The big idea behind hyperconvergence is the combination in a single system of computational power, data storage and network. A hyperconverged infrastructure (HCI) is easier to manage and to adapt than a regular infrastructure.
How does a hyperconverged infrastructure look like?
A classic data center infrastructure is based on three distinct layers: data storage, computational power and the network that links one to the other. With each layer comes a ‘best of breed’ solution. The infrastructure is the addition of different elements coming from multiple suppliers – with each having its own management software.
In a converged infrastructure (CI), things are simplified. We’re still talking about storage, computational power and networks being separate components, but they’re bundled up into a rack – allowing for centralized management. However, a converged infrastructure is also bound by limitations when it comes to hardware and management.
Silos no more
In a hyperconverged infrastructure (HCI), silos are no more. In their stead, a standard x86 server provides both computational power and storage space. HCI combine the ideas of a converged infrastructure and a Software Defined Data Center (SDDC). Hyperconverged infrastructures are the guarantee of a better integration between a data center’s different parts by using a software. That’s the reason why a lot of people go for an HCI.
The corner stone of hyperconvergence
There a few essential elements to an HCI:
- x86 servers: these modern servers are extremely powerful and have both potent processors and large data storage capacities in the form flash disks for better performances.
- A Software defined storage: this feature allows to virtualize direct attach storage in different segments that can be centrally managed. The software takes care of the host-provisioning, load-balancing and even data security.
- Virtualization: this makes informatic and network layers abstract in order to optimize physical resources. It improves efficiency and security.
- High speed Ethernet: Ethernet switches are used to connects HCI junction points in order to create a clustered and virtual resource group with both a high IOPS and a low latency.
All these technologies are regrouped under one server allowing to create a virtual SAN with higher performance levels, more availability and more resilience than a regular SAN.
The upsides of hyperconvergence
An HCI combines computational power, storage space and a network in a single automated layer with centralized management. This results in improved efficiency. There’s no need to buy single servers and no Storage Area Network (SAN) or Network Attach Storage (NAS) need to be installed or maintained. The system has its own virtualization software and its own management tools – all supported by a single provider.
One of the advantages of combining computation power, storage and a network in a single system is that it makes the infrastructure dynamically manageable et reconfigurable on multiple servers. No expertise required. In the meantime, you reduce clutter and reduce costs for better overall performances for all your primary business workloads.
Businesses wanting to benefit from cloud capabilities for increase speeds, reduced costs and great ease of use – with no compromise on control and security. A hyperconverged infrastructure can help achieve this objective.
According to a report from The Enterprise Strategy Group, HCI users have seen their infrastructure costs reduced by 25%. Daily maintenance times lowered by 31% and deployment time by 52%.
HCI, Edge and SDDC
When talking about HCI, Software Define Data Center (SDDC) and Edge Computing are often mentioned. Let’s look at their similarities and differences.
An HCI can be considered precursory to a Software Defined Data Center. A hyperconverged infrastructure has the same capabilities as a SDDC based on a classic x86 architecture and a storage space. If the configuration and management of the network are virtualized (SDN) while the IT management is automated (Automated), the HCI becomes a mature SDDC.
When it comes to Edge Computing, data center functionalities such as computing and storage are close to end points – such as point of sales and IoTs. This is made possible by using smarter peripheral hardware. An HCI facilitates its management and evolution towards Edge computing.