Hyperconvergence. The word itself sounds futuristic, like a term from a Mystery Science Theatre 3000 parody of a 1950s sci-fi flick. Although the word may be as unapproachable as The Forbidden Planet (starring the completely approachable Leslie Neilson), hyperconvergence is no longer a niche architecture that only supports homogeneous virtual desktop infrastructure (VDI) environments. It can now support mainstream, mission-critical production workloads across a vast spectrum of software technologies, and it’s on track for ubiquity.
Like the famed Id monster in The Forbidden Planet, hyperconvergence came to its present place by ruthlessly eliminating provisioning complexity and by delivering economic benefits and resource efficiencies. Looking forward to 2017 and beyond, hyperconvergence will become pervasive—and sooner rather than later.
Today, the adoption rate of hyperconvergence is still in the single digits, so a quick definition makes sense before we go too far. Hyperconvergence is simply an infrastructure and architecture for which compute, storage, networking and virtualization (or containers) are provisioned using a single, central management system.
Although adoption is low right now, the only thing keeping it from taking off is the traditional IT refresh cycle. Every major technology manufacturer now has hyperconverged offerings and is focused on delivering more value into those platforms to help clients simplify their IT operations. How is it possible to simplify something that many believe to be so complex? Hyperconvergence can reduce the amount of hardware, cut the cost of maintenance and guarantee fewer potential points of failure, enabling more data center resiliency. Organizations that have successfully deployed hyperconverged solutions have been extremely pleased with the results.
No matter the type of task or overall technology you work with, there is a hyperconverged offering for your environment. That’s just one of the reasons why hyperconvergence is set to take off in 2017 and beyond. Three other trends will lead toward wide adoption in the coming years.
1. Hyperconvergence Will Become Highly Secure
With the sheer number of security threats out there, it’s no exaggeration to say that security is the number-one priority for most organizations. We learn of a new data breach or cyberattack every day in the news, and these hackers are only getting smarter and more advanced in their methods. They aren’t picky, either. All industries and businesses of all sizes have been affected in one way or another. It follows that hyperconverged architectures can’t gain traction if they aren’t secure.
Traditionally, security has been provided externally through additional technology solutions that surround a company’s integrated environment. With security threats coming from everywhere, this model isn’t enough. Embedded security technologies will become default options from all of the major providers. This approach will address issues such as inspection of east-west traffic flows, encryption and even physical security.
Expanded offerings from firewall, end-point, malware, encryption, DDoS and data-loss-prevention companies will become integrated into the platforms and offered either as bundled solutions or as modular elements to create a security platform that’s further integrated into the hyperconverged architectures.
2. Hyperconvergence Will Adopt Machine Learning
Historically, it’s taken well-trained, well-compensated people to plan for how workloads consume compute resources. As workloads and patterns naturally change over time, and as workloads move for various reasons, ensuring service-level agreements (SLAs) for performance and other metrics requires continuous evaluation, and it risks workload disruptions.
Here is where machine learning can provide major benefits and dramatically reduce the need for human management of IT architectures. As policy-driven and intelligent workload-placement technology adopts machine learning, hyperconverged environments will be able to first recommend and then provide the best possible performance for an organization’s entire IT environment.
The scheduling technology is already deployed in many hyperscale cloud architectures, and as it evolves and commercializes, look for OpenLava, Platform Computing and other on-premises technologies from the high-performance computing world to become integrated into hyperconverged platforms. Then look for them to deploy AI functions to learn from the behavior of the workloads. The real payoff will come, however, when the behavioral patterns of billions of workloads are analyzed in a cloud architecture and digests are then delivered back to on-premises hyperconverged environments to enable not just reactive adaptations, but active autonomic workload management.
3. Hyperconvergence Thrives on Next-Generation Technology
Of course, hyperconvergence isn’t the only thing we’ll be watching take hold over the next few years. Next-generation hardware innovations such as Non-Volatile Memory Express (NVMe), 3D XPoint, Purley and others will also hit the mainstream during this time. It just so happens that hyperconverged architectures are perfectly set up to take advantage of all these innovations have to offer, as well as help deliver the initial promise of these platforms.
Of those listed above, the most compelling is 3D XPoint. When very dense persistent memory reaches the fiscal milestone that meets today’s external flash-storage pricing, hyperconvergence will explode. There will be little incentive to manage an external storage architecture if a completely integrated, hyperconverged solution can solve for almost all commercial workload deployments and include block, file and object architectures at scale.
Unless you are Robbie the Robot (my favorite of all B-movie sci-fi characters), it’s tough to predict the future. It’s particularly true of technology, where a small market move can mean big things for what’s next. The move toward hyperconvergence, however, has already begun, and all the factors point toward it becoming a standard. The more an organization can learn about this technology today, the better off it will be in the next year, and into the future.
About the Author
Lief Morin is CEO of Key Information Systems.