Do we really have to go over this again? Based on a new study from ServerPronto University (yes, really), private cloud (read: legacy data centers dressed up in cloudy clothes) are 3x cheaper than Amazon Web Services (AWS). Dell founder Michael Dell took the bait, touting EMC-Dell-based VxRail as 2-4x cheaper than AWS.
It's a nice thought, but reminds me of Trautman's declaration to Rambo: "It's over Johnny. It's over!" to which Rambo replies: "Nothing is over! Nothing! You just don't turn it off!"
Meaning, the silly cloud price war is over and, really, it never begun. The cloud has never been a question of cost, but rather one of convenience.
Garbage in, garbage out
Even so, it's worth digging into the cost claims, if briefly. ServerPronto is (wait for it!) a dedicated server hosting company. That means it gets paid to push servers on enterprises, even as the world increasingly thinks about serverless computing (as the natural extension of cloud computing, wherein the server disappears entirely and only services/functions matter). One other post it published recently suggests a disconnect from reality: The Simple Reason Companies are Abandoning the Cloud.
Because, um, that's happening?...
If you look at ServerPronto's numbers, it's a wonder that anyone would ever consider running anything in the public cloud. After all, the company finds that AWS costs $2,762.81 a month for a comparable configuration, while the private cloud offering costs a mere $899 a month (even the pricing is optimized for optics—it's not $900 per month. It's $899).
SEE: AWS isn't the cheapskate's cloud, and Amazon doesn't care (TechRepublic)
Things enter bizarro-land, however, when ServerPronto spells out the reasons private server (I mean, "cloud") hosting manages to be so much cheaper:
The difference between public cloud hosting and private cloud hosting is simple. A public cloud is utilized by multiple tenants who each rent out some of the cloud's resources. This is the service offered by Amazon Web Services (AWS), Microsoft's Azure, Google Cloud Platform and more. A private cloud is a cloud setup which is utilized by a single tenant. The difference is small, but it does make a big impact.
Well, yes. But that "impact" is completely in the public cloud's favor. It's a truism that private servers sitting in an average enterprise data center get used just 5-10% of the time. ServerPronto's $899/month covers just a fraction of what you'd need to pay to get remotely close to the public cloud's levels of utilization.
The cost of convenience
The company might respond: "But that's not the point! Even with our profound waste of money, energy, and materials to bulk up a data center, we're still cheaper!" To which I'd say: "Doubtful at best, but irrelevant, anyway."
Irrelevant, because enterprises aren't simply buying raw storage and compute from the public clouds. They're buying into Amazon Aurora, Google BigTable, Microsoft Azure Machine Learning, etc. They're buying services and convenience.
SEE: Why public cloud R&D is making lock-in worse, not better (TechRepublic)
Is that cheap? No, but it's demonstrably cheaper, for example, to run big data workloads on public clouds than on dedicated servers. Why? Because the very nature of data science requires elastic infrastructure, as AWS product chief Matt Wood told me:
Those that go out and buy expensive infrastructure find that the problem scope and domain shift really quickly. By the time they get around to answering the original question, the business has moved on. You need an environment that is flexible and allows you to quickly respond to changing big data requirements. Your resource mix is continually evolving—if you buy infrastructure, it's almost immediately irrelevant to your business because it's frozen in time. It's solving a problem you may not have or care about any more.
Even if all an enterprise buys from the public cloud is storage and compute, it's going to be cost-advantageous compared to bulking up on under-utilized, quickly obsolete servers. But, as mentioned, enterprises are increasingly looking to the cloud for the next stage of convenience, including powerful services they can rent by the hour (or minute) instead of signing long-term contracts for dumb infrastructure that requires the extra cost (and expertise) of server-side software.
As former VMware executive Mathew Lodge tweeted, the ServerPronto University study "Displays a staggering lack of understanding of the drivers for public cloud." That, or a profound need to keep peddling servers in a world that increasingly doesn't care. To be clear, there are very good reasons to run private clouds, as I've written before, but saving 3x on infrastructure is not one of them.
- Data science demands elastic infrastructure (TechRepublic)
- Why public cloud R&D is making lock-in worse, not better (TechRepublic)
- Here's the single biggest reason telcos have failed in the cloud (TechRepublic)
- AWS isn't the cheapskate's cloud, and Amazon doesn't care (TechRepublic)
- Oracle hopes to win the cloud by going cheap on data centers. Good luck with that (TechRepublic)
Matt is currently head of the developer ecosystem at Adobe. The views expressed are his own, not those of his employer.
Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.