As cloud computing services become more powerful and varied, most senior IT managers are contemplating a hybrid approach to the cloud. This is not a new topic, but the continuing evolution of cloud solutions should lead to refresher on the subject every six to twelve months. Last year’s hybrid cloud is not the same as the one you might be considering today.

What it is Hybrid Cloud?

In simple terms, “Hybrid Cloud” refers to an architecture that blends cloud computing with traditional, on-premises IT. This usually means extending a legacy system into the cloud. Integration is implied, though not essential, for a hybrid cloud. For instance, a hybrid cloud architecture usually involves deploying some functions or modules of an enterprise application in the cloud, while other modules remain in the data center. However, some IT departments run completely separate, independent systems in the cloud and on-premises. This is sometimes call “Hybrid IT” or “Two-Track IT.”

Hybrid Cloud vs. Multi-Cloud

There is some confusion among IT professionals about the respective definitions of Hybrid Cloud and “Multi-Cloud.” The confusion is legitimate. People have adopted different definitions for the same terms. For some, a multi-cloud architecture is one that combines more than one cloud service provider (CSP), e.g. AWS, Azure and Google Cloud all running at once for a single corporation’s IT needs. Others call this arrangement hybrid cloud. For our purposes, hybrid cloud means linking on-premises systems with those in the cloud.

Modes of Hybrid Cloud Deployment

Hybrid cloud deployment can occur in multiple modes. It’s sort of like a recipe, with a few basic ingredients, but many different combinations. The main ingredients include:

  • Public cloud platforms, e.g. AWS or Microsoft Azure
  • Private cloud, hosted on the system owner’s premises (e.g. cloud infrastructure, with a cloud software architecture, running in a private data center)
  • Private cloud hosted by a third party, but managed by the system owner, e.g. at Rackspace
  • Managed private cloud, where a dedicated third party manages the private cloud for the system owner
  • Cloud variants:
    • Infrastructure-as-a-Service (IaaS), where the system owner sets up whatever infrastructure elements he or she needs, e.g. compute, storage, operating system, etc. on a cloud platform (public or private)
    • Platform-as-a-Service (PaaS), where the system owner uses a platform like Windows Server in the cloud (public or private)
    • Software-as-a-Service (SaaS), which comprises a complete software application running from the cloud, e.g. Salesforce.com. SaaS is almost always a public cloud-based service.

With these various options available, you can imagine how complex a hybrid cloud can get. The combinations are many. You could have an on-premises application integrated with an application running on PaaS or IaaS. It could integrate with SaaS, or with software or data in a private cloud, and so forth.

Hybrid Cloud Use Cases

There are as many potential use cases for hybrid cloud as there are architecture choices. Each organization will have to assess its IT needs, current and future, and choose the best path forward. This may mean avoiding hybrid cloud for the time being. However, it worthwhile to understand how others are using the architecture.

Some companies are utilizing hybrid cloud for general-purpose workloads. For example, it is possible to deploy one module of an Enterprise Resource Planning (ERP) solution in the cloud, while others remain on-premises. You could have your accounting and financial management application in the data center, but connect it with a warehouse management module in the cloud.

Others are making the most of the hybrid model for specialized use cases that are difficult to replicate in a private data center. For instance, if you want to start using Hadoop for analytics on massive data lakes, it may be preferable to set that up in the cloud. Or, if you want to manage a large number of Internet of Things (IoT) devices, a cloud IoT platform could be a better option than setting up such a system on-premises.

Reasons it’s worth considering Hybrid Cloud

Why should you consider hybrid cloud? A number of compelling reasons have emerged as companies adopt this architecture. The most basic reason is practical reality. It’s simply not realistic to move a company’s complete collection of digital assets to the cloud all at once. Even if you wanted, it would be a nearly insurmountable task. And, it would not be a wise move. You’d be asking for business disruption.

Thus, it makes sense to move some of your systems to the cloud and run on a hybrid basis for as long as it takes to move everything into the cloud—if that happens to be your strategy. It doesn’t have to be. There’s a lot of wisdom in keeping certain things running in-house for as long as they are operating. A system that’s running well on-premises and enjoys the attention of experienced IT staff does not have an urgent need to migrate to the cloud. It’s only when systems reach end of life or people start retiring that the cloud starts to look appealing.

There are also some platforms that are difficult to run in the cloud, or are unavailable, e.g. mainframes or iSeries (AS400). Or, if you have a specific performance requirement, it may be challenging to get what you need in the cloud. The financial industry, for example, expects stock trading transactions to clear in milliseconds. Even very small delays can cost money. These institutions usually run “bare metal” servers on-premises to deliver the service levels they need. However, these companies may have other systems that could easily run in the cloud, e.g. customer marketing databases. They don’t need to tie up expensive Wall Street data center racks with such systems, so they can put them in the cloud and go hybrid.

Security and compliance concerns keep some IT assets on-site. Healthcare records, for example, may not be stored in a public, multi-tenant architecture like AWS. National security is another case of this. Companies subject to privacy laws like GDPR may want to keep personal data on-premises so they can be 100% sure they know where it’s located.

Alternatively, some solutions are nearly impossible to do on-premises. Disaster Recovery-as-a-Service (DRaaS) is one such workload. Cloud platforms are great for setting up mirror instances of on-premises systems and databases. You can replicate your IT assets in real time to more than one geographically remote backup site. Upon failover, you can bring up the cloud DR instance with little or no disruption to your enterprise system availability.

The cloud has a learning curve. The hybrid approach gives your IT organization a chance to get familiar with cloud concepts in an incremental way. By moving a few applications or databases to the cloud at first, everyone can learn about configuring virtual machines in the cloud, monitoring them, securing them, setting up access controls and so forth.

Economics is a big driver of hybrid cloud. For a company that has maxed out its data center, setting up cloud instances is a very appealing alternative to constructing a costly new facility. Similarly, some businesses want to get out of the capital expense (CapEx) cycle required to run IT on-premises. The cloud is all operational expenses (OpEx), which gets the company out of borrowing, leasing or using cash assets to upgrade IT every three to five years.

Having a Cloud Strategy

As you can see, there’s a lot to the hybrid cloud topic. This article has barely delivered into the nuances and complexities of the issue. The important thing, though, is to start thinking about what it means to your business, if you haven’t already. It’s essential to have a cloud strategy today, even if that strategy is “we’ll start making a plan next year.” Standing still is not a sound choice at this point.