Hyperconvergence as a service offers hands-off data-center management

Hyperconverged infrastructure (HCI) has made substantial inroads in enterprise environments, and vendors have responded with new use cases and purchasing scenarios, including an emerging deployment option: HCI as a service.

Conventional HCI combines servers, storage and network resources into a single box, providing adopters with a gateway to simplified, centralized data-center management. HCI as a service (HCIaaS) ups the ante by enabling data-center operators to adopt HCI in a manner that promises to reduce both operational and financial overhead.

Several HCI vendors, including VMware, Nutanix, Dell, and HPE, offer a managed service option, says Naveen Chhabra, a senior analyst at IT research firm Forrester. “It basically turns the capital expenditure and one-time investment into an [operating expense],” Chhabra says. “In most cases, the vendor will also manage the HCI’s day-to-day operations.”

When an organization acquires a conventional HCI, it’s essentially buying or leasing hardware and licensing the software that makes it run, says Fred Chagnon, principal research director, IT service provider sector, for IT research firm Info-Tech Research Group. “HCIaaS combines the HCI product with an outsourced engagement model, providing a capability to procure the necessary infrastructure, as well as have it installed, and operated, by a single vendor.”

Previously, in cases when an organization lacked the skillset necessary to operate the HCI stack, it could outsource the operation to an external IT service provider offering the capability as a service, Chagnon says. HCIaaS vendors are now providing the ability to acquire the necessary infrastructure as well as to have it installed and operated by a single vendor. “If we accept that HCI is the collapsing of traditional server, network, and storage components into an all-in-one infrastructure node, HCIaaS is simply a change in the nature of how this block is procured,” he says.

In essence, HCI as a service eliminates the need for an enterprise to provision, manage and troubleshoot its data-center infrastructure. “They no longer have to manage servers, networks, et cetera,” says Rajiv Mirani, CTO at Nutanix. “Instead, IT teams can focus on their HCI policies to optimally deliver their applications.”

HCIaaS at work

HCIaaS can be deployed on prem or in a public cloud, explains Amr Ahmed, a managing director in the IT advisory and technology transformation practice at EY.

“The real benefits of HCIaaS on prem can be realized much faster [than conventional on-prem HCI] due to simplification of deployment, ease of management and administration,” he says.

In the case of public cloud HCIaaS, the cloud provider masks all administration overhead through its internal automation and orchestration capabilities, allowing administrative tasks to function invisibly in background.

Public cloud HCIaaS deployment provides agility, virtually unlimited scalability and much shorter lead time to production, Ahmed says. On the other hand, HCIaaS deployment using on-prem hardware, running in a private cloud, generally offers much better device/user responsiveness and user experience, but can be challenged by limited scalability, he notes. 

HCI already supports many of the features cloud providers offer, such as multitenancy and deep access control, a security measure put in place to restrict users from manipulating or accessing data from public, unknown and unauthorized devices, says David Linthicum, chief cloud strategy officer at business advisory firm Deloitte Consulting.

HCIaaS builds upon this foundation. “Typically, there’s a centralized console that provides control for all of the components, and you’ll see that as a service,” Linthicum says. “In the HCI cloud, that software tightly couples the components, and the tightly coupled infrastructure is able to remove much of the latency, resulting in faster running systems, generally speaking.”

A hypervisor, also provided as a service, allows users to run multiple virtual machines, making the VMs relatively easy to deploy for new applications and configurations as requirements change over time. “Moreover, multi-node clusters provide redundancy to eliminate downtime, and software-defined storage eliminates the risk of data loss when drives and/or nodes fail,” Linthicum says. Finally, since HCI is a modular technology, adopters can scale-out their infrastructure for additional processing, memory and storage resources as their needs increase. “Keep in mind this all occurs in the cloud,” he says.

Like any cloud-based service, HCIaaS is typically billed on a subscription basis. The management and maintenance of the appliances are handled by the provider. “Customers only pay for what they’re using, including RAM, CPU, and storage,” Ahmed says. “With a cloud model, HCI pricing can be flexible and based on hourly or monthly usage.”

