A hyper-converged network is a network that combines compute and storage network resources into a single preconfigured stack, server or brick with heavy use of virtualization.
Typically, a hyper-converged network is built on server hardware that uses virtualized storage, unifying all hard disks or SSD’s into a single storage stack. These servers are often supplied, managed and supported by vendors. Additional capacity can be added by adding another node or server box and connecting it through the virtual-SAN-appliance. These nodes allow management by an administrator or virtual service.
For IT, these preconfigured arrangements offer greatly reduced setup time compared to making individually bought components interoperable. Setup and support requirements are so reduced that many predict that hyper-convergence will likewise reduce the demand for IT staff.
Hyper-convergence has been driven by the change and increase in data traffic on enterprise networks. Trends in office communication have changed the nature of workloads. For example, in UCC (unified collaboration and communication), data can go between text, audio, video and even virtual white boards. In a network with compute and storage separate, these dynamic workloads can result in congestion. Hyper-convergence helps address these issues with something IT old timers may find ironic: a return to the centralized, multi-purpose server.
See a video introduction to hyper-convergence: