dcsimg
 

Hyperconverged Infrastructure is Reshaping IT Management

by Drew Turney
article-image

The emerging software-defined infrastructure model is redefining approaches to managing hardwired systems.

Hyperconverged Infrastructure (HCI) is the natural evolution of converged infrastructure, the practice of controlling and changing computer, network, and storage functions in the same server or architecture. Hyperconvergence supercharges this process by using only software to build and deploy the elements of your information and communication technology (ICT) architecture virtually across a network, usually through a hypervisor environment, rather than in discrete and separate units of hardware, giving you much more control over customization. It can be run using off-the-shelf servers like you'd find at any mass market cloud or data center provider.

The most common use case for the enterprise is to simplify infrastructure, particular for large players who might have teams, offices, servers, and equipment at multiple locations. 

A recent study found that HCI had been used primarily for new net-native applications when it first gained traction, but  it's now often used to replace existing hardware systems as they cycle out of use.

So far, the most widely used enterprise applications of HCI have been databases. HCI offers the ability to more tightly enmesh management, particularly with easy-to-use hypervisor interface technology, with data protection and disaster recovery protocols. They're all built right into the DNA of the system, making it easier to protect widely dispersed datasets.

Also read: The Growing Value of a Microservice Architecture

HCI Advantages

The primary advantage of HCI is ease of access. With all the elements of your infrastructure defined by software and accessible through a single management environment, control and deployment of resources can be shared or centralized as IT managers see fit according to the needs of particular departments of teams.

The ability to “spin up” a complete copy of the business gives you the chance to rigorously evaluate it for statistical, sales, or marketing opportunities without using or compromising live customer or company data.

There’s also the ability to scale quickly. Imagine you publish a mobile game an influential celebrity likes on Instagram and you’re suddenly faced with millions of downloads and players. In a HCI world, you need only allocate more resources to database management, compute and storage on the fly to cope with the explosive growth. By contrast, if you have to wait even 24 hours for a new physical server to come online, you might have missed your window.

Challenges of HCI

Although they're gradually falling away, there are still downsides to HCI. For instance, there's an argument to be made that it is unsuitable for wide-scale deployment. When you have dozens of servers storing dozens of terabytes, segregated subsystems might be preferable because cost savings can be found in their discrete functions without committing the entire business to the same solution.

Which leads to the matter of vendor lock-in. The HCI market tends to be an all-or-nothing proposition, and although you may have control of your infrastructure built through virtualization control, there's a lack of transparency about what's actually inside the black box at the vendor end.

If variables crop up that would be better served by separate compute or data instances, good luck wresting control of your resources back from an iron-clad vendor contract.

Also read: Understanding the Role of Enterprise Data Fabrics

Effective HCI Deployment

The main change factor facing HCI is customer demand. HCI deployments tend to be rigid with little room to either make changes or connect other services or infrastructures. But in the increasingly hybridized enterprise, many organizations want the ability to connect public or other cloud environments to their HCI builds to port workloads as they see fit or assure redundancy in the case of a data disaster.

There'll also be more emphasis behind the scenes on smart network automation. Because of the lack of control organizations have with the inner workings of a vendor's HCI system, they expect exceptional network performance. Algorithms, some of them managed by machine learning, that smartly direct traffic and computing power to the most demanding processes on a 24/7/365 basis will be the most basic tenet of any HCI offering. If customers can't be guaranteed of uptime and efficient use of network resources by one provider, they have plenty of other choices among other providers who can.

Like most anything ICT-related, HCI is about agility and, ultimately, cost savings. Anecdotal evidence suggests most HCI builds take days to deploy rather than the months it would take to develop a completely virtualized, software-driven infrastructure yourself.

It's also a way of effectively outsourcing your entire computing and storage needs. With HCI, you can negotiate a minimum SLA standard that will ensure your continued ability to operate, set and forget while looking forward to new projects and goals.

Read next: Transforming Networks: From Virtualization to Cloudification

This article was originally published on Monday Feb 22nd 2021