Way back in the dot.com bubble, the idea of running applications over the internet was born. Companies sprouted up all over Silicon Valley that offered all sorts of services essentially running in a cloud, although it wasn’t called “the cloud” at that time. Unfortunately, the technology wasn’t available to deliver these services and we all know what happened to most of the dot.coms. One of the biggest issues was a lack of broadband connectivity for consumers and businesses.
But over the past dozen years or so, technology has largely caught up with the whiteboard dreams of these companies and the cloud has successfully taken up where the dot.coms failed. Today, most people have one or more email accounts, online banking, file sharing, photo sharing, etc, etc. And for businesses the result is largely the same, as the cloud is slowly replacing more and more on-site applications.
This decision to move from the corporate data center to the cloud usually involves some kind of detailed cost analysis to justify the disruption and potential risk. For many companies this was a TCO analysis comparing the costs of on-site infrastructure with that of services from providers like Amazon, Microsoft or Google. Based on its popularity with many companies, it would seem the TCO of the cloud was indeed compelling.
It took a while but the cloud has become a mainstream alternative for many IT organizations. This popularity and the amount of coverage cloud-based services have gotten seems to have fostered a “cloud assumption”, the premise that somehow a company couldn’t possibly compete with the cloud. As the thinking goes, cloud providers have the economies scale and the resources companies can’t match, basically, the expertise and the headcount to run all the infrastructure that’s needed. But is that really accurate?
Is the cloud a more economical solution than anything a company could stand up in their data center? With some of the changes in the IT Infrastructure space, specifically the convergence of open systems servers, storage and hypervisors, maybe it’s time to look at the details behind the cloud assumption.
This is what Evaluator Group did. We looked at the costs associated with buying, implementing and operating a hyperconverged infrastructure and the cost of running a similar system in Amazon’s AWS.
As it turns out, the components that make up a TCO model for IT infrastructure, things like design, evaluation and installation are quite a bit different with a hyperconverged system than they are with traditional open systems compute and storage. And after these systems get up and running, the level of effort required to operate, maintain and especially, to upgrade them is also much different.
Another aspect that’s changed quite a bit in the last decade is the proliferation of server virtualization, a technology that makes server consolidation possible, but also enables the mobility of workloads. Closely related is the disaggregation of software and hardware made possible by software-defined storage technologies. Together these developments have given rise to “commodity” server hardware, one of the economic drivers in hyper-scale infrastructure pioneered by the big cloud vendors.
Hyperconverged appliances combine storage, compute, hypervisor and a host of advanced features and functionality into a modular package that makes implementation simple and scaling much less complex than with traditional infrastructure. They also leverage the aforementioned economics of industry-standard server hardware to keep costs down.
In this White Paper, “Is Hyperconverged Cost-Competitive with the Cloud?”, Evaluator Group goes into the details of a TCO model used to compare the AWS cloud with a SimpliVity OmniCube hyperconverged solution. The results clearly show that long-held assumptions about the economic advantages of the cloud need to be reconsidered.
Many products have long lists of features that sound the same but work very differently. It’s important to think outside of the checkbox of similar-sounding features and understand how technologies and products differ.
Back to Analyst Blogs