstream we will need solutions that deal with collecting and sorting all the data collected by IoT devices, and sending back for analysis summarized or most critical data. While there are multiple options for this, one of the most likely will be a ‘hub and spoke’ model where IoT devices communicate with a local, private server, and that device then decides what information needs to be shared to the cloud. This hub and spoke is an instance of Edge Computing, where selected processing is moved closer to the data being collected. There is a great example of how this will work using an example of IoT data collection of a motorcycle and rider example here via TechBeacon.
Some would say that coming rollout of 5G networks will resolve this capacity problem, but particularly in a large country like Canada the rollout of 5G will likely be focused on the largest cities and some narrow corridors between them. So the hub and spoke model will continue to be in use for a long long time.
Hybrid Cloud is clearly a flexible and evolving term. The concept of private or dedicated technology working together with systems built or migrated to the public cloud will persist for a long time. Indeed, Hybrid Cloud usage is growing as companies continue their migration to the cloud. “The global hybrid cloud market is expected to grow from USD 44.6 billion in 2018 to USD 97.6 billion by 2023, at a CAGR of 17.0% during the forecast period.” (Markets and Markets ).
Emerging technologies like IoT and Edge computing will bring new Hybrid Cloud use cases over time. The term Hybrid Cloud will continue to evolve but the concept – your own technology interacting with public cloud solutions – will continue to be a valuable part of your IT environment for years to come.
Addendum – Other Hybrid Cloud definitions
As I did some research for this blog, I saw that there is no consistent definition for Hybrid Cloud. And its not just the words that vary, the scope described varies. See the examples below, all of which differ in meaning from the NIST shared at the start of this article. Some would argue this comes from the fact that ‘Hybrid Cloud’ is a term that comes from the vendor community, not from clients or some academic pursuit. But I think when people say that they are trying to imply that Hybrid cloud isn’t valuable or useful or that it ‘isn’t really cloud’. I hope I’ve shown that there is more to hybrid than a short-term or interim step on the way to Hybrid Cloud. The multiple definitions point more to the fact that hybrid cloud is evolving, like most cloud technologies are today.
IBM definition “Hybrid cloud is a computing environment that connects a company’s on-premises private cloud services and third-party public cloud into a single, flexible infrastructure for running the organization’s applications and workloads.”
Forrester Research‘ definition of Hybrid Cloud: “One or more public clouds connected to something in my data center. That thing could be a private cloud; that thing could just be traditional data center infrastructure.” (As found on this BackBlaze blog post. )
TechTarget definition: “Hybrid Cloud is a cloud computing environment that uses a mix of on-premises, private cloud and third-party, public cloud services with orchestration between the two platforms. By allowing workloads to move between private and public clouds as computing needs and costs change, hybrid cloud gives businesses greater flexibility and more data deployment options.”
If you want a broader discussion of the variation in the meaning of the term Hybrid Cloud, there is a good summary in this IDC Report.
Other Links on Hybrid Cloud