Cloud-First is Often a Mistake. Here's Why.
Enterprises that have predictable and stateful workload might be better served using on-premise computing options.
Join the DZone community and get the full member experience.Join For Free
For some enterprises, a “cloud-first” policy can seem like a no-brainer, especially when compared to the quagmire of traditional data center infrastructure. Yet new software-defined infrastructure solutions like hyperconverged infrastructure (HCI) also offer IT agility, as well as greater security and control than what’s available in a public cloud. Perhaps surprisingly, many actually cite cost as the key incentive for using public cloud, despite the fact that, in most cases, it is significantly more expensive than on-premises HCI solutions like Enterprise Cloud.
IDC published a study that found predictable workloads, which account for the majority of all enterprise workloads, on average were about twice as expensive to run in the public cloud as compared to running on-premises on Nutanix. And a 2018 IDC survey entitled Cloud Repatriation Accelerates in a Multicloud World reported that 80 percent of organizations had repatriated applications out of the public cloud back to on-premises, and that 50 percent of all public cloud applications installed today will move back on-premises over the next two years.
In his latest book, The ROI Story: A Guide for IT Leaders, Steve Kaplan explains that some enterprises make incorrect assumptions about overall cost due to the lower upfront capital expense of public cloud. Yet CAPEX is merely one element of the larger equation, and IT decision-makers should do a sound financial analysis that accounts for all of the factors that add up to the true cost of IT investment.
When comparing TCO of public cloud versus HCI, Kaplan recommends that you consider the following:
Rent vs. Buy. This is the best-known consideration—if you plan on using a service for a short time, then renting probably makes the most economic sense; for the long term, though, buying is often much cheaper. For some elastic, burstable workloads, public cloud may be a good option. But many, if not most, enterprise workloads are persistent and predictable, and thus usually far more economical to run on HCI.
Staffing Requirements. Most organizations do not fully consider the investment necessary to properly staff for cloud expertise, including cloud security, cloud databases, cloud networking, and cloud monitoring, as well as the expertise required to migrate to public cloud.
Unpredictable Costs. Thanks to highly complex pricing models, many organizations suffer “sticker shock” when their monthly cloud bill arrives. The Nutanix Enterprise Cloud Index, which is based on surveys of 2,300 global IT decision-makers, has found that a mere 6 percent of the respondents using public cloud services said they stayed under budget, while 35 percent overspent.
Performance Concerns. Kaplan points out that many companies suffer performance issues while migrating mission-critical applications. Latency issues between on-prem and off-premises applications, dependencies on other apps not in the cloud, and resource contention—the “noisy neighbor” effect—can all be a problem. These issues generate costs, from VM and storage usage, to data egress charges, and more.
Security Considerations. Public clouds often share hardware among many customers, which can make them more vulnerable. Deloitte’s David Linthicum found that “one in four organizations using Infrastructure-as-a-Service (IaaS) or Software-as-a-Service (SaaS) have experienced cybersecurity threats that compromised some data. Moreover, one in five were infiltrated by advanced attackers targeting their public cloud infrastructures.” A secure enterprise must resolve these flaws and vulnerabilities, which add another often unexpected expense.
So, is private, enterprise cloud a better choice? Perhaps hybrid cloud or multi-cloud? Kaplan reminds us that when it comes to something as vital as enterprise infrastructure, the only “no-brainer” is to do a thorough financial analysis, one that accounts for the full range of variables associated with disruptive technologies like public cloud and hyperconverged infrastructure.
Published at DZone with permission of Jordan McMahon. See the original article here.
Opinions expressed by DZone contributors are their own.