Hyper-convergence: the good, the bad and the ugly

Hyper-converged technologies are gaining traction, but what can they do for your organisation

As organisations look beyond cloud for the next technology trend to boost their business efficiency and agility, an investment in hyper-converged infrastructure (HCI) is now on the radar for many.

According to market watcher IDC, sales of HCI grew by 104.3% year on year in the third quarter of 2016, generating $570.5m worth of sales, amounting to 22% of the total converged systems market value.

Meanwhile, figures from rival analyst house Gartner suggest the market for HCI will be worth $5bn by 2019, with the technology gaining traction in enterprises as more businesses investigate its potential benefits.

In a recent Computer Weekly article, Simon Robinson from analyst 451 Research noted that the presence of incumbent IT suppliers in the HCI market meant the technology was becoming more mainstream (see box, p20).

Step up from convergence

But before any organisation embarks on a move to adopt HCI, it needs to get a firm grasp on what the term really means. It is generally accepted that HCI is a step up from converged infrastructure, which, in basic terms, is where compute, storage, network and virtualisation components are brought together and packaged as a pre-tested and integrated offering.

HCI builds on this concept by including components with little distinction from each other and which are further combined with software-defined functions that run on these appliances. When capacity dries up, you simply add more systems.

When upgrading or changing IT systems, disruption must be kept to a minimum, because every second of downtime represents lost money.

Albie Attias, managing director of IT supplier King of Servers, says hyper-convergence counters this by offering an out-of-the-box infrastructure that is readily scalable. “Its quick, easy deployment is very attractive to enterprises as the drain on IT resource to install and configure the hardware is minimal,” he says.

Tackle rising costs

Hyper-convergence can also tackle rising IT costs, says Matt Foley, director of Hewlett Packard Enterprise’s European solutions and technology group, because entry-level commitments are very low, with users able to start scaling at just two nodes.

“Due to their all-in-one configuration, hyper-converged systems allow you to build, scale and protect IT infrastructure more affordably and effectively than any other option available,” he says. “And software-defined intelligence reduces operational management, providing automated provisioning of compute and storage capacity for dynamic workloads.”

Neil Thurston, chief technologist of hybrid IT at service provider Logicalis UK, cites simplified maintenance as another advantage of HCI. “The stack self-builds itself initially, and when new units are added the stack self-patches the hardware and software layers of the platform and self-upgrades the software,” he says.

The need for data protection is also catered for in HCI, says Thurston. “Data is typically spread across the stack using erasure coding, which is far more efficient on disk than traditional storage array methods of replication and keeping multiple full copies of the same data.”

Another benefit of HCI is that the hardware is often simpler to manage, says John Abbott, founder and research vice-president at analyst 451 Research. “The idea is to eliminate the need for specialist expertise in storage, servers and network management,” he says.

“To avoid the introduction of yet another management console, plug-ins and APIs [application programming interfaces] will increasingly be developed for integration with existing toolsets. As support for multiple hypervisors becomes more common, more consistent virtual machine-level management tools will emerge that work in the same way across multiple hypervisors, be it ESXi, Hyper-V or KVM.

“Infrastructure is made up of standard x86 server technology without the need for external storage. Everything is accessed in a standard way through the virtual machine hypervisor, which helps to reduce complexity.”

Security concern

But although adopting HCI has many benefits, enterprises must also be mindful of the drawbacks, one of which is security, says Liviu Arsene, senior e-threat analyst at Bitdefender.

Such systems will face attacks against the control plane, data plane and management infrastructure, and mitigating these threats will not only require understanding the environment in which they operate, but also an appreciation of how this may affect the security tools’ functionality, he says. “A security architecture that is able to support automated infrastructures while being able to spin up and spin down, maintaining performance, needs to be an integral part of the migration toward hyper-convergence and software-defined datacentre environments.”

These drawbacks can be overcome by using security tools that are compatible with any type of hypervisor technology, as this guarantees they can be deployed and scaled regardless of how many nodes are added to the hyper-converged infrastructure.

Changing landscape

Gradually, hyper-converged infrastructure software platforms are beginning to support a broader range of hardware from multiple sources and hypervisors, says 451 Research’s Abbott. “The initially limited recommended workloads are also broadening to include analytics and mission-critical enterprise applications,” he says. “Test/dev, DR [disaster recovery], Tier 3 and Tier 4 workloads are common, but adoption is also growing in more business-critical applications, such as Oracle, SAP and Exchange.”

There is also an emergence of more specialised forms of hyper-converged infrastructure, optimised for specific rather than general-purpose use cases – Hadoop-based analytics, for example.

In the future, we should also expect hyper-converged platforms to support microservices and cloud-native services and also to support platform as a service (PaaS), allowing organisations to create and run applications without being troubled with the core infrastructure and platform services or locationn.

Hyper-convergence goes mainstream

Hyper-convergence is now becoming a core weapon in the armoury of suppliers looking to demonstrate to customers that they can meaningfully change the fundamental economics of running core and edge IT infrastructure.

This is not only a lynchpin of any IT transformation strategy, but absolutely essential for any supplier that wants to demonstrate ongoing relevance in the era of public cloud.

In January 2017, Hewlett Packard Enterprise announced it would pay $650m cash for seven-year-old startup Simplivity.

VMware recently said it has exceeded 7,000 customers for VSAN, the hyper-converged infrastructure software that is optimised for VMware environments.

Meanwhile, the combined Dell-EMC has highlighted hyper-converged infrastructure as a key growth market.

Other major players are still hoping to make a run at hyper-converged infrastructure, including Cisco with HyperFlex, and there are still startups with momentum, including Pivot3 and Scale Computing.

One crucial test of whether a market is becoming mainstream is the extent to which it is embraced by large incumbent suppliers. As far as hyper-converged infrastructure is concerned, we are now approaching that point.

Source: Simon Robinson, 451 Research.

This was last published in June 2017

CW+

Features

Enjoy the benefits of CW+ membership, learn more and join.

Essential Guide

Essential guide to optimising hybrid IT infrastructure

Join the conversation

1 comment

Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.

As the writer Rene notes, hyperconvergence has many advantages, but the complex nature of a hyperconverged infrastructure and its tightly integrated compute, storage, networking and virtualisation resources, can cause performance issues. Applications that appear to work independently of each other are in fact connected and a spike in activity in one application, can cause a slow-down or outage in another. And if a business-critical application is impacted, that can have serious implications.

I’m not suggesting that hyperconvergence should be rejected – but there is a strong argument for monitoring and managing the performance of applications across the infrastructure. By getting a full view across the infrastructure – not merely how each component is performing on its own – app-centric infrastructure performance management gives IT teams real insight into which applications are under pressure and why. This gives them much more control over their hyperconverged infrastructure.

Chris James

Virtual Instruments EMEA marketing director

Cancel

-ADS BY GOOGLE

SearchCIO

SearchSecurity

SearchNetworking

SearchDataCenter

SearchDataManagement

Close