Hyper-Converged Infrastructure Systems Overview
What are Hyper-Converged Infrastructure Systems?
Hyper-converged infrastructure is a software-centric architecture that integrates compute, storage and virtualization resources on commodity hardware.
From Converged to Hyper-Converged
Converged Infrastructure: The first phase
Before discussing hyper-converged infrastructure, it’s necessary to understand converged infrastructure and the problem that it tries to solve. Traditionally, IT infrastructure has been relatively siloed with different IT groups. These groups are responsible for provisioning, managing and supporting server, storage, virtualization, and network resources. For example, the IT storage team is responsible for all storage hardware and software purchase and support.
The concept of converged infrastructure seeks to break down these silos by combining them in various configurations. This means that compute, networking, virtualization, and storage resources are tightly coupled in pre-engineered packages linked together with software.
How Hyper-converged infrastructure differs
Hyper-converged Infrastructure takes this a step further. While converged infrastructure is separate components engineered to work well together, hyper-converged systems are modular appliances sold by a single vendor. These appliances are usually commodity hardware boxes, containing computing, virtualization, and storage resources. These can be added together for simple scaling; the more appliances, the greater the capacity.
With simple converged infrastructure, a vendor provides a pre-configured bundle of hardware and software which can be used separately. Hyper-converged infrastructure does not allow the components to be used independently. Each appliance is so tightly integrated that it functions as an indivisible unit.
Advantages and Disadvantages of Hyper-Converged Infrastructure
The modular approach of hyperconvergence presents many advantages and a few disadvantages. Among the advantages: ease of scaling, decreased maintenance overhead, simplified vendor management, and lower cost.
The price for these advantages is an inability to tweak the systems. For example, if more storage is needed for a particular cluster but not more compute, it’s not possible to just add storage. Instead a new appliance including all resources must be added. There can also be challenges in breaking down entrenched IT silos, although this is likely to be worth the effort in the long term.
Hyper-converged appliances usually come with single administration console that allows management and control of all resources. The administration console usually provides visibility into compute and storage statistics. It also provides integration with other datacenter tools and cloud orchestration solutions via APIs.