The cloud infrastructure is a composition of two or more distinct cloud infrastructures (private, community, or public) that remain unique entities, but are bound together by standardized or proprietary technology that enables data and application portability (e.g., cloud bursting for load balancing between clouds).
Are We Ready for Hybrid Cloud Computing Environments
There is a view that customers are neither demanding, nor even considering, the development or implementation of such connected systems that deliver data and application portability. The complexity of implementing and maintaining such a system is significant and suited for those at the higher end of the IT Maturity model. There may come a time when this is the distinguishing element for customers who need to incorporate integrated processes, but at present connectivity and usability between on-premises and public cloud entities are the primary focus of most customers.
How Cloud Computing Vendors View Hybrid Clouds
Vendors seem to take such liberty with what “hybrid” means that it has come to imply any connectivity to any cloud or hosted environment. At the September 2015 AzureCon event, Microsoft described the four hybrid workloads that make up 50% of customer workloads in the cloud: Backup & Disaster Recovery, Development/Test, Lift & Shift, and Cloud Native Application Development. VMware vCloud Air defines Hybrid Use Cases very similarly: Development/Testing, Extending Existing Application (migration), Disaster Recovery and Creating Next-Generation Applications.
It is true that Microsoft and VMware intended those workloads to be examples of use cases that can make hybrid cloud a valuable asset. Many system integrators make reference to hybrid solutions. In a study done by Avanade in 2014, Global Study: Hybrid Cloud – From Hype to Reality, the C-level suite, including IT decision-makers, did not understand what hybrid cloud really was. However, 74% surveyed said they’d “be likely to adopt a hybrid cloud solution.” They are willing to adopt a solution that they don’t even understand!
Does that even matter? I like the term “hybrid computing” because it removes the cloud portion of the requirement. This is important because most customers do not have a true private cloud. They have virtualization to be sure, but it is usually not automatic or elastic and is rarely a measured or on-demand self-service. Usually, what they do have is a solid, cost saving, efficient virtual platform. This is the platform the executives and IT decision-makers see value in extending to the cloud to reap advantages. These are the scenarios that Microsoft and VMware talk about when discussing their workloads in a hybrid environment. Can you do a disaster-recovery workload to Azure? Definitely! You can use Azure Site Recovery to connect to Azure (or another site) to replicate virtual machines. You can replicate P2V, Hyper-V or VMware, and database replication in many forms, all tied together in a Site Recovery Plan, or runbook. Similar solutions exist with VMware, AWS and most major vendors.
Those solutions do not require the NIST definition of hybrid cloud to be successful. They just need a virtualization stack and connectivity. This leads us to the question, Is an SaaS solution a hybrid compute solution? It certainly does not fall into the category of a hybrid cloud. What is hybrid computing? I define as follows:
Hybrid computing is the integration of one or more public clouds with an on-premises virtualization environment.
How Critical Is Cloud Virtualization to Hybrid Clouds?
Some would argue that the virtual environment is critical. Given that a true hybrid cloud connects a public cloud with a private cloud, in this definition, the private cloud does not have to be a true cloud. The public cloud does, however, need to be connected to a virtualized environment. If we expand the definition to include any connectivity, then hybrid would fit a lot of configurations. For example, colocation, with connectivity directly to the data center, could be considered hybrid computing. For that matter, another branch office could be so considered as well. In addition, SaaS solutions like Box.com, Cisco Meraki, Salesforce.com and even WebEx can all be considered hybrid computing. The limitation of imposing virtualization tends to filter out a lot of noise and keeps IT focused on specific architectures that are designed to eventually all work together. That is the goal.
In the four workloads described above, Microsoft assumes the use of virtual servers. Disaster recovery replicates virtual machines to an Azure (or other location) subscription. With the integration of InMage, it also supports a physical to virtual (P2V) replication. It wouldn’t make sense to do a P2V without the associated virtual server replication. The application development/test environments are also intended to work with virtualization since it would not make sense to develop applications in a public cloud that cannot work on virtual servers. Are you required to include the components and capabilities that NIST suggests? No. You can back up to the cloud, replicate your VMs for disaster recovery, develop new applications or test old ones quite nicely without having data portability or cloud bursting. The goal is to get to a point where there is a true private cloud connecting, sharing and integrating with a public cloud. When an application can reach out to the cloud, grab what it needs, share its data and move on (maybe to another cloud for another purpose), then we will be in a true hybrid cloud architecture.
Sometimes our industry gets granular on specific definitions, and IT needs to technically understand what is being talked about and how to implement and support it. When cloud first emerged, there were very confusing conversations going on about what the cloud is. That is when NIST stepped in to try to focus the conversation in order to move forward with solutions. In the case of hybrid cloud, the definition became so specific that it rarely fit the average business. In such situations, people started using the term “hybrid” in a generic sense that fit their situation. Hybrid computing is a term that can help separate the purist from the average business owner and allow solutions to move forward in a more fluid state. The strategy is to move to a true hybrid architecture; the goal is to solve problems. What we call it sometimes doesn’t matter.