The future of cloud – true private, public or a hybrid – is yet to be decided, but Wikibon research indicates that “true” private cloud is making inroads into the market.
In order for IT departments to deliver business value more quickly in a tech-driven world, it is becoming necessary for enterprises to look to external vendors for cloud infrastructure. But while public clouds generally offer simplicity and low upfront investment costs, concerns about security and long-term storage persist, making on-premises options more attractive to enterprises.
But building an in-house private cloud from scratch comes with its own costs in the form of personnel, time and infrastructure, making “true” private cloud solutions an attractive alternative. In this vision of the future, private cloud vendors would provide highly individualized solutions (including cloud-adjacent datacenters and hybrid cloud) while also ensuring high-quality updates, support and maintenance at a lower cost than managing everything in-house.
Some companies are also trying to bridge the gap with “enterprise” or “hybrid” clouds, although opinions in a recent CrowdChat discussion were mixed on whether these options are actually meaningful alternatives to “true” private cloud.
Private cloud roadblocks
John Furrier, founder of SiliconANGLE Media, posed a foundational question: “What are the key problem statements that practitioners face when looking at #privatecloud?” These pain points provide some of the impetus for the shift to outsourced management of cloud infrastructure.
Dave Vellante, senior analyst with Wikibon, suggested that the “economic pressure,” “agility,” and “more modern” infrastructures of public clouds are forcing IT departments to look externally for solutions, as it takes a not-inconsequential amount of resources to compete with public options. Furrier quoted Floyd Strimling of the PlatenReport: “Finding the resources and talent to build and operate a private cloud,” and he added that the cultural changes necessary to successfully operate a private cloud are “a lot to ask an enterprise.”
What enterprises want
One CrowdChat participant asked about how, exactly, enterprises are currently adopting cloud technology. Are they “building new cloud-native apps directly in the public cloud?” Are they refactoring their on-premises applications to operate in the cloud? Or are they opting to “forklift” their apps directly into the new environment?
Bert Latamore, a freelance writer covering the intersection of IT and business for SiliconANGLE, said that it’s important to remember SaaS contributions as well, given that it “represents well over 50% of public business cloud and is often subscribed to by business, not IT.” Kenneth Hui, of cloudarchitectmusings.com, responded that building new, native apps in the public cloud is often the most likely option “given the cost of refactoring legacy apps or forklifting to public clouds,” although that conversion cost could change in the future.
Leo Leung, VP of marketing at Scality, Inc., agreed, saying that if an enterprise wants to reuse an existing app, virtualizing is the more common option as it increases flexibility. Hui later reinforced this point, asking whether “enterprise cloud” is simply another form of virtualization with a portal and additional professional services. Most of the CrowdChat participants agreed, although several contributors said that virtualization can be a step on the way to full public cloud adoption, given that it allows enterprises to transition away from their IT investments slowly and strategically.
Some users believe that many of today’s definitions of private, public and hybrid cloud will disappear entirely as the overlap continues to grow. Justin Augat, head of global services marketing at Hitachi Data Systems Corp., offered: “Too many of the public cloud characteristics are now available on-prem – and vice versa. Thoughts? What really defines public vs. private? Ownership? Financing? Geography?” He also included ideas like automation, self-service and management – all things that can be achieved using both private and public cloud options.
Participant Peter Herdman-Grant, a database solutions architect, said that the definition used to be very clear: “Hybrid requires an element of both public and private cloud, with private being exclusively on-premise.” But now, some are saying that “private” clouds can be “hosted or even segmented in the public cloud,” blurring the lines.
Augat suggested that a flexible mixture of terms could help businesses rethink the data lifecycle. “I think there are economic arguments for data to reside on-prem, eventually move off-prem (and be readily accessible) – all as part of [data] lifecycle management.” Another solution, offered by Phil Dunn, sales consultant at Oracle Corp., might be to only have one cloud, but allow for differences in where the data and compute actually reside.
The many faces of hybrid
Hybrid cloud is also tough to pin down under a single definition, especially as the big players (Amazon Web Services, Azure and Google Cloud Services) “all seem to approach the problem differently,” according to a comment by Hui. He continued: “AWS is building virtual private clouds, Azure is providing full hardware/software on-prem stack, GCE [Google Compute Engine] hopes federation of on-prem Kubernetes becomes their private cloud solution. Which one will work?”
Strimling said that he hoped to see the Microsoft Azure Stack get more traction, especially since it seems to be geared toward partnering with Microsoft. Hui suggested that Microsoft will “seed accounts with the Azure stack [and] provide parity over time” between the stack and the Azure cloud. Furrier pointed out that Azure seems to be keeping its legacy on-prem, while Oracle seems to be aiming the other way, although, as another commenter suggested, it may simply be that Azure realizes it would be impractical to rewrite old applications, so it must “live in two worlds.”
Impact: Pre-IPO unicorns
Furrier livened up the conversation with a dose of controversy: “Does true private cloud help or hurt the pre-IPO unicorns?” He asked whether Nutanix, Inc. and similar companies were holding off on IPOs because of a soft private cloud market or due to a larger, looming “Arctic Winter” in tech.
Augat suggested that pre-IPO startups may be a special case that doesn’t reflect the field as a whole: “Many newbie startups use public cloud because it’s just easier to ‘get up and running’ (vs. building a DC with no revenue) — but there is a big inflection point as many companies realize costs, SLAs, compliance, etc., require on-prem.”
Strimling, on the other hand, said that he wouldn’t invest in a company that built a private cloud at all. “Unless you are making billions of dollars of revenue, you have no business doing it yourself. Use #AWS and get over it.” But Vellante pointed out that even Uber Technologies, Inc. is “clawing back toward private cloud” – which Strimling acknowledged that it has the revenue to do so, although scaling quickly may be a challenge. “There is not an endless supply of cloud talent,” he said, “and cloud is hard,” a point echoed by other participants.
According to the Wikibon 10-year forecast, true private cloud will see adoption increase over the next decade. Furrier quoted system engineer Geoff Hughes as saying, “There will always be true private clouds based on particular business requirements,” giving private cloud developers a hopeful long-term market.
But as Vellante pointed out, the forecasted $300 billion shift from “traditional infrastructure” spending to public and private cloud could have much broader implications: “$300 billion shift from undifferentiated labor into solutions … this resource can move to creating value” across the industry.
Read full crowdchat below: