What is Hyper-Converged Infrastructure?

Jonathan Mizzi is the manager of Alert Digital by Deloitte Data Centre. For more info, go to www.alert.com.mt

The term Hyper-Convergence Infrastructure (HCI) grew out of the concept of Converged Infrastructure (CI) where it brought compute, storage, networking and server virtualisation into a single chassis which could centrally manage. The hardware which one gets in a converged infrastructure bundle is pre-configured to run whatever workload it is bought for; whether to support a virtual desktop infrastructure (VDI), a database, a specific application or something else. The main limiting factor with CI, is that there is not much flexibility to alter that configuration once this is live. Adding components separately becomes complex, taking away from many of CI’s benefits. Moreover, adding desktops and capacity to in-house infrastructure can be just as expensive, thus suggesting the importance of proper planning for any deployment.

Hyper-convergence is the latest step in the now multiyear pursuit to build an infrastructure that is flexible and simpler to manage. HCI seeks to do the same as CI, which is to collapse compute, storage, and networking into a single SKU, with the added benefit of throwing in a software-defined storage and not placing much emphasis on networking. The idea is to keep focus on the data control and management layers.

HCI has become so popular in such a short space of time due to the pressure on IT departments to deliver faster turnaround in the provision of resources. More and more applications are best-suited for scale-out systems which could be built using commodity components. Today’s HCI can come as an appliance, a reference architecture, or as software that is flexible in terms of the platform it runs on. Furthermore, software-defined storage promises great efficiency gains with an unprecedented data volume.

Although HCI offers an array of benefits, companies currently seldom use it to deploy generalpurpose workloads, virtual desktop infrastructure or analytics (like Hadoop clusters). In even fewer instances, HCI underlies private or hybrid cloud or those agile environments that support rapid software-release cycles. Gartner expects this to change, as the market evolves and users become more familiar with the architecture. They expect it to go from zero (back in 2012) to a $5 billion market by 2019, becoming the category leader by revenue in pre-integrated full-stack infrastructure products.

The fact that HCI solutions are based on proven technology dramatically reduces the risk to one’s business and the migration can be seen as a shift in operational practices rather than a shift in technology. Additionally, hyper-converged solutions won’t bring with them large amounts of innovation and technology, but the bedrock of the solution being proven should inspire confidence. It also means that the migration should be relatively simple, as the virtualisation platform in use in the organisation would most probably already have a mechanism to facilitate it. In the case where a migration is set to cross platforms, for instance a vSphere shop migrating to a hyper-converged platform that uses a fork of KVM, the vendor typically provides tools to help with the migration. Those tools help take existing workloads and swap out virtual hardware, OS drivers, while also automate moving them to the new platform.

With the money and time invested by companies like VMWare, HP, DELL and Cisco amongst others, it’s clear that HCI is of interest. For many, there is a perception that this is still a young market that has a lot of growth and change ahead of it. We expect to see HCI adoption increase rapidly as more mainstream vendors enter the market. VMware’s venture into this space in 2014 was just the beginning and was a big step toward pushing acceptance of the approach. Unlike many other technologies in the past which promised much but failed to deliver, the concept of the Hyper-Converged Infrastructure is living up to its promise.