More Reasons for HCI at the Edge

Like the proverbial runaway home improvement project, IT initiatives can quickly take on a life of their own – and amass an array of suppliers to manage in the process.

The adoption of a single public cloud service for CapEx savings, for example, can fast morph into an IT optimization project that spans a proliferating number of infrastructure options. Enterprises may end up with many best-of-breed public cloud services, private cloud infrastructure and, increasingly, remote or edge computing infrastructure. This leaves many searching for a way to manage the growing demands and complexity of these different systems, including the need for real-time processing, data storage and governance requirements.

“The computing pendulum is swinging away from centralized hyperscalers in 2022, in the form of edge computing,” Anay Nawathe, principal consultant at research and advisory firm ISG, told TechTarget earlier this year. 

“Organizations increasingly employ edge computing across a network of data centers to provide high-performance workloads closer to the end user,” he said.

But how many different IT infrastructures, cloud platforms, locations, and vendors can a given enterprise reasonably deploy and manage in parallel? And at what cost?

Consistency is Key 

Statista reported results from a poll that show as of 2020, 69% of companies using hybrid clouds that it polled had adopted hyperconverged infrastructure (HCI) platforms or were in the process of doing so to make mixed-cloud management easier. HCI combines server computing, storage and networking capabilities into a software-defined platform designed to easily scale for changing needs.

As HCI matures and proliferates, it's moving beyond centralized data centers to edge computing. That’s because the core capabilities of HCI, such as provisioning, monitoring, management and on-demand scaling, can significantly reduce the complexities associated with edge computing.

Earlier this year, CRN cited a 451 Research study indicating that in 2021, 33% of organizations had deployed HCI in remote and branch office (ROBO) locations, up from 19% in 2020. ResearchandMarkets expects a compound annual growth rate (CAGR) of nearly 39% for edge computing through 2030.

HCI’s tight hardware component integration makes the technology a strong fit for small yet powerful edge sites that can spring up anywhere data is collected, generated and needed for employee or customer access. Its virtualized architecture enables the software-defined, unified management of IT resources remotely across sites, which is more efficient and less error-prone than manually configuring and maintaining separate hardware appliances.

Why the Edge Now?

While ROBO sites have been around for some time, there are a few drivers now pushing new types of edge solutions into mainstream deployment and management practices. One is a function of companies seeking higher-performance IT support for distributed customers, stores, and the large remote workforces created by the pandemic. Local infrastructure that’s geographically closer to users helps deliver that improvement by reducing distance-induced latency. There is an increasing variety of needs from thick (enterprise edge) to thin (and IoT) edge approaches.

In addition, an escalating number of real-time applications rely on local sensors, cameras, processing and storage in edge locations, many of which are unmanned and wholly contained. In particular, edge data centers often process streaming data to support emerging Internet of Things (IoT) and artificial intelligence (AI)-driven analytics and automation applications.

Some edge sites are served by emerging 5G wireless networks, which are highly distributed by design. 5G connections may carry partial data to locations with greater compute power for aggregated, multisite analytics.

The different types and sizes of edge data centers depend on whether the location is autonomous or connects to other sites, shares resources with other edge deployments and serves users or simply collects or generates data for analytics. And with increasing business-critical workloads at the edge, there is an increasing need for better availability, data protection and disaster recovery there too. Whatever their purpose, enterprises need a way to seamlessly integrate them into their IT support and management infrastructure.

Using HCI to Tackle Challenges

Greg White, senior director of product and solutions marketing at Nutanix, described infrastructure cost, management complexity, and mitigating risk as primary ROBO/edge challenges.

“Currently, there are islands of infrastructure with server, storage, network, hypervisor, container, and backup resources provided by different vendors and managed separately,” said White.

“Customers are also running many different workloads and applications at these sites. It’s been challenging to support their requirements with the same infrastructure, often resulting in many different appliances on site.”

As enterprises’ edge locations grow from tens to hundreds to thousands of sites, this approach no longer scales, he explained. 

“At ROBO sites, there often isn’t a dedicated IT person,” he said.

“Managing sites remotely, if it can be done at all, with different management platforms can lead to errors and makes it difficult to respond quickly to issues or adapt to changing business needs.”

Consolidating infrastructure and application management on a single platform and eliminating multiple management consoles using an HCI-based approach enables simplicity, flexibility and rapid scalability.

“Remote management is especially powerful when customers are managing multiple locations, particularly if they can use the platform to perform upgrades, add applications, and enforce security and data protection without local resources,” said White.

Amid a growing number of infrastructures to manage, enterprises also need affordable, scalable storage capacity to address an expected tripling in the amount of data created in the world between 2020 and 2025. By then, Gartner predicts that three-fourths (75%) of enterprise-generated data will be created and processed outside a traditional data center or cloud.

HCI-based platforms that unify object, file and block storage under the management and same software-defined infrastructure as their VM and container-based workloads help simplify and lower the cost of storing multiple types of data in smaller locations, White said.

Disaster Recovery at Edge Sites

Lastly, said White, with few or no IT resources in edge data centers, customers struggle to mitigate risk and downtime.

“They’re challenged to maintain the same security and uptime as the central data center and minimize the impact of breaches and unexpected outages on users and the business,” he explained.

For self-contained edge and ROBO sites where speed of recovery is critical, White suggested, companies can maintain a local disaster recovery cluster. There are flexible options to meet different needs too, from running a few VMs to having all the workloads available again in a specific recovery time objective (RTO).  Integrated local snapshots and clones are very helpful here as well as backup software from partners. 

For sites with good WAN connectivity, customers can back up their most critical applications and data to their main data centers periodically. Another option is to back up essential data from edge devices to the cloud; however, the connectivity and amount of data will impact how quickly they can recover. Often this is a limiting factor on recovery speed as it can take significant time to move data back on-site from the cloud. An alternative possibility is to back data up to another, nearby edge location.

IT shops with distributed edge sites that are likely to multiply need to plan how they will integrate those locations into their hybrid cloud infrastructure and platform services strategies. White advised treating these remote sites as an extension of the data center, “not a separate silo.”  

He noted that an HCI platform such as the Nutanix Cloud Infrastructure integrates management, data backup and storage of different data formats. Such architectures, he said, enable enterprises to simplify and scale their IT environments across private clouds, hosted/managed clouds, public clouds, and now, the increasing diversity of remote edge and ROBO locations needed for local IT performance and storage that satisfies changing business demands.

Joanie Wexler is a contributing writer and editor with more than 25 years of experience covering the business implications of IT and computer networking technologies.

© 2022 Nutanix, Inc. All rights reserved. For additional legal information, please go here.