Benefits of HCI as a service

A key attraction of HCIaaS is that adopters aren’t required to buy any hardware or software. “This brings with it many of the reasons we leverage cloud computing: resources on demand, scalability, better efficiency and business agility,” Linthicum says. “The end game is to save money, facilitate the business’ ability to change and grow, and have others handle operations and maintenance.”

The model appeals to organizations that seek the benefits of running operations in the public cloud—including a rapidly elastic infrastructure as well as the advantages inherent in tapping into pooled resources. “They also like the fact that, in the public cloud, they don’t have to run the infrastructure; that burden belongs to the cloud service provider,” Chagnon says.

With on-prem HCIaaS, vendors are positioning their solutions as a “private cloud infrastructure delivered as a service,” Chagnon says. “With HCIaaS, [customers] get all the benefits of the public cloud in a box [that’s] run for them in their very own data center,” he says.

Other important benefits, Ahmed says, include rapid deployment at all hosting locations and a built-in multi-cloud management capability. Adopters can choose from a wide range of pricing models, obtain a reduced build time enabled through automation and standardization, and gain the ability to scale in accordance with business- and location-specific demands by simply adding or removing nodes, Ahmed says.

Watch for infrastructure outsourcing pitfalls

Along with its numerous potential benefits, HCIaaS also presents some notable disadvantages, including a high failure rate associated with infrastructure outsourcing.

“Our research in the infrastructure outsourcing market has shown routinely that almost half, 47%, of all customers who outsource their infrastructure operations to a partner do not realize the full value of that engagement in the first year,” Chagnon reports. Insufficient due diligence leading up to the engagement is a key reason for sub-par savings, he says. “Whatever the case, these engagements are fraught and complicated, and aren’t the slam-dunk they purport to be,” he says.

HCIaaS can tends to fall short when it comes to tasks that rely on physical control, such as when a compliance mandate dictates that data must be stored on premises, Linthicum notes. “Also, you’re relying on the HCI provider to keep the HCI system operational with few outages,” he says. “If they can’t do that, then you’re pretty much at their mercy, or have to pay the migration costs to move back to on-premises or another HCI provider.”

Chhabra warns that one-sided service pacts can also significantly diminish the overall value and appeal of HCI as a service. “Hardware and services are not available by the drink,” Chhabra says. A customer is typically locked into a defined-term contract with its service provider. “Firms think they can exit out of the contract any time like they could in the public cloud, but it’s not that easy and simple in a private on-prem deployment,” Chhabra explains.

The as-a-service model also doesn’t relieve enterprises of the need to develop a comprehensive data center strategy. “Firms must still plan for the applicability of the technology for their business use-case,” Chhabra says. “Firms must [also] plan for capacity expansions, [and] be more involved in those decisions and operations.”

HCIaaS use cases

HCIaaS can be particularly attractive to organizations that would like to leverage cloud technology but are currently unable to use a public cloud due to data residency and/or minimum latency requirements. “Nor do [many organizations] possess the technical aptitude to run and operate their own private cloud,” Chagnon says.

Just about any task that’s appropriate for an on-premises HCI deployment could also qualify as a suitable HCIaaS use case.

“Anything that requires a great deal of performance and needs to avoid latency would be in play, such as banking systems, data processing and analysis, transaction processing—even machine learning and AI,” Linthicum says. Other use cases include virtual desktop infrastructure (VDI) deployments, and private/hybrid cloud and edge/remote localization services, Ahmed states.

Some of the most popular use cases among Nutanix customers are capacity bursting, data center consolidation, lift and shift of legacy application, disaster recovery, application development and VDI delivery, Mirani says.

Future outlook: Cloud-based HCI

HCIaaS’s immediate future appears to be anything but cloudy. “Going forward, we believe more customers will adopt a subscription-based model, as it will allow them to be flexible, only pay for what they use and potentially reduce hardware costs, similar to a public cloud model,” Ahmed says. “Cloud-based HCI will become a preferred solution by which organizations can get out of the data center business and improve their IT operation cost, agility and time to market capabilities.”