Since the days of the mainframe, enterprise data centers have evolved through several crucial pivot points. Think the x86 revolution, virtualization, and more recently the cloud.
Organizations see the benefits of the cloud and are eager to achieve them, yet the pull to keep some workloads on premises is strong. However, on-premises architecture can be unwieldy and complex. Traditional data centers have isolated, separately managed silos of compute, storage, and networking resources. These resources are on distinct life cycles, are sold and supported by separate vendors, and are individually managed.
Many organizations look to public cloud offerings thinking they are a quick affordable way to simplify and achieve cloud benefits. But many workloads are not a good fit for public cloud, or would require costly migration. So public cloud is simply not a complete solution for some companies.
Enterprises can still take advantage of cloud-like scalability while using on-premises infrastructure in what could very well be the next major pivot point in data center evolution: hyper-convergence.
You’ve heard of converged infrastructure: it brings compute, storage, and networking together in a single chassis as a unit of infrastructure that is sold as a single solution. VCE Vblock Systems are a typical example. Converged infrastructure is designed for virtualization, and it is easy to scale out. When your business needs more infrastructure, you can simply add more building blocks.
Hyper-convergence is a newer concept which some say is the next logical evolution of the enterprise data center. This post explains the difference between converged and hyper-converged infrastructure and tells you why you might want to pay attention.
What Is Hyper-converged Infrastructure?
As with converged infrastructure, hyper-converged infrastructure combines compute and storage in a single chassis, or node. It then adds a layer of virtualization software that manages all of the resources in a node or cluster as flexible pools. The building-block unit is typically smaller than in converged infrastructure, so you are less likely to overprovision when adding infrastructure.
Where convergence is hardware centric, hyper-convergence is software centric. The approach presumes that the basic constructs of a data center are virtual machines and workloads rather than the server, storage area network (SAN), and networking gear.
Two Hyper-convergence Leaders: Nutanix and Ericsson
Prowess recently had the pleasure of working with two companies that build hyper-converged solutions: Nutanix and Ericsson. Both companies’ solutions enable web-scale architecture that brings to on-premises clouds the scalability and economics that power leading Internet and cloud infrastructures, such as Google, Facebook, and Amazon. Ericsson and Nutanix solutions feature software-driven building blocks containing compute, storage, and virtualization. When an organization needs more infrastructure, it can simply add more building blocks, which are ready to run any workload out of the box. Additional benefits include:
- Ability to centrally manage resources as flexible pools
- Easy deployment
- Data-center simplification
- Predictable performance and scalability
Is your organization ready for the next evolution of the data center? Find out by learning more about hyper-converged infrastructure and Nutanix and Ericsson solutions.