What Are Converged and Hyperconverged Infrastructures?
In the old days of computing, a huge mainframe dominated the data center. Then came the microprocessor revolution, and the one became many. A separate box, costing very little, took care of each distinct service. Even processing and storage separated, with the advent of network attached storage (NAS). Local networks grew rapidly, with each device having its own IP address and access configuration.
CI: Everything in one package
But having so many pieces became chaotic as computing needs grew. The infrastructure broke up into silos that didn't readily connect with each other. Shifting resources from one purpose to another was difficult.
Administrators wanted a better way to manage all the components. The answer to this was the converged infrastructure (CI). It wasn't really a return to the mainframe. It still consisted of many separate components. But the components became available as a package, often in one chassis. A dashboard managed the entire infrastructure, letting administrators combine processing, storage, and network connections as needed.
The advantages of this approach include simplified procurement, ease of management, a high level of compatibility, and the economy of buying a whole package. Configurations are available to suit particular business use cases. On the negative side, a CI isn't as flexible as acquiring discrete components, and the customer can end up locked in with a vendor.
HCI: Everything virtualized
With CI, the components remain separate, and the administrator can treat them as distinct units. But with the growth of cloud computing, a new way of looking at resources has emerged. Each machine has enough power for many machines. Why not use it that way? Virtualization is the new trend, replacing discrete hardware resources with virtual ones that come and go as needed. The infrastructure is defined by software. This is what's known as the hyper-converged infrastructure (HCI).
HCI sits on top of any compatible hypervisor, or even more than one. It can connect with remote services through their APIs and make them part of the infrastructure at the application level. The infrastructure can span multiple local networks as well as cloud services. A single dashboard manages all of its aspects.
HCI lets an organization use generic hardware for almost everything. A data center can add machines from different vendors, as long as they run a compatible operating system. Each machine contributes whatever combination of processing power, memory, and storage is necessary. Administrators can allocate virtual machines and storage from the control panel without any hardware changes.
Even network components such as switches and firewalls become part of the virtualized infrastructure. This is called network function virtualization (NFV). Subnets can be defined as needed. The network gains far more flexibility than one with hardware-based components.
HCI and the private cloud
Hyper-converged infrastructure and private cloud services are the perfect match. In the classic model of infrastructure as a service (IaaS), the customer can allocate virtual machines up to whatever limits the service sets, but doesn't have control over the infrastructure. The servers are still siloed, even if they're in the cloud.
HCI takes IaaS to another level, giving control over a pool of virtualized resources that can be configured and combined at will. With HCI on a private cloud, an organization has exclusive access to a set of hardware resources and can arrange them into virtual servers, storage units, and network resources as they're needed.
HCI on a private cloud service lets a company achieve the greatest flexibility while keeping costs down. The benefits include more opportunities for automation and orchestration, globally meaningful analytics, and ready adaptation to changing needs. With so many things going for it, it's no surprise that businesses are rapidly adopting HCI.