The number of applications in organizations, and the size of their data sets, continue to grow at an alarming rate. IT professionals are looking for technologies to simplify their lives, and hyper-convergence promises to do just that. Each time a node is added in a hyper-converged data center, the availability of compute, networking and storage resources increases. As new applications or updates to old applications are requested and processed, the new nodes come with seemingly everything an IT professional could need to support the request.
To support all the services being converged, a hyper-converged data center requires not only the right platforms, but:
- CPU power. Luckily, vendors like Intel are far out in front of the need, and multicore, high clock-speed processors are plentiful.
- A solid hypervisor foundation to allocate CPU resources to a variety of applications and processes. VMware leads the way here, and there are numerous quality hypervisor alternatives.
- The ability to turn storage hardware and networking functions into software. There will always be a physical component to storage and networking in a hyper-converged data center, but the more management and allocation capabilities that can go into software, the better.
Hyper-converged data center focuses on abstraction
Abstracting the operating system and applications from the physical CPU is the role of the hypervisor, while abstracting storage and networking services from physical hardware is the role of software-defined storage and networking. When these factors are combined, they are labeled a software-defined data center. When the three software components run on the same physical system, it is known as hyper-convergence.
But abstraction by itself is not enough. The storage software plays a key role in how well hyper-convergence delivers on its promise. The software-defined storage offering must run well within the hypervisor construct and provide access to a shared pool of storage so that virtual machine mobility functions correctly.
Adoption challenges with hyper-convergence
There are many benefits to a hyper-converged data center. It can simplify the management of the expanding data center, creating a natural scale of the most problematic resources. In some cases, the right vendor selection can also reduce the hard costs of data center expansion. But there are several roadblocks affecting hyper-converge infrastructure adoption:
- These architectures are shared-everything and guarantee specific performance to specific applications. If performance guarantees are required, IT professionals need to understand what vendors can provide.
- Many hyper-converged architectures ignore the existing investment in storage and create a new tier in the data center without considering the existing SAN or NAS systems. If the goal is to keep the existing tier of storage, a vendor that knows how to work with that resource should be placed higher on the consideration list.
- Vendor lock-in. Many hyper-converged vendors require organizations to buy both hardware and software from them. While this can lead to a faster turn-up of equipment, it also limits future flexibility.
Introducing hyper-convergence into the data center
See how these enterprises feel about switching to hyper-converged