BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
The term hyper-converged cloud can be applied to a public cloud provider using hyper-converged infrastructure, an on-premises hybrid cloud using HCI or even a system that links the virtualization in an enterprise HCI system with a public cloud's own virtualized servers.
The core principles of HCI are also core to the cloud: the virtualization of resources to virtual machines (VMs). In the case of the enterprise, those VMs are most commonly end-user machines in a virtual desktop infrastructure (VDI). In the cloud, they are usually virtual storage and file servers, and even virtual production machines.
HCI in the cloud
Cloud giants, such as Amazon with Simple Storage Service and Google Cloud Platform, buy the technology for their virtualized servers in massive quantities. This makes the HCI appliance less appealing for the giants than it would be for a smaller cloud service provider, since the larger CSPs wouldn't get the same cost benefit of adopting HCI.
A hyper-converged cloud can also help alleviate one of the problems a CSP faces: the noisy neighbor. This problem happens when one tenant in a shared virtual server eats up so many resources that its neighbor on the server experiences performance problems. By scaling out in discrete nodes that have a limit to how many VMs each hypervisor can use, there is less chance of a noisy neighbor problem arising.
HCI and private clouds
If your organization plans to set up a private cloud, it would benefit from the ease of implementation and management of using HCI. Such a hyper-converged cloud could have the virtualized resources of storage and compute shared in different ways for different uses. Private cloud storage would need less of the CPU resource and more of the storage. A virtual development server running in a hyper-converged cloud requires more CPU resources than a VM for storage.
The simplicity of setup and management of HCI might come at a premium cost versus buying standard hardware and building a private cloud around it. This can be offset somewhat via the savings in administrator-level staff, since the management of a hyper-converged cloud would require less specialized IT staff.
The hybrid cloud and HCI
The area where the hyper-converged cloud shines is the hybrid cloud. Integrating an on-premises private cloud with a public cloud can be easier than trying to integrate a standard enterprise IT system with the public cloud. The virtualization of hardware resources is at the heart of both platforms.
Vendors such as Hewlett Packard Enterprise and Nutanix sell HCI products that offer built-in data protection services that use the public cloud. Cohesity and Rubrik specialize in this area, selling hyper-converged secondary appliances intended specifically for data backup, which can also be used to back up to the cloud.
Even VDI can take advantage of HCI in a hybrid cloud. Vendors Ericom, Parallels and Workspot provide cloud-based systems that you can use to deploy virtual desktops both to the cloud and on premises.
Integrating the public cloud into an on-premises private cloud requires a robust network. While most on-premises uses of a hyper-converged cloud will work well with a 10 Gigabit Ethernet network, 40 GbE would be better suited for high-performance applications. If you add the latency and throughput issues of connecting to a public cloud, you can eliminate even the advantages of a speedy 40 GbE network infrastructure.