Why there's no compelling use case for hybrid cloud

Don't confuse hybrid infrastructure with hybrid cloud. With no real reason to integrate public and private cloud solutions, hybrid cloud isn't a worthwhile undertaking.

Image: iStockphoto/JackyLeung

The VMware Cloud on AWS solution has reignited the cloud washing conversations. One such conversation has centered around hybrid cloud, which begs the question of if there's a viable path to hybrid cloud in the enterprise data center.

Alternatively, is the pursuit of managing public and private cloud as an integrated solution a fool's errand?

The hybrid myth

One of the myths of hybrid cloud is the manufactured desire for enterprises to expand their private cloud resources to the public cloud. A given typical use case is the ability to cloud burst--a concept where enterprises "burst" to leverage public cloud when private cloud resources run low.

SEE: Hybrid cloud isn't cloudy enough, suggests new Forrester data

In a recent podcast conversation, IBM Cloud developer advocate Tyler Britten questioned the business value of expanding workloads from a private data center to the public cloud. In his experience, customers don't extend to public cloud during the most critical periods of business. The example he gave was a marathon organizer.

The one time of year that the marathon needs capacity, they don't rely on an external provider to host their most critical workload--the registration site. The organization does the opposite and steals capacity from other internal workloads to ensure availability and performance. The organizers are much more comfortable with the systems they are most familiar with operating.

Another use case is the centralized management console. In the centralized management console model, developers write to a single API and the cloud management platform determines the best environment to run the target workload. Again, an interesting idea, but overly difficult to implement in practice. It's too difficult for the average enterprise, at least.

Limited private cloud

I'm not in the camp that believes private cloud is a non-existent or shrinking market. Private clouds have a purpose in the enterprise. According to Britten, a common use case for IBM Cloud's managed OpenStack solution is to bring AWS-hosted apps into the private data center. Point solutions are another use case for private cloud.

During the 2015 HPE Discover conference, the Helion Cloud team shared that the typical use case for Helion Private Cloud was for single applications. While a platform such as Helion is capable of hosting multiple applications, it's not uncommon for customers to deploy independent Helion solutions for each cloud-native application.

The use case that never took off is the migration of legacy workloads to a private cloud operating model. The desire to reduce or eliminate vSphere licensing was an early driver for private cloud. Some customers had a misconception that platforms such as OpenStack would allow the seamless workload management across public and private clouds for traditional workloads. Just as public cloud isn't an efficient location for legacy applications, private cloud proved to be a poor fit for legacy applications.

Without the need to manage a large private cloud, there's no incentive to build a complex integration between private and public cloud for a non-existent management requirement. As such, there's no need for chasing the hybrid cloud use case at scale. The hybrid cloud use case involves using a single pane of glass to manage or proxy public and private cloud infrastructure.

In short, there's no compelling need for hybrid cloud.

Also see